, Ned Horning [aut],
Jakob Schwalb-Willmann [aut]
, Robert J. Hijmans
[ctb] , Konstantin
Mueller [aut, cre] | Title: | Remote Sensing Data Analysis |
|---|---|
| Description: | Toolbox for remote sensing image processing and analysis such as calculating spectral indexes, principal component transformation, unsupervised and supervised classification or fractional cover analyses. |
| Authors: | Benjamin Leutner [aut]
, Ned Horning [aut],
Jakob Schwalb-Willmann [aut]
, Robert J. Hijmans
[ctb] , Konstantin
Mueller [aut, cre] |
| Maintainer: | Konstantin Mueller <[email protected]> |
| License: | GPL (>= 3) |
| Version: | 1.0.0 |
| Built: | 2024-11-22 06:54:27 UTC |
| Source: | CRAN |
extracts five classes from QA band: background, cloud, cirrus, snow and water.
classifyQA( img, type = c("background", "cloud", "cirrus", "snow", "water"), confLayers = FALSE, sensor = "OLI", legacy = "collection1", ... )classifyQA( img, type = c("background", "cloud", "cirrus", "snow", "water"), confLayers = FALSE, sensor = "OLI", legacy = "collection1", ... )
img |
SpatRaster. Landsat 8 OLI QA band. |
type |
Character. Classes which should be returned. One or more of c("background", "cloud", "cirrus","snow", "water"). |
confLayers |
Logical. Return one layer per class classified by confidence levels, i.e. cloud:low, cloud:med, cloud:high. |
sensor |
Sensor to encode. Options: |
legacy |
Encoding systematic Options: |
... |
further arguments passed to writeRaster |
By default each class is queried for *high* confidence. See encodeQA for details. To return the different confidence levels per condition use confLayers=TRUE.
This approach corresponds to the way LandsatLook Quality Images are produced by the USGS.
Returns a SpatRaster with maximal five classes:
| class | value |
| background | 1L |
| cloud | 2L |
| cirrus | 3L |
| snow | 4L |
| water | 5L |
Values outside of these classes are returned as NA.
If confLayers = TRUE then a RasterStack with one layer per condition (except 'background') is returned, whereby each layer contains the confidence level of the condition.
| Confidence | value |
| low | 1L |
| med | 2L |
| high | 3L |
library(terra) qa <- rast(ncol = 100, nrow=100, val = sample(1:2^14, 10000)) ## QA classes qacs <- classifyQA(img = qa) ## Confidence levels qacs_conf <- classifyQA(img = qa, confLayers = TRUE)library(terra) qa <- rast(ncol = 100, nrow=100, val = sample(1:2^14, 10000)) ## QA classes qacs <- classifyQA(img = qa) ## Confidence levels qacs_conf <- classifyQA(img = qa, confLayers = TRUE)
Developed for use with Landsat data cloudMask relies on the distinctive difference between the blue (or any other short-wave band)
and thermal band for semi-automated creation of a cloud mask. Since it relies on thermal information it doesn't work well for sensors without
thermal bands.
cloudMask( x, threshold = 0.2, blue = "B1_sre", tir = "B6_sre", buffer = NULL, plot = FALSE, verbose )cloudMask( x, threshold = 0.2, blue = "B1_sre", tir = "B6_sre", buffer = NULL, plot = FALSE, verbose )
x |
SpatRaster with reflectance and brightness temperature OR the mask of a previous run of |
threshold |
Numeric. cloud detection threshold. If not provided it will be guessed. Everything *below* this threshold will be considered a cloud pixel (unless it is removed by filtering afterwards). |
blue |
Character or integer. Bandname or number for the blue band |
tir |
Character or integer. Bandname or number for the thermal band |
buffer |
Integer. Number of pixels to use as a buffer that will be added to the identified cloud centers. |
plot |
Logical. Plots of the cloud mask for all sub-steps (sanitizing etc.) Helpful to find proper parametrization. |
verbose |
Logical. Print messages or suppress. |
Returns a SpatRaster with two layers: CMASK contains the binary cloud mask (1 = cloud, NA = not-cloud) and NDTCI contains the cloud index.
Typically clouds are cold in the thermal region and have high reflectance in short wavelengths (blue). By calculating a normalized difference index between the two bands and thresholding a rough cloud mask can be obtained. Before calculating the spectral cloud index (let's call it Normalized Difference Thermal Cloud Index (NDTCI)) the thermal band will be matched to the same value range as the blue band. Therefore, it doesn't matter whether you provide DN, radiance or brightness temperature.
This approach to cloud masking is very simplistic. And aims only at rough removal of potentially clouded areas. Nevertheless, it is able to provide satisfactory results. More sophisticated approaches, including cloud cast shadow detection can be found elsewhere, e.g. fmask.
It can make sense to find a suitable threshold on a cropped version of the scene. Also make sure you make use of the returnDiffLayer argument to save yourself one processing step.
Buffering should be seen as final polishing, i.e. as long as the pure cloud centers are not detected properly, you might want to turn it off. since it takes some time to calculate.
Once your mask detects obvious cloud pixels properly re-enable buffering for fine tuning if desired. Finally, once a suitable threshold is established re-run cloudMask on the whole scene with this threshold and go get a coffee.
library(ggplot2) ## Import Landsat example subset ## We have two tiny clouds in the east ggRGB(lsat, stretch = "lin") ## Calculate cloud index cldmsk <- cloudMask(lsat, blue = 1, tir = 6) ggR(cldmsk, 2, geom_raster = TRUE) ## Define threshold (re-use the previously calculated index) ## Everything above the threshold is masked ## In addition we apply a region-growing around the core cloud pixels cldmsk_final <- cloudMask(cldmsk, threshold = 0.1, buffer = 5) ## Plot cloudmask ggRGB(lsat, stretch = "lin") + ggR(cldmsk_final[[1]], ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("red"), na.value = NA) #' ## Estimate cloud shadow displacement ## Interactively (click on cloud pixels and the corresponding shadow pixels) ## Not run: shadow <- cloudShadowMask(lsat, cldmsk_final, nc = 2) ## Non-interactively. Pre-defined shadow displacement estimate (shiftEstimate) shadow <- cloudShadowMask(lsat, cldmsk_final, shiftEstimate = c(-16,-6)) ## Plot csmask <- terra::merge(cldmsk_final[[1]], shadow) ggRGB(lsat, stretch = "lin") + ggR(csmask, ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("blue", "yellow"), labels = c("shadow", "cloud"), na.value = NA)library(ggplot2) ## Import Landsat example subset ## We have two tiny clouds in the east ggRGB(lsat, stretch = "lin") ## Calculate cloud index cldmsk <- cloudMask(lsat, blue = 1, tir = 6) ggR(cldmsk, 2, geom_raster = TRUE) ## Define threshold (re-use the previously calculated index) ## Everything above the threshold is masked ## In addition we apply a region-growing around the core cloud pixels cldmsk_final <- cloudMask(cldmsk, threshold = 0.1, buffer = 5) ## Plot cloudmask ggRGB(lsat, stretch = "lin") + ggR(cldmsk_final[[1]], ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("red"), na.value = NA) #' ## Estimate cloud shadow displacement ## Interactively (click on cloud pixels and the corresponding shadow pixels) ## Not run: shadow <- cloudShadowMask(lsat, cldmsk_final, nc = 2) ## Non-interactively. Pre-defined shadow displacement estimate (shiftEstimate) shadow <- cloudShadowMask(lsat, cldmsk_final, shiftEstimate = c(-16,-6)) ## Plot csmask <- terra::merge(cldmsk_final[[1]], shadow) ggRGB(lsat, stretch = "lin") + ggR(csmask, ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("blue", "yellow"), labels = c("shadow", "cloud"), na.value = NA)
Intended for interactive use, cloudShadowMask will ask the user to select a few
corresponding cloud/cloudShadow pixels which will be used to estimate coordinates
for a linear cloudmask shift.
cloudShadowMask( img, cm, nc = 5, shiftEstimate = NULL, preciseShift = NULL, quantile = 0.2, returnShift = FALSE )cloudShadowMask( img, cm, nc = 5, shiftEstimate = NULL, preciseShift = NULL, quantile = 0.2, returnShift = FALSE )
img |
SpatRaster containing the scene |
cm |
SpatRaster. Cloud mask (typically the result of |
nc |
Integer. Number of control points. A few points (default) are fine because the final shift is estimated by coregisterImages. |
shiftEstimate |
NULL or numeric vector of length two (x,y). Estimated displacement of shadows in map units. If |
preciseShift |
NULL or numeric vector of length two (x,y). Use this if cloud/cloud-shadow displacement is already known, e.g. from a previous run of |
quantile |
Numeric (between 0 and 1). Quantile threshold used for image co-registration. By default the 20% quantile of the total intensity (sum) of the image is used as potential shadow mask. |
returnShift |
Logical. Return a numeric vector containing the shift parameters. Useful if you estimate parameters on a subset of the image. |
This is a very simplistic approach to cloud shadow masking (simple shift of the cloud mask). It is not image based and accuracy will suffer from clouds at different altitudes. However, just as cloudMask this is a quick and easy to use tool for Landsat data if you're just working on a few scenes and don't have fMask or CDR data at hand. Although for some test scenes it does perform surprisingly well.
Returns a RasterLayer with the cloud shadow mask (0 = shadow, NA = not-shadow).
library(ggplot2) ## Import Landsat example subset ## We have two tiny clouds in the east ggRGB(lsat, stretch = "lin") ## Calculate cloud index cldmsk <- cloudMask(lsat, blue = 1, tir = 6) ggR(cldmsk, 2, geom_raster = TRUE) ## Define threshold (re-use the previously calculated index) ## Everything above the threshold is masked ## In addition we apply a region-growing around the core cloud pixels cldmsk_final <- cloudMask(cldmsk, threshold = 0.1, buffer = 5) ## Plot cloudmask ggRGB(lsat, stretch = "lin") + ggR(cldmsk_final[[1]], ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("red"), na.value = NA) #' ## Estimate cloud shadow displacement ## Interactively (click on cloud pixels and the corresponding shadow pixels) ## Not run: shadow <- cloudShadowMask(lsat, cldmsk_final, nc = 2) ## Non-interactively. Pre-defined shadow displacement estimate (shiftEstimate) shadow <- cloudShadowMask(lsat, cldmsk_final, shiftEstimate = c(-16,-6)) ## Plot csmask <- terra::merge(cldmsk_final[[1]], shadow) ggRGB(lsat, stretch = "lin") + ggR(csmask, ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("blue", "yellow"), labels = c("shadow", "cloud"), na.value = NA)library(ggplot2) ## Import Landsat example subset ## We have two tiny clouds in the east ggRGB(lsat, stretch = "lin") ## Calculate cloud index cldmsk <- cloudMask(lsat, blue = 1, tir = 6) ggR(cldmsk, 2, geom_raster = TRUE) ## Define threshold (re-use the previously calculated index) ## Everything above the threshold is masked ## In addition we apply a region-growing around the core cloud pixels cldmsk_final <- cloudMask(cldmsk, threshold = 0.1, buffer = 5) ## Plot cloudmask ggRGB(lsat, stretch = "lin") + ggR(cldmsk_final[[1]], ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("red"), na.value = NA) #' ## Estimate cloud shadow displacement ## Interactively (click on cloud pixels and the corresponding shadow pixels) ## Not run: shadow <- cloudShadowMask(lsat, cldmsk_final, nc = 2) ## Non-interactively. Pre-defined shadow displacement estimate (shiftEstimate) shadow <- cloudShadowMask(lsat, cldmsk_final, shiftEstimate = c(-16,-6)) ## Plot csmask <- terra::merge(cldmsk_final[[1]], shadow) ggRGB(lsat, stretch = "lin") + ggR(csmask, ggLayer = TRUE, forceCat = TRUE, geom_raster = TRUE) + scale_fill_manual(values = c("blue", "yellow"), labels = c("shadow", "cloud"), na.value = NA)
Shifts an image to match a reference image. Matching is based on maximum mutual information.
coregisterImages( img, ref, shift = 3, shiftInc = 1, nSamples = 100, reportStats = FALSE, verbose, nBins = 100, master = deprecated(), slave = deprecated(), ... )coregisterImages( img, ref, shift = 3, shiftInc = 1, nSamples = 100, reportStats = FALSE, verbose, nBins = 100, master = deprecated(), slave = deprecated(), ... )
img |
SpatRaster. Image to shift to match reference image. |
ref |
SpatRaster. Reference image. |
shift |
Numeric or matrix. If numeric, then shift is the maximal absolute radius (in pixels of |
shiftInc |
Numeric. Shift increment (in pixels, but not restricted to integer). Ignored if |
nSamples |
Integer. Number of samples to calculate mutual information. |
reportStats |
Logical. If |
verbose |
Logical. Print status messages. Overrides global RStoolbox.verbose option. |
nBins |
Integer. Number of bins to calculate joint histogram. |
master |
DEPRECATED! Argument was renamed. Please use |
slave |
DEPRECATED! Argument was renamed. Please use |
... |
further arguments passed to |
Currently only a simple linear x - y shift is considered and tested. No higher order shifts (e.g. rotation, non-linear transformation) are performed. This means that your imagery should already be properly geometrically corrected.
Mutual information is a similarity metric originating from information theory. Roughly speaking, the higher the mutual information of two data-sets, the higher is their shared information content, i.e. their similarity. When two images are exactly co-registered their mutual information is maximal. By trying different image shifts, we aim to find the best overlap which maximises the mutual information.
reportStats=FALSE returns a SpatRaster (x-y shifted image).
reportStats=TRUE returns a list containing a data.frame with mutual information per shift ($MI), the shift of maximum MI ($bestShift),
the joint histograms per shift in a list ($jointHist) and the shifted image ($coregImg).
library(terra) library(ggplot2) library(reshape2) reference <- rlogo ## Shift reference 2 pixels to the right and 3 up missreg <- shift(reference, 2, 3) ## Compare shift p <- ggR(reference, sat = 1, alpha = .5) p + ggR(missreg, sat = 1, hue = .5, alpha = 0.5, ggLayer=TRUE) ## Coregister images (and report statistics) coreg <- coregisterImages(missreg, ref = reference, nSamples = 500, reportStats = TRUE) ## Plot mutual information per shift ggplot(coreg$MI) + geom_raster(aes(x,y,fill=mi)) ## Plot joint histograms per shift (x/y shift in facet labels) df <- melt(coreg$jointHist) df$L1 <- factor(df$L1, levels = names(coreg$jointHist)) df[df$value == 0, "value"] <- NA ## don't display p = 0 ggplot(df) + geom_raster(aes(x = Var2, y = Var1,fill=value)) + facet_wrap(~L1) + scale_fill_gradientn(name = "p", colours = heat.colors(10), na.value = NA) ## Compare correction ggR(reference, sat = 1, alpha = .5) + ggR(coreg$coregImg, sat = 1, hue = .5, alpha = 0.5, ggLayer=TRUE)library(terra) library(ggplot2) library(reshape2) reference <- rlogo ## Shift reference 2 pixels to the right and 3 up missreg <- shift(reference, 2, 3) ## Compare shift p <- ggR(reference, sat = 1, alpha = .5) p + ggR(missreg, sat = 1, hue = .5, alpha = 0.5, ggLayer=TRUE) ## Coregister images (and report statistics) coreg <- coregisterImages(missreg, ref = reference, nSamples = 500, reportStats = TRUE) ## Plot mutual information per shift ggplot(coreg$MI) + geom_raster(aes(x,y,fill=mi)) ## Plot joint histograms per shift (x/y shift in facet labels) df <- melt(coreg$jointHist) df$L1 <- factor(df$L1, levels = names(coreg$jointHist)) df[df$value == 0, "value"] <- NA ## don't display p = 0 ggplot(df) + geom_raster(aes(x = Var2, y = Var1,fill=value)) + facet_wrap(~L1) + scale_fill_gradientn(name = "p", colours = heat.colors(10), na.value = NA) ## Compare correction ggR(reference, sat = 1, alpha = .5) + ggR(coreg$coregImg, sat = 1, hue = .5, alpha = 0.5, ggLayer=TRUE)
Intended for use with Landsat 16-bit QA bands. Decodes pixel quality flags from integer to bit-words.
decodeQA(x)decodeQA(x)
x |
Integer (16bit) |
Returns the decoded QA values from an integer
decodeQA(53248)decodeQA(53248)
Intended for use with Landsat 16-bit QA bands. Converts pixel quality flags from human readable to integer, which can then be used to
subset a QA image. Please be aware of the default settings which differ for different parameters.
Depending on, which sensor and legacy is selected, some quality parameters are not used, since the sequences of available bitwise quality designations differ per sensor and collection.
encodeQA( fill = "no", terrainOcclusion = "no", radSaturation = "na", cloudMask = "all", cloud = "all", cloudShadow = "all", snow = "all", cirrus = "all", droppedPixel = "no", water = "all", droppedFrame = "no", sensor = "OLI", legacy = "collection1" )encodeQA( fill = "no", terrainOcclusion = "no", radSaturation = "na", cloudMask = "all", cloud = "all", cloudShadow = "all", snow = "all", cirrus = "all", droppedPixel = "no", water = "all", droppedFrame = "no", sensor = "OLI", legacy = "collection1" )
fill |
Designated fill. Options: |
terrainOcclusion |
Terrain induced occlusion. Options: |
radSaturation |
Number of bands that contain radiometric saturation. Options: |
cloudMask |
Cloud mask. Options: |
cloud |
Cloud confidence. Options: |
cloudShadow |
Cloud shadow confidence. Options: |
snow |
Snow / ice confidence. Options: |
cirrus |
Cirrus confidence. Options: |
droppedPixel |
Dropped pixel. Options: |
water |
Water confidence. Options: |
droppedFrame |
Dropped frame. Options: |
sensor |
Sensor to encode. Options: |
legacy |
Encoding systematic Options: |
Returns the Integer value for the QA values
Only currently populated bits are available as arguments.
https://www.usgs.gov/landsat-missions/landsat-collection-1-level-1-quality-assessment-band for Collection 1 quality designations (legacy = "collection1")
encodeQA(snow = "low", cirrus = c("med", "high"), cloud = "high")encodeQA(snow = "low", cirrus = c("med", "high"), cloud = "high")
estimates the digital number (DN) pixel value of *dark* objects for the visible wavelength range.
estimateHaze( x, hazeBands, darkProp = 0.01, maxSlope = TRUE, plot = FALSE, returnTables = FALSE )estimateHaze( x, hazeBands, darkProp = 0.01, maxSlope = TRUE, plot = FALSE, returnTables = FALSE )
x |
SpatRaster or a previous result from |
hazeBands |
Integer or Character. Band number or bandname from which to estimate atmospheric haze (optional if x contains only one layer) |
darkProp |
Numeric. Proportion of pixels estimated to be dark. |
maxSlope |
Logical. Use |
plot |
Logical. Option to display histograms and haze values |
returnTables |
Logical. Option to return the frequency table per layer. Only takes effect if x is a SpatRaster. If x is a result of estimateHaze tables will always be returned. |
It is assumed that any radiation originating from *dark* pixels is due to atmospheric haze and
not the reflectance of the surface itself (the surface is dark, i.e. it has a reflectance close to zero).
Hence, the haze values are estimates of path radiance, which can be subtracted in a dark object subtraction (DOS) procedure (see radCor)
Atmospheric haze affects almost exclusively the visible wavelength range. Therefore, typically, you'd only want to estimate haze in blue, green and red bands, occasionally also in the nir band.
If returnTables is FALSE (default). Then a vector of length(hazeBands) containing the estimated haze DNs will be returned. If returnTables is TRUE a list with two components will be returned. The list element 'SHV' contains the haze values, while 'table' contains another list with the sampled frequency tables. The latter can be re-used to try different darkProp thresholds without having to sample the raster again.
## Estimate haze for blue, green and red band haze <- estimateHaze(lsat, hazeBands = 1:3, plot = FALSE) haze ## Find threshold interactively #### Return the frequency tables for re-use #### avoids having to sample the Raster again and again haze <- estimateHaze(lsat, hazeBands = 1:3, returnTables = TRUE) ## Use frequency table instead of lsat and fiddle with haze <- estimateHaze(haze, hazeBands = 1:3, darkProp = .1, plot = FALSE) haze$SHV## Estimate haze for blue, green and red band haze <- estimateHaze(lsat, hazeBands = 1:3, plot = FALSE) haze ## Find threshold interactively #### Return the frequency tables for re-use #### avoids having to sample the Raster again and again haze <- estimateHaze(lsat, hazeBands = 1:3, returnTables = TRUE) ## Use frequency table instead of lsat and fiddle with haze <- estimateHaze(haze, hazeBands = 1:3, darkProp = .1, plot = FALSE) haze$SHV
fCover takes a classified high resolution image, e.g. vegetation and non-vegetation based on Landsat and calculates cover fractions for pixels of a coarser resolution, e.g. MODIS.
fCover( classImage, predImage, nSamples = 1000, classes = 1, maxNA = 0, clamp = TRUE, model = "rf", tuneLength = 3, trControl = trainControl(method = "cv"), method = deprecated(), filename = NULL, overwrite = FALSE, verbose, ... )fCover( classImage, predImage, nSamples = 1000, classes = 1, maxNA = 0, clamp = TRUE, model = "rf", tuneLength = 3, trControl = trainControl(method = "cv"), method = deprecated(), filename = NULL, overwrite = FALSE, verbose, ... )
classImage |
high resolution SpatRaster containing a landcover classification, e.g. as obtained by superClass. |
predImage |
coarse resolution SpatRaster for which fractional cover will be estimated. |
nSamples |
Integer. Number of pixels to sample from predImage to train the regression model |
classes |
Integer. Classes for which fractional cover should be estimated (one or more). |
maxNA |
Numeric. Maximal proportion of NAs allowed in training pixels. |
clamp |
Logical. Enforce results to stay within [0,1] interval. Values <0 are reset to 0 and values >1 to 1. |
model |
Character. Which model to fit for image regression. See train for options. Defaults to randomForest ('rf') |
tuneLength |
Integer. Number of levels for each tuning parameters that should be generated by train. See Details and |
trControl |
|
method |
DEPREACTED! in favor of |
filename |
Character. Filename of the output raster file (optional). |
overwrite |
Logical. if |
verbose |
Logical. Print progress information. |
... |
further arguments to be passed to |
fCover gets the pixel values in a high resolution classified image that correspond to randomly selected moderate resolution pixels and then calculates the percent of the classified image pixels that represent your cover type of interest. In other words, if your high resolution image has a pixel size of 1m and your moderate resolution image has a pixel size of 30m the sampling process would take a block of 900 of the 1m resolution pixels that correspond to a single 30m pixel and calculate the percentage of the 1m pixels that are forest. For example, if there were 600 forest pixels and 300 non-forest pixels the value given for the output pixel would be 0.67 since 67
fCover relies on the train() function from the caret package which provides access to a huge number of classifiers. Please see the available options at train. The default classifier (randomForest) we chose has been shown to provide very good results in image regression and hence it is recomended you start with this one. If you choose a different classifier, make sure it can run in regression mode.
Many models require tuning of certain parameters. Again, this is handled by train from the caret package. With the tuneLength argument you can specify how many different values of each tuning parameter should be tested. The Random Forest algorithm for example can be tuned by varying the mtry parameter. Hence, by specifying tuneLength = 10, ten different levels of mtry will be tested in a cross-validation scheme and the best-performing value will be chosen for the final model.
If the total no-data values for a block of high resolution pixels is greater than maxNA then it will not be included in the training data set since there is too much missing data to provide a reliable cover percentage. If the no-data proporton is less then maxNA the no-data pixels are removed from the total number of pixels when calculating the percent cover.
Returns a list with two elements: models contains the fitted models evaluated in tenfold cross-validation (caret train objects); fCover contains a SpatRaster with a fractional cover layer for each requested class.
library(terra) library(caret) ## Create fake input images agg.level <- 9 modis <- terra::aggregate(rlogo, agg.level) ## Perform an exemplary classification lc <- unsuperClass(rlogo, nClass=2) ## Calculate the true cover, which is of course only possible in this example, ## because the fake corse resolution imagery is exactly res(rlogo)*9 trueCover <- terra::aggregate(lc$map, agg.level, fun = function(x, ...){sum(x == 1, ...)/sum(!is.na(x))}) ## Run with randomForest and support vector machine (radial basis kernel) ## Of course the SVM is handicapped in this example, ## due to poor tuning (tuneLength) par(mfrow=c(2,3)) for(model in c("rf", "svmRadial")){ fc <- fCover( classImage = lc$map , predImage = modis, classes=1, trControl = trainControl(method = "cv", number = 3), model=model, nSample = 50, tuneLength=2 ) ## How close is it to the truth? compare.rf <- trueCover - fc$map plot(fc$map, main = paste("Fractional Cover: Class 1\nModel:", model)) plot(compare.rf, main = "Diffence\n true vs. predicted") plot(trueCover[],fc$map[], xlim = c(0,1), ylim =c(0,1), xlab = "True Cover", ylab = "Predicted Cover" ) abline(coef=c(0,1)) rmse <- sqrt(global(compare.rf^2, "sum", na.rm = TRUE))/ncell(compare.rf) r2 <- cor(trueCover[], fc$map[], "complete.obs") text(0.9,0.1, adj=1, labels = paste(c("RMSE:","\nR2:"), round(unlist(c(rmse, r2)),3), collapse="")) } ## Reset par par(mfrow=c(1,1))library(terra) library(caret) ## Create fake input images agg.level <- 9 modis <- terra::aggregate(rlogo, agg.level) ## Perform an exemplary classification lc <- unsuperClass(rlogo, nClass=2) ## Calculate the true cover, which is of course only possible in this example, ## because the fake corse resolution imagery is exactly res(rlogo)*9 trueCover <- terra::aggregate(lc$map, agg.level, fun = function(x, ...){sum(x == 1, ...)/sum(!is.na(x))}) ## Run with randomForest and support vector machine (radial basis kernel) ## Of course the SVM is handicapped in this example, ## due to poor tuning (tuneLength) par(mfrow=c(2,3)) for(model in c("rf", "svmRadial")){ fc <- fCover( classImage = lc$map , predImage = modis, classes=1, trControl = trainControl(method = "cv", number = 3), model=model, nSample = 50, tuneLength=2 ) ## How close is it to the truth? compare.rf <- trueCover - fc$map plot(fc$map, main = paste("Fractional Cover: Class 1\nModel:", model)) plot(compare.rf, main = "Diffence\n true vs. predicted") plot(trueCover[],fc$map[], xlim = c(0,1), ylim =c(0,1), xlab = "True Cover", ylab = "Predicted Cover" ) abline(coef=c(0,1)) rmse <- sqrt(global(compare.rf^2, "sum", na.rm = TRUE))/ncell(compare.rf) r2 <- cor(trueCover[], fc$map[], "complete.obs") text(0.9,0.1, adj=1, labels = paste(c("RMSE:","\nR2:"), round(unlist(c(rmse, r2)),3), collapse="")) } ## Reset par par(mfrow=c(1,1))
Fortify method for classes from the terra package.
fortifySpatRaster(x, maxpixels = 50000)fortifySpatRaster(x, maxpixels = 50000)
x |
|
maxpixels |
Integer. Maximum number of pixels to sample |
Returns a data.frame with coordinates (x,y) and corresponding raster values.
r_df <- fortifySpatRaster(rlogo) head(r_df)r_df <- fortifySpatRaster(rlogo) head(r_df)
This is an accessor function to quickly access information stored in ImageMetaData, e.g. scale factor per band. Intended for use with imagery which was imported using stackMeta. Will return parameters using the actual band order in img.
getMeta(img, metaData, what)getMeta(img, metaData, what)
img |
SpatRaster or character vector with band names. |
metaData |
ImageMetaData or path to meta data file. |
what |
Character. Parameter to extract. Either data descriptors, or conversion parameters (see Details for options) |
Possible metadata parameters (what argument):
Data descriptors
| 'FILES' | |
| 'QUANTITY' | |
| 'CATEGORY' | |
| 'NA_VALUE' | |
| 'SATURATE_VALUE' | |
| 'SCALE_FACTOR' | |
| 'DATA_TYPE' | |
| 'SPATIAL_RESOLUTION' | |
Conversion parameters
| 'CALRAD' | Conversion parameters from DN to radiance |
| 'CALBT' | Conversion parameters from radiance to brightness temperature |
| 'CALREF' | Conversion parameters from DN to reflectance (Landsat 8 only) |
If what is one of c('CALRAD', 'CALBT', 'CALREF') a data.frame is returned with bands in rows (order corresponding to img band order).
Otherwise a named numeric vector with the corresponding parameter is returned (layernames as names).
## Import example data mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") meta <- readMeta(mtlFile) lsat_t <- stackMeta(mtlFile) ## Get integer scale factors getMeta(lsat_t, metaData = meta, what = "SCALE_FACTOR") ## Conversion factors for brightness temperature getMeta("B6_dn", metaData = meta, what = "CALBT") ## Conversion factors to top-of-atmosphere radiance ## Band order not corresponding to metaData order getMeta(lsat_t[[5:1]], metaData = meta, what = "CALRAD") ## Get integer scale factors getMeta(lsat_t, metaData = meta, what = "SCALE_FACTOR") ## Get file basenames getMeta(lsat_t, metaData = meta, what = "FILES")## Import example data mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") meta <- readMeta(mtlFile) lsat_t <- stackMeta(mtlFile) ## Get integer scale factors getMeta(lsat_t, metaData = meta, what = "SCALE_FACTOR") ## Conversion factors for brightness temperature getMeta("B6_dn", metaData = meta, what = "CALBT") ## Conversion factors to top-of-atmosphere radiance ## Band order not corresponding to metaData order getMeta(lsat_t[[5:1]], metaData = meta, what = "CALRAD") ## Get integer scale factors getMeta(lsat_t, metaData = meta, what = "SCALE_FACTOR") ## Get file basenames getMeta(lsat_t, metaData = meta, what = "FILES")
Extract validation results from superClass objects
getValidation(x, from = "testset", metrics = "overall")getValidation(x, from = "testset", metrics = "overall")
x |
superClass object or caret::confusionMatrix |
from |
Character. 'testset' extracts the results from independent validation with testset. 'cv' extracts cross-validation results. |
metrics |
Character. Only relevant in classification mode (ignored for regression models). Select 'overall' for overall accuracy metrics, 'classwise' for classwise metrics, 'confmat' for the confusion matrix itself and 'caret' to return the whole caret::confusionMatrix object. |
Returns a data.frame with validation results. If metrics = 'confmat' or 'caret' will return a table or the full caret::confusionMatrix object, respectively.
library(pls) ## Fit classifier (splitting training into 70\% training data, 30\% validation data) train <- readRDS(system.file("external/trainingPoints_rlogo.rds", package="RStoolbox")) SC <- superClass(rlogo, trainData = train, responseCol = "class", model="pls", trainPartition = 0.7) ## Independent testset-validation getValidation(SC) getValidation(SC, metrics = "classwise") ## Cross-validation based getValidation(SC, from = "cv")library(pls) ## Fit classifier (splitting training into 70\% training data, 30\% validation data) train <- readRDS(system.file("external/trainingPoints_rlogo.rds", package="RStoolbox")) SC <- superClass(rlogo, trainData = train, responseCol = "class", model="pls", trainPartition = 0.7) ## Independent testset-validation getValidation(SC) getValidation(SC, metrics = "classwise") ## Cross-validation based getValidation(SC, from = "cv")
Plot single layer imagery in grey-scale. Can be used with a SpatRaster.
ggR( img, layer = 1, maxpixels = 5e+05, alpha = 1, hue = 1, sat = 0, stretch = "none", quantiles = c(0.02, 0.98), ext = NULL, coord_equal = TRUE, ggLayer = FALSE, ggObj = TRUE, geom_raster = FALSE, forceCat = FALSE )ggR( img, layer = 1, maxpixels = 5e+05, alpha = 1, hue = 1, sat = 0, stretch = "none", quantiles = c(0.02, 0.98), ext = NULL, coord_equal = TRUE, ggLayer = FALSE, ggObj = TRUE, geom_raster = FALSE, forceCat = FALSE )
img |
SpatRaster |
layer |
Character or numeric. Layername or number. Can be more than one layer, in which case each layer is plotted in a subplot. |
maxpixels |
Integer. Maximal number of pixels to sample. |
alpha |
Numeric. Transparency (0-1). |
hue |
Numeric. Hue value for color calculation [0,1] (see |
sat |
Numeric. Saturation value for color calculation [0,1] (see |
stretch |
Character. Either 'none', 'lin', 'hist', 'sqrt' or 'log' for no stretch, linear, histogram, square-root or logarithmic stretch. |
quantiles |
Numeric vector with two elements. Min and max quantiles to stretch to. Defaults to 2% stretch, i.e. c(0.02,0.98). |
ext |
Extent object to crop the image |
coord_equal |
Logical. Force addition of coord_equal, i.e. aspect ratio of 1:1. Typically useful for remote sensing data (depending on your projection), hence it defaults to TRUE.
Note however, that this does not apply if ( |
ggLayer |
Logical. Return only a ggplot layer which must be added to an existing ggplot. If |
ggObj |
Logical. Return a stand-alone ggplot object (TRUE) or just the data.frame with values and colors |
geom_raster |
Logical. If |
forceCat |
Logical. If |
When img contains factor values and annotation=TRUE, the raster values will automatically be converted
to numeric in order to proceed with the brightness calculation.
The geom_raster argument switches from the default use of annotation_raster to geom_raster. The difference between the two is that geom_raster performs
a meaningful mapping from pixel values to fill colour, while annotation_raster is simply adding a picture to your plot. In practice this means that whenever you
need a legend for your raster you should use geom_raster = TRUE. This also allows you to specify and modify the fill scale manually.
The advantage of using annotation_raster (geom_raster = TRUE) is that you can still use the scale_fill* for another variable. For example you could add polygons and
map a value to their fill colour. For more details on the theory behind aestetic mapping have a look at the ggplot2 manuals.
ggObj = TRUE: |
ggplot2 plot |
ggLayer = TRUE: |
ggplot2 layer to be combined with an existing ggplot2 |
ggObj = FALSE: |
data.frame in long format suitable for plotting with ggplot2, includes the pixel values and the calculated colors |
library(ggplot2) library(terra) ## Simple grey scale annotation ggR(rlogo) ## With linear stretch contrast enhancement ggR(rlogo, stretch = "lin", quantiles = c(0.1,0.9)) ## ggplot with geom_raster instead of annotation_raster ## and default scale_fill* ggR(rlogo, geom_raster = TRUE) ## with different scale ggR(rlogo, geom_raster = TRUE) + scale_fill_gradientn(name = "mojo", colours = rainbow(10)) + ggtitle("**Funkadelic**") ## Plot multiple layers ggR(lsat, 1:6, geom_raster=TRUE, stretch = "lin") + scale_fill_gradientn(colors=grey.colors(100), guide = "none") + theme(axis.text = element_text(size=5), axis.text.y = element_text(angle=90), axis.title=element_blank()) ## Don't plot, just return a data.frame df <- ggR(rlogo, ggObj = FALSE) head(df, n = 3) ## Layermode (ggLayer=TRUE) data <- data.frame(x = c(0, 0:100,100), y = c(0,sin(seq(0,2*pi,pi/50))*10+20, 0)) ggplot(data, aes(x, y)) + ggR(rlogo, geom_raster= FALSE, ggLayer = TRUE) + geom_polygon(aes(x, y), fill = "blue", alpha = 0.4) + coord_equal(ylim=c(0,75)) ## Categorical data ## In this case you probably want to use geom_raster=TRUE ## in order to perform aestetic mapping (i.e. a meaningful legend) rc <- rlogo rc[] <- cut(rlogo[[1]][], seq(0,300, 50)) ggR(rc, geom_raster = TRUE) ## Legend cusomization etc. ... ggR(rc, geom_raster = TRUE) + scale_fill_continuous(labels=paste("Class", 1:6)) ## Creating a nicely looking DEM with hillshade background terr <- terrain(srtm, c("slope", "aspect")) hill <- shade(terr[["slope"]], terr[["aspect"]]) ggR(hill) ggR(hill) + ggR(srtm, geom_raster = TRUE, ggLayer = TRUE, alpha = 0.3) + scale_fill_gradientn(colours = terrain.colors(100), name = "elevation")library(ggplot2) library(terra) ## Simple grey scale annotation ggR(rlogo) ## With linear stretch contrast enhancement ggR(rlogo, stretch = "lin", quantiles = c(0.1,0.9)) ## ggplot with geom_raster instead of annotation_raster ## and default scale_fill* ggR(rlogo, geom_raster = TRUE) ## with different scale ggR(rlogo, geom_raster = TRUE) + scale_fill_gradientn(name = "mojo", colours = rainbow(10)) + ggtitle("**Funkadelic**") ## Plot multiple layers ggR(lsat, 1:6, geom_raster=TRUE, stretch = "lin") + scale_fill_gradientn(colors=grey.colors(100), guide = "none") + theme(axis.text = element_text(size=5), axis.text.y = element_text(angle=90), axis.title=element_blank()) ## Don't plot, just return a data.frame df <- ggR(rlogo, ggObj = FALSE) head(df, n = 3) ## Layermode (ggLayer=TRUE) data <- data.frame(x = c(0, 0:100,100), y = c(0,sin(seq(0,2*pi,pi/50))*10+20, 0)) ggplot(data, aes(x, y)) + ggR(rlogo, geom_raster= FALSE, ggLayer = TRUE) + geom_polygon(aes(x, y), fill = "blue", alpha = 0.4) + coord_equal(ylim=c(0,75)) ## Categorical data ## In this case you probably want to use geom_raster=TRUE ## in order to perform aestetic mapping (i.e. a meaningful legend) rc <- rlogo rc[] <- cut(rlogo[[1]][], seq(0,300, 50)) ggR(rc, geom_raster = TRUE) ## Legend cusomization etc. ... ggR(rc, geom_raster = TRUE) + scale_fill_continuous(labels=paste("Class", 1:6)) ## Creating a nicely looking DEM with hillshade background terr <- terrain(srtm, c("slope", "aspect")) hill <- shade(terr[["slope"]], terr[["aspect"]]) ggR(hill) ggR(hill) + ggR(srtm, geom_raster = TRUE, ggLayer = TRUE, alpha = 0.3) + scale_fill_gradientn(colours = terrain.colors(100), name = "elevation")
Calculates RGB color composite raster for plotting with ggplot2. Optional values for clipping and and stretching can be used to enhance the imagery.
ggRGB( img, r = 3, g = 2, b = 1, scale, maxpixels = 5e+05, stretch = "none", ext = NULL, limits = NULL, clipValues = "limits", quantiles = c(0.02, 0.98), ggObj = TRUE, ggLayer = FALSE, alpha = 1, coord_equal = TRUE, geom_raster = FALSE, nullValue = 0 )ggRGB( img, r = 3, g = 2, b = 1, scale, maxpixels = 5e+05, stretch = "none", ext = NULL, limits = NULL, clipValues = "limits", quantiles = c(0.02, 0.98), ggObj = TRUE, ggLayer = FALSE, alpha = 1, coord_equal = TRUE, geom_raster = FALSE, nullValue = 0 )
img |
SpatRaster |
r |
Integer or character. Red layer in x. Can be set to |
g |
Integer or character. Green layer in x. Can be set to |
b |
Integer or character. Blue layer in x. Can be set to |
scale |
Numeric. Maximum possible pixel value (optional). Defaults to 255 or to the maximum value of x if that is larger than 255 |
maxpixels |
Integer. Maximal number of pixels used for plotting. |
stretch |
Character. Either 'none', 'lin', 'hist', 'sqrt' or 'log' for no stretch, linear, histogram, square-root or logarithmic stretch. |
ext |
Extent or SpatExtent object to crop the image |
limits |
Vector or matrix. Can be used to reduce the range of values. Either a vector of two values for all bands (c(min, max)) or a 3x2 matrix with min and max values (columns) for each layer (rows). |
clipValues |
Matrix, numeric vector, string or NA. Values to reset out of range (out of |
quantiles |
Numeric vector with two elements. Min and max quantiles to stretch. Defaults to 2% stretch, i.e. c(0.02,0.98). |
ggObj |
Logical. If |
ggLayer |
Logical. If |
alpha |
Numeric. Transparency (0-1). |
coord_equal |
Logical. Force addition of coord_equal, i.e. aspect ratio of 1:1. Typically useful for remote sensing data (depending on your projection), hence it defaults to TRUE.
Note howver, that this does not apply if ( |
geom_raster |
Logical. If |
nullValue |
Numeric. Intensity value used for NULL layers in color compositing. E.g. set g=NULL and fix green value at 0.5 (defaults to 0). |
Functionality is based on plotRGB from the raster package.
ggObj = TRUE: |
ggplot2 plot |
ggLayer = TRUE: |
ggplot2 layer to be combined with an existing ggplot2 |
ggObj = FALSE: |
data.frame in long format suitable for plotting with ggplot2, includes the pixel values and the calculated colors |
library(ggplot2) ggRGB(rlogo, r=1, g=2, b=3) ## Define minMax ranges ggRGB(rlogo, r=1,g=2, b=3, limits = matrix(c(100,150,10,200,50,255), ncol = 2, by = TRUE)) ## Perform stong linear contrast stretch ggRGB(rlogo, r = 1, g = 2, b = 3,stretch = "lin", quantiles = c(0.2, 0.8)) ## Use only two layers for color calculation ggRGB(rlogo, r = 1, g = 2, b = NULL) ## Return only data.frame df <- ggRGB(rlogo, ggObj = FALSE) head(df) ## Use in layer-mode, e.g. to add to another plot wave <- data.frame(x = c(0, 0:100,100), y = c(0,sin(seq(0,2*pi,pi/50))*10+20, 0)) p <- ggplot(wave, aes(x, y)) p + ggRGB(rlogo, ggLayer = TRUE) + geom_polygon(aes(x, y), fill = "blue", alpha = 0.4) + coord_equal(ylim=c(0,75))library(ggplot2) ggRGB(rlogo, r=1, g=2, b=3) ## Define minMax ranges ggRGB(rlogo, r=1,g=2, b=3, limits = matrix(c(100,150,10,200,50,255), ncol = 2, by = TRUE)) ## Perform stong linear contrast stretch ggRGB(rlogo, r = 1, g = 2, b = 3,stretch = "lin", quantiles = c(0.2, 0.8)) ## Use only two layers for color calculation ggRGB(rlogo, r = 1, g = 2, b = NULL) ## Return only data.frame df <- ggRGB(rlogo, ggObj = FALSE) head(df) ## Use in layer-mode, e.g. to add to another plot wave <- data.frame(x = c(0, 0:100,100), y = c(0,sin(seq(0,2*pi,pi/50))*10+20, 0)) p <- ggplot(wave, aes(x, y)) p + ggRGB(rlogo, ggLayer = TRUE) + geom_polygon(aes(x, y), fill = "blue", alpha = 0.4) + coord_equal(ylim=c(0,75))
Performs image to image contrast adjustments based on histogram matching using empirical cumulative distribution functions from both images.
histMatch( x, ref, xmask = NULL, refmask = NULL, nSamples = 1e+05, intersectOnly = TRUE, paired = TRUE, forceInteger = FALSE, returnFunctions = FALSE, ... )histMatch( x, ref, xmask = NULL, refmask = NULL, nSamples = 1e+05, intersectOnly = TRUE, paired = TRUE, forceInteger = FALSE, returnFunctions = FALSE, ... )
x |
SpatRaster. Source raster which is to be modified. |
ref |
SpatRaster. Reference raster, to which x will be matched. |
xmask |
RasterLayer or SpatRaster. Mask layer for |
refmask |
RasterLayer or SpatRaster. Mask layer for |
nSamples |
Integer. Number of random samples from each image to build the histograms. |
intersectOnly |
Logical. If |
paired |
Logical. If |
forceInteger |
Logical. Force integer output. |
returnFunctions |
Logical. If |
... |
Further arguments to be passed to writeRaster. |
A SpatRaster of x adjusted to the histogram of ref. If returnFunctions = TRUE a list of functions (one for each layer) will be returned instead.
x and ref must have the same number of layers.
Richards and Jia: Remote Sensing Digital Image Analysis. Springer, Berlin, Heidelberg, Germany, 439pp.
library(ggplot2) library(terra) ## Original image a (+1 to prevent log(0)) img_a <- rlogo + 1 ## Degraded image b img_b <- log(img_a) ## Cut-off half the image (just for better display) img_b[, 1:50] <- NA ## Compare Images before histMatching ggRGB(img_a,1,2,3)+ ggRGB(img_b, 1,2,3, ggLayer = TRUE, stretch = "lin", q = 0:1) + geom_vline(aes(xintercept = 50))+ ggtitle("Img_a vs. Img_b") ## Do histogram matching img_b_matched <- histMatch(img_b, img_a) ## Compare Images after histMatching ggRGB(img_a, 1, 2, 3)+ ggRGB(img_b_matched, 1, 2, 3, ggLayer = TRUE, stretch = "lin", q = 0:1) + geom_vline(aes(xintercept = 50))+ ggtitle("Img_a vs. Img_b_matched") ## Histogram comparison opar <- par(mfrow = c(1, 3), no.readonly = TRUE) img_a[,1:50] <- NA redLayers <- c(img_a, img_b, img_b_matched)[[c(1,4,7)]] names(redLayers) <- c("img_a", "img_b", "img_b_matched") hist(redLayers) ## Reset par par(opar)library(ggplot2) library(terra) ## Original image a (+1 to prevent log(0)) img_a <- rlogo + 1 ## Degraded image b img_b <- log(img_a) ## Cut-off half the image (just for better display) img_b[, 1:50] <- NA ## Compare Images before histMatching ggRGB(img_a,1,2,3)+ ggRGB(img_b, 1,2,3, ggLayer = TRUE, stretch = "lin", q = 0:1) + geom_vline(aes(xintercept = 50))+ ggtitle("Img_a vs. Img_b") ## Do histogram matching img_b_matched <- histMatch(img_b, img_a) ## Compare Images after histMatching ggRGB(img_a, 1, 2, 3)+ ggRGB(img_b_matched, 1, 2, 3, ggLayer = TRUE, stretch = "lin", q = 0:1) + geom_vline(aes(xintercept = 50))+ ggtitle("Img_a vs. Img_b_matched") ## Histogram comparison opar <- par(mfrow = c(1, 3), no.readonly = TRUE) img_a[,1:50] <- NA redLayers <- c(img_a, img_b, img_b_matched)[[c(1,4,7)]] names(redLayers) <- c("img_a", "img_b", "img_b_matched") hist(redLayers) ## Reset par par(opar)
ImageMetaData Class
ImageMetaData( file = NA, format = NA, sat = NA, sen = NA, scene = NA, colNum = NA, colTier = NA, proj = NA, date = NA, pdate = NA, path = NA, row = NA, az = NA, selv = NA, esd = NA, files = NA, bands = NA, quant = NA, cat = NA, na = NA, vsat = NA, scal = NA, dtyp = NA, calrad = NA, calref = NA, calbt = NA, radRes = NA, spatRes = NA )ImageMetaData( file = NA, format = NA, sat = NA, sen = NA, scene = NA, colNum = NA, colTier = NA, proj = NA, date = NA, pdate = NA, path = NA, row = NA, az = NA, selv = NA, esd = NA, files = NA, bands = NA, quant = NA, cat = NA, na = NA, vsat = NA, scal = NA, dtyp = NA, calrad = NA, calref = NA, calbt = NA, radRes = NA, spatRes = NA )
file |
Character. Metadata file |
format |
Character. Metadata format, e.g. xml, mtl |
sat |
Character. Satellite platform |
sen |
Character. Sensor |
scene |
Character. Scene_ID |
colNum |
Character Collection number |
colTier |
Character Collection tier |
proj |
CRS. Projection. |
date |
POSIXct. Aquisition date. |
pdate |
POSIXct. Processing date. |
path |
Integer. Path. |
row |
Integer. Row. |
az |
Numeric. Sun azimuth |
selv |
Numeric. Sun elevation |
esd |
Numeric. Earth-sun distance |
files |
Character vector. Files containing the data, e.g. tiff files |
bands |
Character vector. Band names |
quant |
Character vector. Quantity, one of c("dn", "tra", "tre", "sre", "bt", "idx", "angle") |
cat |
Character vector. Category, e.g. c("image", "pan", "index", "qa", "aux") |
na |
Numeric vector. No-data value per band |
vsat |
Numeric vector. Saturation value per band |
scal |
Numeric vector. Scale factor per band. e.g. if data was scaled to 1000*reflectance for integer conversion. |
dtyp |
Character vector. Data type per band. |
calrad |
data.frame. Calibration coefficients for dn->radiance conversion. Must have columns 'gain' and 'offset'. Rows named according to |
calref |
data.frame. Calibration coefficients for dn->reflectance conversion. Must have columns 'gain' and 'offset'. Rows named according to |
calbt |
data.frame. Calibration coefficients for dn->brightness temperature conversion. Must have columns 'K1' and 'K2'. Rows named according to |
radRes |
Numeric vector. Radiometric resolution per band. |
spatRes |
Numeric vector. Spatial resolution per band. |
Returns a structured, fully customizable meta-data table of a file
Subset of Landsat 5 TM Scene: LT52240631988227CUB02 Contains all seven bands in DN format.
lsatlsat
ggRGB(lsat, stretch = "sqrt")ggRGB(lsat, stretch = "sqrt")
mesma performs a spectral mixture anlylsis (SMA) or multiple endmember spectral mixture analysis (MESMA) on a multiband raster image.
mesma( img, em, method = "NNLS", iterate = 400, tolerance = 1e-08, n_models = 5, sum_to_one = TRUE, ..., verbose )mesma( img, em, method = "NNLS", iterate = 400, tolerance = 1e-08, n_models = 5, sum_to_one = TRUE, ..., verbose )
img |
SpatRaster. Remote sensing imagery (usually hyperspectral). |
em |
Matrix or data.frame with spectral endmembers. Columns represent the spectral bands (i.e. columns correspond to number of bands in |
method |
Character. Select an unmixing method. Currently, only "NNLS" is implemented. Default is "NNLS".
|
iterate |
Integer. Set maximum iteration per pixel. Processing time could increase the more iterations are made possible. Default is 400. |
tolerance |
Numeric. Tolerance limit representing a nearly zero minimal number. Default is 1e-8. |
n_models |
Logical. Only applies if |
sum_to_one |
Logical. Defines whether a sum-to-one constraint should be applied so that probabilities of endmember classes sum to one (a constraint not covered by NNLS) to be interpretable as fractions. Defaults to |
... |
further arguments passed to writeRaster. |
verbose |
Logical. Prints progress messages during execution. |
Argument em determines whether an SMA (each row representing a single endmember per class) or a MESMA (multiple endmembers per class differentiate using the class column) is computed.
If multiple endmembers per class are provided, mesma will compute a number of SMA (determined by argument n_models) for multiple endmember combinations drawn from em and will select the best fit per pixel based on the lowest RMSE, based on the MESMA approach proposed by Roberts et al. (1998).
SpatRaster. The object will contain one band per class, with each value representing the estimated probability of the respective endmember class per pixel, and an RMSE band. If sum_to_one is TRUE (default), values of the class bands can be interpreted as fractions per endmember class (0 to 1).
Depending on iterate and tolerance settings and the selected endmembers, the sum of estimated probabilities per pixel varies around 1. NNLS does not account for a sum-to-one constraint. Use sum_to_one to sum to one post-NNLS.
To achieve best results, it is recommended to adjust n_models in accordance to the number of endemembers per class provided through em so that as many endmember combinations as possible (with each endmember being used once) are computed. The more models are being calculated, the more processing and memory recourses are needed.
Jakob Schwalb-Willmann
Franc, V., Hlaváč, V., & Navara, M. (2005). Sequential coordinate-wise algorithm for the non-negative least squares problem. In: International Conference on Computer Analysis of Images and Patterns (pp. 407-414). Berlin, Heidelberg.
Roberts, D. A., Gardner, M., Church, R., Ustin, S., Scheer, G., & Green, R. O. (1998). Mapping chaparral in the Santa Monica Mountains using multiple endmember spectral mixture models. Remote sensing of environment, 65(3), 267-279.
library(RStoolbox) library(terra) # to perform a SMA, use a single endmember per class, row by row: em <- data.frame(lsat[c(5294, 47916)]) rownames(em) <- c("forest", "water") # umix the lsat image probs <- mesma(img = lsat, em = em) plot(probs) # to perform a MESMA, use multiple endmembers per class, differntiating them # by a column named 'class': ## Not run: em <- rbind( data.frame(lsat[c(4155, 17018, 53134, 69487, 83704)], class = "forest"), data.frame(lsat[c(22742, 25946, 38617, 59632, 67313)], class = "water") ) # unmix the lsat image probs <- mesma(img = lsat, em = em) plot(probs) # MESMA can also be performed on more than two endmember classes: em <- rbind( data.frame(lsat[c(4155, 17018, 53134, 69487, 83704)], class = "forest"), data.frame(lsat[c(22742, 25946, 38617, 59632, 67313)], class = "water"), data.frame(lsat[c(4330, 1762, 1278, 1357, 17414)], class = "shortgrown") ) # unmix the lsat image probs <- mesma(img = lsat, em = em) plot(probs) ## End(Not run)library(RStoolbox) library(terra) # to perform a SMA, use a single endmember per class, row by row: em <- data.frame(lsat[c(5294, 47916)]) rownames(em) <- c("forest", "water") # umix the lsat image probs <- mesma(img = lsat, em = em) plot(probs) # to perform a MESMA, use multiple endmembers per class, differntiating them # by a column named 'class': ## Not run: em <- rbind( data.frame(lsat[c(4155, 17018, 53134, 69487, 83704)], class = "forest"), data.frame(lsat[c(22742, 25946, 38617, 59632, 67313)], class = "water") ) # unmix the lsat image probs <- mesma(img = lsat, em = em) plot(probs) # MESMA can also be performed on more than two endmember classes: em <- rbind( data.frame(lsat[c(4155, 17018, 53134, 69487, 83704)], class = "forest"), data.frame(lsat[c(22742, 25946, 38617, 59632, 67313)], class = "water"), data.frame(lsat[c(4330, 1762, 1278, 1357, 17414)], class = "shortgrown") ) # unmix the lsat image probs <- mesma(img = lsat, em = em) plot(probs) ## End(Not run)
For each pixel subtracts the mean of the raster layer and optionally divide by its standard deviation.
normImage(img, norm = TRUE, ...)normImage(img, norm = TRUE, ...)
img |
SpatRaster. Image to transform. Transformation will be performed separately for each layer. |
norm |
Logical. Perform normalization (scaling) in addition to centering, i.e. divide by standard deviation. |
... |
further arguments passed to writeRaster. |
Returns a SpatRaster with the same number layers as input layers with each layer being centered and optionally normalized.
library(terra) ## Load example data ## Normalization: Center and Scale rlogo_center_norm <- normImage(rlogo) hist(rlogo_center_norm) ## Centering rlogo_center <- normImage(rlogo, norm = FALSE)library(terra) ## Load example data ## Normalization: Center and Scale rlogo_center_norm <- normImage(rlogo) hist(rlogo_center_norm) ## Centering rlogo_center <- normImage(rlogo, norm = FALSE)
Splits a categorical raster layer (or a vector) into a multilayer raster (or matrix).
oneHotEncode(img, classes, background = 0, foreground = 1, na.rm = FALSE, ...)oneHotEncode(img, classes, background = 0, foreground = 1, na.rm = FALSE, ...)
img |
SpatRaster or integer/numeric vector containing multiple classes |
classes |
integer: vector of classes which should be extracted |
background |
integer: background value (default = 0) |
foreground |
integer: foreground value (default = 1) |
na.rm |
logical: if |
... |
further arguments passed to writeRaster. Ignored if img is not a SpatRaster, but a numeric/integer vector or matrix |
A SpatRaster with as many layers as there are classes. Pixels matching the class of interest are set to 1, backround values by default are set to 0 (see background argument)
sc <- unsuperClass(rlogo, nClasses = 3) ## one-hot encode sc_oneHot <- oneHotEncode(sc$map, classes = c(1,2,3)) ## check results sc_oneHotsc <- unsuperClass(rlogo, nClasses = 3) ## one-hot encode sc_oneHot <- oneHotEncode(sc$map, classes = c(1,2,3)) ## check results sc_oneHot
provides different methods for pan sharpening a coarse resolution (typically multispectral) image with a higher reolution panchromatic image. Values of the pan-chromatic and multispectral images must be of the same scale, (e.g. from 0:1, or all DNs from 0:255)
panSharpen(img, pan, r, g, b, pc = 1, method = "brovey", norm = TRUE)panSharpen(img, pan, r, g, b, pc = 1, method = "brovey", norm = TRUE)
img |
SpatRaster. Coarse resolution multispectral image |
pan |
SpatRaster. High resolution image, typically panchromatic. |
r |
Character or Integer. Red band in |
g |
Character or Integer. Green band in |
b |
Character or Integer. Blue band in |
pc |
Integer. Only relevant if |
method |
Character. Choose method from c("pca", "ihs", "brovey"). |
norm |
Logical. Rescale pan image to match the 1st PC component. Only relevant if |
Pan sharpening options:
method='pca': Performs a pca using rasterPCA. The first component is then swapped for the pan band an the PCA is rotated backwards.
method='ihs': Performs a color space transform to Intensity-Hue-Saturation space, swaps intensity for the histogram matched pan and does the backwards transformation.
method='brovey': Performs Brovey reweighting. Pan and img must be at the same value scale (e.g. 0:1, or 0:255) otherwise you'll end up with psychodelic colors.
pan-sharpened SpatRaster
library(terra) library(ggplot2) ## Fake panchromatic image (30m resolution covering ## the visible range (integral from blue to red)) pan <- sum(lsat[[1:3]]) ggR(pan, stretch = "lin") ## Fake coarse resolution image (150m spatial resolution) lowResImg <- aggregate(lsat, 5) ## Brovey pan sharpening lowResImg_pan <- panSharpen(lowResImg, pan, r = 3, g = 2, b = 1, method = "brovey") lowResImg_pan ## Plot ggRGB(lowResImg, stretch = "lin") + ggtitle("Original") ggRGB(lowResImg_pan, stretch="lin") + ggtitle("Pansharpened (Brovey)")library(terra) library(ggplot2) ## Fake panchromatic image (30m resolution covering ## the visible range (integral from blue to red)) pan <- sum(lsat[[1:3]]) ggR(pan, stretch = "lin") ## Fake coarse resolution image (150m spatial resolution) lowResImg <- aggregate(lsat, 5) ## Brovey pan sharpening lowResImg_pan <- panSharpen(lowResImg, pan, r = 3, g = 2, b = 1, method = "brovey") lowResImg_pan ## Plot ggRGB(lowResImg, stretch = "lin") + ggtitle("Original") ggRGB(lowResImg_pan, stretch="lin") + ggtitle("Pansharpened (Brovey)")
Match one scene to another based on linear regression of pseudo-invariant features (PIF).
pifMatch( img, ref, method = "cor", quantile = 0.95, returnPifMap = TRUE, returnSimMap = TRUE, returnModels = FALSE )pifMatch( img, ref, method = "cor", quantile = 0.95, returnPifMap = TRUE, returnSimMap = TRUE, returnModels = FALSE )
img |
SpatRaster. Image to be adjusted. |
ref |
SpatRaster. Reference image. |
method |
Method to calculate pixel similarity. Options: euclidean distance ('ed'), spectral angle ('sam') or pearson correlation coefficient ('cor'). |
quantile |
Numeric. Threshold quantile used to identify PIFs |
returnPifMap |
Logical. Return a binary raster map ot pixels which were identified as pesudo-invariant features. |
returnSimMap |
Logical. Return the similarity map as well |
returnModels |
Logical. Return the linear models along with the adjusted image. |
The function consists of three main steps:
First, it calculates pixel-wise similarity between the two rasters and identifies pseudo-invariant pixels based on
a similarity threshold.
In the second step the values of the pseudo-invariant pixels are regressed against each other in a linear model for each layer.
Finally the linear models are applied to all pixels in the img, thereby matching it to the reference scene.
Pixel-wise similarity can be calculated using one of three methods: euclidean distance (method = "ed"), spectral angle ("sam") or pearsons correlation coefficient ("cor").
The threshold is defined as a similarity quantile. Setting quantile=0.95 will select all pixels with a similarity above the 95% quantile as pseudo-invariant features.
Model fitting is performed with simple linear models (lm); fitting one model per layer.
Returns a List with the adjusted image and intermediate products (if requested).
img: the adjusted image
simMap: pixel-wise similarity map (if returnSimMap = TRUE)
pifMap: binary map of pixels selected as pseudo-invariant features (if returnPifMap = TRUE)
models: list of linear models; one per layer (if returnModels = TRUE)
library(terra) ## Create fake example data ## In practice this would be an image from another acquisition date lsat_b <- log(lsat) ## Run pifMatch and return similarity layer, invariant features mask and models lsat_b_adj <- pifMatch(lsat_b, lsat, returnPifMap = TRUE, returnSimMap = TRUE, returnModels = TRUE) ## Pixelwise similarity ggR(lsat_b_adj$simMap, geom_raster = TRUE) ## Pesudo invariant feature mask ggR(lsat_b_adj$pifMap) ## Histograms of changes par(mfrow=c(1,3)) hist(lsat_b[[1]], main = "lsat_b") hist(lsat[[1]], main = "reference") hist(lsat_b_adj$img[[1]], main = "lsat_b adjusted") ## Model summary for first band summary(lsat_b_adj$models[[1]])library(terra) ## Create fake example data ## In practice this would be an image from another acquisition date lsat_b <- log(lsat) ## Run pifMatch and return similarity layer, invariant features mask and models lsat_b_adj <- pifMatch(lsat_b, lsat, returnPifMap = TRUE, returnSimMap = TRUE, returnModels = TRUE) ## Pixelwise similarity ggR(lsat_b_adj$simMap, geom_raster = TRUE) ## Pesudo invariant feature mask ggR(lsat_b_adj$pifMap) ## Histograms of changes par(mfrow=c(1,3)) hist(lsat_b[[1]], main = "lsat_b") hist(lsat[[1]], main = "reference") hist(lsat_b_adj$img[[1]], main = "lsat_b adjusted") ## Model summary for first band summary(lsat_b_adj$models[[1]])
applies a kmeans cluster model to all pixels of a raster. Useful if you want to apply a kmeans model of scene A to scene B.
## S3 method for class 'unsuperClass' predict(object, img, output = "classes", ...)## S3 method for class 'unsuperClass' predict(object, img, output = "classes", ...)
object |
unsuperClass object |
img |
Raster object. Layernames must correspond to layernames used to train the superClass model, i.e. layernames in the original raster image. |
output |
Character. Either 'classes' (kmeans class; default) or 'distances' (euclidean distance to each cluster center). |
... |
further arguments to be passed to writeRaster, e.g. filename |
Returns a raster with the K-means distances base on your object passed in the arguments.
## Load training data ## Perform unsupervised classification uc <- unsuperClass(rlogo, nClasses = 10) ## Apply the model to another raster map <- predict(uc, rlogo)## Load training data ## Perform unsupervised classification uc <- unsuperClass(rlogo, nClasses = 10) ## Apply the model to another raster map <- predict(uc, rlogo)
Implements several different methods for radiometric calibration and correction of Landsat data. You can either specify a metadata file, or supply all neccesary values manually. With proper parametrization apref and sdos should work for other sensors as well.
radCor( img, metaData, method = "apref", bandSet = "full", hazeValues, hazeBands, atmosphere, darkProp = 0.01, clamp = TRUE, verbose )radCor( img, metaData, method = "apref", bandSet = "full", hazeValues, hazeBands, atmosphere, darkProp = 0.01, clamp = TRUE, verbose )
img |
SpatRaster |
metaData |
object of class ImageMetaData or a path to the meta data (MTL) file. |
method |
Radiometric conversion/correction method to be used. There are currently four methods available (see Details): "rad", "apref", "sdos", "dos", "costz". |
bandSet |
Numeric or character. original Landsat band numbers or names in the form of ("B1", "B2" etc). If set to 'full' all bands in the solar (optical) region will be processed. |
hazeValues |
Numeric. Either a vector with dark DNs per |
hazeBands |
Character or integer. Bands corresponding to |
atmosphere |
Character. Atmospheric characteristics. Will be estimated if not expicilty provided. Must be one of |
darkProp |
Numeric. Estimated proportion of dark pixels in the scene. Used only for automatic guessing of hazeValues (typically one would choose 1 or 2%). |
clamp |
Logical. Enforce valid value range. By default reflectance will be forced to stay within [0,1] and radiance >= 0 by replacing invalid values with the correspinding boundary, e.g. -0.1 will become 0. |
verbose |
Logical. Print status information. |
The atmospheric correction methods (sdos, dos and costz) apply to the optical (solar) region of the spectrum and do not affect the thermal band.
Dark object subtraction approaches rely on the estimation of atmospheric haze based on *dark* pixels. Dark pixels are assumed to have zero reflectance, hence the name. It is then assumed further that any radiation originating from such *dark* pixels is due to atmospheric haze and not the reflectance of the surface itself.
The folloiwing methods are available:
| rad | Radiance |
| apref | Apparent reflectance (top-of-atmosphere reflectance) |
| dos | Dark object subtratction following Chavez (1989) |
| costz | Dark object subtraction following Chavez (1996) |
| sdos | Simple dark object subtraction. Classical DOS, Lhaze must be estimated for each band separately. |
If either "dos" or "costz" are selected, radCor will use the atmospheric haze decay model described by Chavez (1989).
Depending on the atmosphere the following coefficients are used:
| veryClear | |
| clear | |
| moderate | |
| hazy | |
| veryHazy | |
For Landsat 8, no values for extra-terrestrial irradiation (esun) are provided by NASA. These are, however, neccessary for DOS-based approaches. Therefore, these values were derived from a standard reference spectrum published by Thuillier et al. (2003) using the Landsat 8 OLI spectral response functions
The implemented sun-earth distances neglect the earth's eccentricity. Instead we use a 100 year daily average (1979-2070).
SpatRaster with top-of-atmosphere radiance (), at-satellite brightness temperature (K),
top-of-atmosphere reflectance (unitless) corrected for the sun angle or at-surface reflectance (unitless).
This was originally a fork of randcorr() function in the landsat package. This version works on SpatRasters and hence is suitable for large rasters.
S. Goslee (2011): Analyzing Remote Sensing Data in R: The landsat Package. Journal of Statistical Software 43(4).
G. Thuillier et al. (2003) THE SOLAR SPECTRAL IRRADIANCE FROM 200 TO 2400 nm AS MEASURED BY THE SOLSPEC SPECTROMETER FROM THE ATLAS AND EURECA MISSIONS. Solar Physics 214(1): 1-22 (
library(terra) ## Import meta-data and bands based on MTL file mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") metaData <- readMeta(mtlFile) lsat_t <- stackMeta(mtlFile) ## Convert DN to top of atmosphere reflectance and brightness temperature lsat_ref <- radCor(lsat_t, metaData = metaData, method = "apref") ## Correct DN to at-surface-reflecatance with DOS (Chavez decay model) lsat_sref <- radCor(lsat_t, metaData = metaData) ## Correct DN to at-surface-reflecatance with simple DOS ## Automatic haze estimation hazeDN <- estimateHaze(lsat_t, hazeBands = 1:4, darkProp = 0.01, plot = FALSE) lsat_sref <- radCor(lsat_t, metaData = metaData, method = "sdos", hazeValues = hazeDN, hazeBands = 1:4)library(terra) ## Import meta-data and bands based on MTL file mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") metaData <- readMeta(mtlFile) lsat_t <- stackMeta(mtlFile) ## Convert DN to top of atmosphere reflectance and brightness temperature lsat_ref <- radCor(lsat_t, metaData = metaData, method = "apref") ## Correct DN to at-surface-reflecatance with DOS (Chavez decay model) lsat_sref <- radCor(lsat_t, metaData = metaData) ## Correct DN to at-surface-reflecatance with simple DOS ## Automatic haze estimation hazeDN <- estimateHaze(lsat_t, hazeBands = 1:4, darkProp = 0.01, plot = FALSE) lsat_sref <- radCor(lsat_t, metaData = metaData, method = "sdos", hazeValues = hazeDN, hazeBands = 1:4)
Calculates angle and magnitude of change vectors. Dimensionality is limited to two bands per image.
rasterCVA(x, y, tmf = NULL, nct = NULL, ...)rasterCVA(x, y, tmf = NULL, nct = NULL, ...)
x |
SpatRaster with two layers. This will be the reference/origin for the change calculations. Both rasters (y and y) need to correspond to each other, i.e. same resolution, extent and origin. |
y |
SpatRaster with two layers. Both rasters (y and y) need to correspond to each other, i.e. same resolution, extent and origin. |
tmf |
Numeric. Threshold median factor (optional). Used to calculate a threshold magnitude for which pixels are considered stable, i.e. no change. Calculated as |
nct |
Numeric. No-change threshold (optional). Alternative to |
... |
further arguments passed to writeRaster |
Change Vector Analysis (CVA) is used to identify spectral changes between two identical scenes which were acquired at different times. CVA is limited to two bands per image. For each pixel it calculates the change vector in the two-dimensional spectral space. For example for a given pixel in image A and B for the red and nir band the change vector is calculated for the coordinate pairs: (red_A | nir_A) and (red_B | nir_B).
The coordinate system is defined by the order of the input bands: the first band defines the x-axis and the second band the y-axis, respectively. Angles are returned *in degree* beginning with 0 degrees pointing 'north', i.e. the y-axis, i.e. the second band.
Returns a SpatRaster with two layers: change vector angle and change vector magnitude
library(terra) pca <- rasterPCA(lsat)$map ## Do change vector analysis cva <- rasterCVA(pca[[1:2]], pca[[3:4]]) cvalibrary(terra) pca <- rasterPCA(lsat)$map ## Do change vector analysis cva <- rasterCVA(pca[[1:2]], pca[[3:4]]) cva
Shannon entropy is calculated for each pixel based on it's layer values. To be used with categorical / integer valued rasters.
rasterEntropy(img, ...)rasterEntropy(img, ...)
img |
SpatRaster |
... |
additional arguments passed to writeRaster |
Entropy is calculated as -sum(p log(p)); p being the class frequency per pixel.
SpatRaster "entropy"
re <- rasterEntropy(rlogo) ggR(re, geom_raster = TRUE)re <- rasterEntropy(rlogo) ggR(re, geom_raster = TRUE)
Calculates R-mode PCA for SpatRasters and returns a SpatRaster with multiple layers of PCA scores.
rasterPCA( img, nSamples = NULL, nComp = nlyr(img), spca = FALSE, maskCheck = TRUE, ... )rasterPCA( img, nSamples = NULL, nComp = nlyr(img), spca = FALSE, maskCheck = TRUE, ... )
img |
SpatRaster. |
nSamples |
Integer or NULL. Number of pixels to sample for PCA fitting. If NULL, all pixels will be used. |
nComp |
Integer. Number of PCA components to return. |
spca |
Logical. If |
maskCheck |
Logical. Masks all pixels which have at least one NA (default TRUE is reccomended but introduces a slow-down, see Details when it is wise to disable maskCheck). Takes effect only if nSamples is NULL. |
... |
further arguments to be passed to writeRaster, e.g. filename. |
Internally rasterPCA relies on the use of princomp (R-mode PCA). If nSamples is given the PCA will be calculated based on a random sample of pixels and then predicted for the full raster. If nSamples is NULL then the covariance matrix will be calculated first and will then be used to calculate princomp and predict the full raster. The latter is more precise, since it considers all pixels, however, it may be slower than calculating the PCA only on a subset of pixels.
Pixels with missing values in one or more bands will be set to NA. The built-in check for such pixels can lead to a slow-down of rasterPCA. However, if you make sure or know beforehand that all pixels have either only valid values or only NAs throughout all layers you can disable this check by setting maskCheck=FALSE which speeds up the computation.
Standardised PCA (SPCA) can be useful if imagery or bands of different dynamic ranges are combined. SPC uses the correlation matrix instead of the covariance matrix, which has the same effect as using normalised bands of unit variance.
Returns a named list containing the PCA model object ($model) and a SpatRaster with the principal component layers ($object).
library(ggplot2) library(reshape2) ggRGB(rlogo, 1,2,3) ## Run PCA set.seed(25) rpc <- rasterPCA(rlogo) rpc ## Model parameters: summary(rpc$model) loadings(rpc$model) ggRGB(rpc$map,1,2,3, stretch="lin", q=0) if(require(gridExtra)){ plots <- lapply(1:3, function(x) ggR(rpc$map, x, geom_raster = TRUE)) grid.arrange(plots[[1]],plots[[2]], plots[[3]], ncol=2) }library(ggplot2) library(reshape2) ggRGB(rlogo, 1,2,3) ## Run PCA set.seed(25) rpc <- rasterPCA(rlogo) rpc ## Model parameters: summary(rpc$model) loadings(rpc$model) ggRGB(rpc$map,1,2,3, stretch="lin", q=0) if(require(gridExtra)){ plots <- lapply(1:3, function(x) ggR(rpc$map, x, geom_raster = TRUE)) grid.arrange(plots[[1]],plots[[2]], plots[[3]], ncol=2) }
Imports and tidies CSV files exported from EarthExplorer into data.frames and annotates missing fields.
readEE(x)readEE(x)
x |
Character, Character or list. One or more paths to EarthExplorer export files. |
The EarthExplorer CSV file can be produced from the search results page. Above the results click on 'export results' and select 'comma (,) delimited'.
Note that only a subset of columns is imported which was deemed interesting. Please contact the maintainer if you think an omited column should be included.
data.frame
library(ggplot2) ee <- readEE(system.file("external/EarthExplorer_LS8.txt", package = "RStoolbox")) ## Scenes with cloud cover < 20% ee[ee$Cloud.Cover < 20,] ## Available time-series ggplot(ee) + geom_segment(aes(x = Date, xend = Date, y = 0, yend = 100 - Cloud.Cover, col = as.factor(Year))) + scale_y_continuous(name = "Scene quality (% clear sky)")library(ggplot2) ee <- readEE(system.file("external/EarthExplorer_LS8.txt", package = "RStoolbox")) ## Scenes with cloud cover < 20% ee[ee$Cloud.Cover < 20,] ## Available time-series ggplot(ee) + geom_segment(aes(x = Date, xend = Date, y = 0, yend = 100 - Cloud.Cover, col = as.factor(Year))) + scale_y_continuous(name = "Scene quality (% clear sky)")
Reads metadata and deals with legacy versions of Landsat metadata files and where possible adds missing information (radiometric gain and offset, earth-sun distance).
readMeta(file, raw = FALSE)readMeta(file, raw = FALSE)
file |
path to Landsat MTL file (...MTL.txt) |
raw |
Logical. If |
Object of class ImageMetaData
## Example metadata file (MTL) mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") ## Read metadata metaData <- readMeta(mtlFile) ## Summary summary(metaData)## Example metadata file (MTL) mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") ## Read metadata metaData <- readMeta(mtlFile) ## Summary summary(metaData)
read/write support for ENVI spectral libraries
readSLI(path)readSLI(path)
path |
Path to spectral library file with ending .sli. |
ENVI spectral libraries consist of a binary data file (.sli) and a corresponding header (.hdr, or .sli.hdr) file.
The spectral libraries are read into a data.frame. The first column contains the wavelengths and the remaining columns contain the spectra.
## Example data sliFile <- system.file("external/vegSpec.sli", package="RStoolbox") sliTmpFile <- paste0(tempdir(),"/vegetationSpectra.sli") ## Read spectral library sli <- readSLI(sliFile) head(sli) plot(sli[,1:2], col = "orange", type = "l") lines(sli[,c(1,3)], col = "green") ## Write to binary spectral library writeSLI(sli, path = sliTmpFile)## Example data sliFile <- system.file("external/vegSpec.sli", package="RStoolbox") sliTmpFile <- paste0(tempdir(),"/vegetationSpectra.sli") ## Read spectral library sli <- readSLI(sliFile) head(sli) plot(sli[,1:2], col = "orange", type = "l") lines(sli[,c(1,3)], col = "green") ## Write to binary spectral library writeSLI(sli, path = sliTmpFile)
performs linear shifts of value ranges either to match min/max of another image (y)
or to any other min and max value (ymin and ymax).
rescaleImage(x, y, xmin, xmax, ymin, ymax, forceMinMax = FALSE, ...)rescaleImage(x, y, xmin, xmax, ymin, ymax, forceMinMax = FALSE, ...)
x |
SpatRaster or numeric vector. Image to normalise. |
y |
SpatRaster or numeric vector. Reference image. Optional. Used to extract min and max values if ymin or ymax are missing. |
xmin |
Numeric. Min value of x. Either a single value or one value per layer in x. If xmin is not provided it will be extracted from x. |
xmax |
Numeric. Max value of x. Either a single value or one value per layer in x. If xmax is not provided it will be extracted from x. |
ymin |
Numeric. Min value of y. Either a single value or one value per layer in x. If ymin is not provided it will be extracted from y. |
ymax |
Numeric. Max value of y. Either a single value or one value per layer in x. If ymax is not provided it will be extracted from y. |
forceMinMax |
Logical. Forces update of min and max data slots in x or y. |
... |
additional arguments passed to |
Providing xmin and xmax values manually can be useful if the raster contains a variable of a known, fixed value range,
e.g. NDVI from -1 to 1 but the actual pixel values don't encompass this entire range.
By providing xmin = -1 and xmax = 1 the values can be rescaled to any other range,
e.g. 1 to 100 while comparability to other rescaled NDVI scenes is retained.
Returns a SpatRaster of the same dimensions as the input raster x but shifted and stretched to the new limits.
lsat2 <- lsat - 1000 lsat2 ## Rescale lsat2 to match original lsat value range lsat2_rescaled <- rescaleImage(lsat2, lsat) lsat2_rescaled ## Rescale lsat to value range [0,1] lsat2_unity <- rescaleImage(lsat2, ymin = 0, ymax = 1) lsat2_unitylsat2 <- lsat - 1000 lsat2 ## Rescale lsat2 to match original lsat value range lsat2_rescaled <- rescaleImage(lsat2, lsat) lsat2_rescaled ## Rescale lsat to value range [0,1] lsat2_unity <- rescaleImage(lsat2, ymin = 0, ymax = 1) lsat2_unity
Tiny example of raster data used to run examples.
rlogorlogo
ggRGB(rlogo,r = 1,g = 2,b = 3)ggRGB(rlogo,r = 1,g = 2,b = 3)
shortcut to options(RStoolbox.*)
rsOpts(verbose)rsOpts(verbose)
verbose |
Logical. If |
No return, just a setter for the verbosiness of the RStoolbox package
rsOpts(verbose=TRUE)rsOpts(verbose=TRUE)
The RStoolbox package provides a set of functions which simplify performing standard remote sensing tasks in R.
readMeta: import Landsat metadata from MTL or XML files
saveRSTBX & readRSTBX: save and re-import RStoolbox classification objects (model and map)
readEE: import and tidy EarthExplorer search results
radCor: radiometric conversions and corrections. Primarily, yet not exclusively, intended for Landsat data processing. DN to radiance to reflectance conversion as well as DOS approaches
topCor: topographic illumination correction
cloudMask & cloudShadowMask: mask clouds and cloud shadows in Landsat or other imagery which comes with a thermal band
classifyQA: extract layers from Landsat 8 QA bands, e.g. cloud confidence
rescaleImage: rescale image to match min/max from another image or a specified min/max range
normImage: normalize imagery by centering and scaling
oneHotEncode: one-hot encode a raster or vector
histMatch: matches the histograms of two scenes
pifMatch: matches one scene to another based on linear regression of Pseudo-Invariant Features (PIF)
coregisterImages: co-register images based on mutual information
panSharpen: sharpen a coarse resolution image with a high resolution image (typically panchromatic)
estimateHaze: estimate image haze for Dark Object Subtraction (DOS)
spectralIndices: calculate a set of predefined multispectral indices like NDVI
tasseledCap: tasseled cap transformation
sam: spectral angle mapper
rasterPCA: principal components transform for raster data
rasterCVA: change vector analysis
rasterEntropy: calculates shannon entropy
unsuperClass: unsupervised classification
superClass, validateMap, getValidation: supervised classification and validation
fCover: fractional cover of coarse resolution imagery based on high resolution classification
mesma: spectral unmixing using Multiple Endmember Spectral Mixture Analysis (MESMA)
ggR: single raster layer plotting with ggplot2
ggRGB: efficient plotting of remote sensing imagery in RGB with ggplot2
Calculates the angle in spectral space between pixels and a set of reference spectra (endmembers) for image classification based on spectral similarity.
sam(img, em, angles = FALSE, ...)sam(img, em, angles = FALSE, ...)
img |
SpatRaster. Remote sensing imagery. |
em |
Matrix or data.frame with endmembers. Each row should contain the endmember spectrum of a class, i.e. columns correspond to bands in |
angles |
Logical. If |
... |
further arguments to be passed to |
For each pixel the spectral angle mapper calculates the angle between the vector defined by the pixel values and each endmember vector. The result of this is one raster layer for each endmember containing the spectral angle. The smaller the spectral angle the more similar a pixel is to a given endmember class. In a second step one can the go ahead an enforce thresholds of maximum angles or simply classify each pixel to the most similar endmember.
SpatRaster
If angles = FALSE a single Layer will be returned in which each pixel is assigned to the closest endmember class (integer pixel values correspond to row order of em.
library(terra) library(ggplot2) ## Sample endmember spectra ## First location is water, second is open agricultural vegetation pts <- data.frame(x = c(624720, 627480), y = c(-414690, -411090)) endmembers <- extract(lsat, pts) rownames(endmembers) <- c("water", "vegetation") ## Calculate spectral angles lsat_sam <- sam(lsat, endmembers, angles = TRUE) plot(lsat_sam) ## Classify based on minimum angle lsat_sam <- sam(lsat, endmembers, angles = FALSE) ggR(lsat_sam, forceCat = TRUE, geom_raster=TRUE) + scale_fill_manual(values = c("blue", "green"), labels = c("water", "vegetation"))library(terra) library(ggplot2) ## Sample endmember spectra ## First location is water, second is open agricultural vegetation pts <- data.frame(x = c(624720, 627480), y = c(-414690, -411090)) endmembers <- extract(lsat, pts) rownames(endmembers) <- c("water", "vegetation") ## Calculate spectral angles lsat_sam <- sam(lsat, endmembers, angles = TRUE) plot(lsat_sam) ## Classify based on minimum angle lsat_sam <- sam(lsat, endmembers, angles = FALSE) ggR(lsat_sam, forceCat = TRUE, geom_raster=TRUE) + scale_fill_manual(values = c("blue", "green"), labels = c("water", "vegetation"))
Saves objects of classes unsuperClass, superClass, rasterPCA and fCover to file. Useful to archive the fitted models.
saveRSTBX(x, filename, format = "raster", ...) readRSTBX(filename)saveRSTBX(x, filename, format = "raster", ...) readRSTBX(filename)
x |
RStoolbox object of classes c("fCover", "rasterPCA", "superClass", "unsuperClass") |
filename |
Character. Path and filename. Any file extension will be ignored. |
format |
Character. Driver to use for the raster file |
... |
further arguments passed to writeRaster |
The output of writeRSTBX will be at least two files written to disk: a) an .rds file containing the object itself and b) the raster file (depending on the driver you choose this can be more than two files).
saveRSTBX(): Save RStoolbox object to file
readRSTBX(): Read files saved with saveRSTBX
All files must be kept in the same directory to read the full object back into R by means of readRSTBX. You can move them to another location but you'll have to move *all* of them (just like you would with Shapefiles). In case the raster file(s) is missing, readRSTBX will still return the object but the raster will be missing.
writeRSTBX and readRSTBX are convenience wrappers around saveRDS, readRDS. This means you can read all files created this way also with base functionality as long as you don't move your files. This is because x$map is a SpatRaster object and hence contains only a static link to the file on disk.
## Not run: input <- rlogo ## Create filename file <- paste0(tempdir(), "/test", runif(1)) ## Run PCA rpc <- rasterPCA(input, nSample = 100) ## Save object saveRSTBX(rpc, filename=file) ## Which files were written? list.files(tempdir(), pattern = basename(file)) ## Re-read files re_rpc <- readRSTBX(file) ## Remove files file.remove(list.files(tempdir(), pattern = basename(file), full = TRUE)) ## End(Not run)## Not run: input <- rlogo ## Create filename file <- paste0(tempdir(), "/test", runif(1)) ## Run PCA rpc <- rasterPCA(input, nSample = 100) ## Save object saveRSTBX(rpc, filename=file) ## Which files were written? list.files(tempdir(), pattern = basename(file)) ## Re-read files re_rpc <- readRSTBX(file) ## Remove files file.remove(list.files(tempdir(), pattern = basename(file), full = TRUE)) ## End(Not run)
Contains all 13 bands in already converted spectral reflectances
sen2sen2
ggRGB(sen2, r=4, g=3, b=2, stretch = "lin")ggRGB(sen2, r=4, g=3, b=2, stretch = "lin")
Calculate a suite of multispectral indices such as NDVI, SAVI etc. in an efficient way.
spectralIndices( img, blue = NULL, green = NULL, red = NULL, nir = NULL, redEdge1 = NULL, redEdge2 = NULL, redEdge3 = NULL, swir1 = NULL, swir2 = NULL, swir3 = NULL, scaleFactor = 1, skipRefCheck = FALSE, indices = NULL, index = NULL, maskLayer = NULL, maskValue = 1, coefs = list(L = 0.5, G = 2.5, L_evi = 1, C1 = 6, C2 = 7.5, s = 1, swir2ccc = NULL, swir2coc = NULL), ... )spectralIndices( img, blue = NULL, green = NULL, red = NULL, nir = NULL, redEdge1 = NULL, redEdge2 = NULL, redEdge3 = NULL, swir1 = NULL, swir2 = NULL, swir3 = NULL, scaleFactor = 1, skipRefCheck = FALSE, indices = NULL, index = NULL, maskLayer = NULL, maskValue = 1, coefs = list(L = 0.5, G = 2.5, L_evi = 1, C1 = 6, C2 = 7.5, s = 1, swir2ccc = NULL, swir2coc = NULL), ... )
img |
SpatRaster. Typically remote sensing imagery, which is to be classified. |
blue |
Character or integer. Blue band. |
green |
Character or integer. Green band. |
red |
Character or integer. Red band. |
nir |
Character or integer. Near-infrared band (700-1100nm). |
redEdge1 |
Character or integer. Red-edge band (705nm) |
redEdge2 |
Character or integer. Red-edge band (740nm) |
redEdge3 |
Character or integer. Red-edge band (783nm) |
swir1 |
not used |
swir2 |
Character or integer. Short-wave-infrared band (1400-1800nm). |
swir3 |
Character or integer. Short-wave-infrared band (2000-2500nm). |
scaleFactor |
Numeric. Scale factor for the conversion of scaled reflectances to [0,1] value range (applied as reflectance/scaleFactor) Neccesary for calculating EVI/EVI2 with scaled reflectance values. |
skipRefCheck |
Logical. When EVI/EVI2 is to be calculated there is a rough heuristic check, whether the data are inside [0,1]+/-0.5 (after applying a potential |
indices |
Character. One or more spectral indices to calculate (see Details). By default (NULL) all implemented indices given the spectral bands which are provided will be calculated. |
index |
Character. Alias for |
maskLayer |
RasterLayer or SpatRaster containing a mask, e.g. clouds, for which pixels are set to NA. Alternatively a layername or -number can be provided if the mask is part of |
maskValue |
Integer. Pixel value in |
coefs |
List of coefficients (see Details). |
... |
further arguments such as filename etc. passed to writeRaster |
spectralIndices calculates all indices in one go in C++, which is more efficient than calculating each index separately (for large rasters).
By default all indices which can be calculated given the specified indices will be calculated. If you don't want all indices, use the indices argument to specify exactly which indices are to be calculated.
See the table bellow for index names and required bands.
Index values outside the valid value ranges (if such a range exists) will be set to NA. For example a pixel with NDVI > 1 will be set to NA.
| Index | Description | Source | Bands | Formula |
| CLG | Green-band Chlorophyll Index | Gitelson2003 | redEdge3, green |
|
| CLRE | Red-edge-band Chlorophyll Index | Gitelson2003 | redEdge3, redEdge1 |
|
| CTVI | Corrected Transformed Vegetation Index | Perry1984 | red, nir |
|
| DVI | Difference Vegetation Index | Richardson1977 | red, nir |
|
| EVI | Enhanced Vegetation Index | Huete1999 | red, nir, blue |
|
| EVI2 | Two-band Enhanced Vegetation Index | Jiang 2008 | red, nir |
|
| GEMI | Global Environmental Monitoring Index | Pinty1992 | red, nir |
|
| GNDVI | Green Normalised Difference Vegetation Index | Gitelson1998 | green, nir |
|
| KNDVI | Kernel Normalised Difference Vegetation Index | Camps-Valls2021 | red, nir |
|
| MCARI | Modified Chlorophyll Absorption Ratio Index | Daughtery2000 | green, red, redEdge1 |
|
| MNDWI | Modified Normalised Difference Water Index | Xu2006 | green, swir2 |
|
| MSAVI | Modified Soil Adjusted Vegetation Index | Qi1994 | red, nir |
|
| MSAVI2 | Modified Soil Adjusted Vegetation Index 2 | Qi1994 | red, nir |
|
| MTCI | MERIS Terrestrial Chlorophyll Index | DashAndCurran2004 | red, redEdge1, redEdge2 |
|
| NBRI | Normalised Burn Ratio Index | Garcia1991 | nir, swir3 |
|
| NDREI1 | Normalised Difference Red Edge Index 1 | GitelsonAndMerzlyak1994 | redEdge2, redEdge1 |
|
| NDREI2 | Normalised Difference Red Edge Index 2 | Barnes2000 | redEdge3, redEdge1 |
|
| NDVI | Normalised Difference Vegetation Index | Rouse1974 | red, nir |
|
| NDVIC | Corrected Normalised Difference Vegetation Index | Nemani1993 | red, nir, swir2 |
|
| NDWI | Normalised Difference Water Index | McFeeters1996 | green, nir |
|
| NDWI2 | Normalised Difference Water Index | Gao1996 | nir, swir2 |
|
| NRVI | Normalised Ratio Vegetation Index | Baret1991 | red, nir |
|
| REIP | Red Edge Inflection Point | GuyotAndBarnet1988 | red, redEdge1, redEdge2, redEdge3 |
|
| RVI | Ratio Vegetation Index | red, nir |
|
|
| SATVI | Soil Adjusted Total Vegetation Index | Marsett2006 | red, swir2, swir3 |
|
| SAVI | Soil Adjusted Vegetation Index | Huete1988 | red, nir |
|
| SLAVI | Specific Leaf Area Vegetation Index | Lymburger2000 | red, nir, swir2 |
|
| SR | Simple Ratio Vegetation Index | Birth1968 | red, nir |
|
| TTVI | Thiam's Transformed Vegetation Index | Thiam1997 | red, nir |
|
| TVI | Transformed Vegetation Index | Deering1975 | red, nir |
|
| WDVI | Weighted Difference Vegetation Index | Richardson1977 | red, nir |
|
Some indices require additional parameters, such as the slope of the soil line which are specified via a list to the coefs argument.
Although the defaults are sensible values, values like the soil brightnes factor L for SAVI should be adapted depending on the characteristics of the scene.
The coefficients are:
| Coefficient | Description | Affected Indices |
s |
slope of the soil line | DVI, WDVI |
L_evi, C1, C2, G |
various | EVI |
L |
soil brightness factor | SAVI, SATVI |
swir2ccc |
minimum swir2 value (completely closed forest canopy) | NDVIC |
swir2coc |
maximum swir2 value (completely open canopy) | NDVIC |
The wavelength band names are defined following Schowengertd 2007, p10. The last column shows exemplarily which Landsat 5 TM bands correspond to which wavelength range definition.
| Band | Description | Wavl_min | Wavl_max | Landsat5_Band | Sentinel2_Band |
| vis | visible | 400 | 680 | 1,2,3 | 2,3,4 |
| red-edge1 | red-edge1 | 680 | 720 | - | 5 |
| red-edge2 | red-edge2 | 720 | 760 | - | 6 |
| red-edge3 | red-edge3 | 760 | 800 | - | 7 |
| nir | near infra-red | 800 | 1100 | 4 | 8/8a |
| swir1 | short-wave infra-red | 1100 | 1351 | - | 9,10 |
| swir2 | short-wave infra-red | 1400 | 1800 | 5 | 11 |
| swir3 | short-wave infra-red | 2000 | 2500 | 7 | 12 |
| mir1 | mid-wave infra-red | 3000 | 4000 | - | - |
| mir2 | mid-wave infra-red | 45000 | 5000 | - | - |
| tir1 | thermal infra-red | 8000 | 9500 | - | - |
| tir2 | thermal infra-red | 10000 | 140000 | 6 | - |
SpatRaster
library(ggplot2) library(terra) ## Calculate NDVI ndvi <- spectralIndices(lsat, red = "B3_dn", nir = "B4_dn", indices = "NDVI") ndvi ggR(ndvi, geom_raster = TRUE) + scale_fill_gradientn(colours = c("black", "white")) ## Calculate all possible indices, given the provided bands ## Convert DNs to reflectance (required to calculate EVI and EVI2) mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") lsat_ref <- radCor(lsat, mtlFile, method = "apref") SI <- spectralIndices(lsat_ref, red = "B3_tre", nir = "B4_tre") plot(SI)library(ggplot2) library(terra) ## Calculate NDVI ndvi <- spectralIndices(lsat, red = "B3_dn", nir = "B4_dn", indices = "NDVI") ndvi ggR(ndvi, geom_raster = TRUE) + scale_fill_gradientn(colours = c("black", "white")) ## Calculate all possible indices, given the provided bands ## Convert DNs to reflectance (required to calculate EVI and EVI2) mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") lsat_ref <- radCor(lsat, mtlFile, method = "apref") SI <- spectralIndices(lsat_ref, red = "B3_tre", nir = "B4_tre") plot(SI)
DEM for the Landsat example area taken from SRTM v3 tile: s04_w050_1arc_v3.tif
srtmsrtm
ggR(srtm)ggR(srtm)
DEM for the Sentinel 2 example area taken from SRTM v4
srtm_sen2srtm_sen2
ggR(srtm_sen2)ggR(srtm_sen2)
Reads Landsat MTL or XML metadata files and loads single Landsat Tiffs into a rasterStack. Be aware that by default stackMeta() does NOT import panchromatic bands nor thermal bands with resolutions != 30m.
stackMeta(file, quantity = "all", category = "image", allResolutions = FALSE)stackMeta(file, quantity = "all", category = "image", allResolutions = FALSE)
file |
Character. Path to Landsat MTL metadata (*_MTL.txt) file or an Landsat CDR xml metadata file (*.xml). |
quantity |
Character vector. Which quantity should be returned. Options: digital numbers ('dn'), top of atmosphere reflectance ('tre'), at surface reflectance ('sre'), brightness temperature ('bt'), spectral index ('index'), all quantities ('all'). |
category |
Character vector. Which category of data to return. Options 'image': image data, 'pan': panchromatic image, 'index': multiband indices, 'qa' quality flag bands, 'all': all categories. |
allResolutions |
Logical. if |
Returns one single SpatRaster comprising all requested bands.
If allResolutions = TRUE *and* there are different resolution layers (e.g. a 15m panchromatic band along wit 30m imagery) a list of RasterStacks will be returned.
Be aware that by default stackMeta() does NOT import panchromatic bands nor thermal bands with resolutions != 30m. Use the allResolutions argument to import all layers. Note that nowadays the USGS uses cubic convolution to resample the TIR bands to 30m resolution.
## Example metadata file (MTL) mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") ## Read metadata metaData <- readMeta(mtlFile) summary(metaData) ## Load rasters based on metadata file lsat <- stackMeta(mtlFile) lsat## Example metadata file (MTL) mtlFile <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") ## Read metadata metaData <- readMeta(mtlFile) summary(metaData) ## Load rasters based on metadata file lsat <- stackMeta(mtlFile) lsat
Supervised classification both for classification and regression mode based on vector training data (points or polygons).
superClass( img, trainData, valData = NULL, responseCol = NULL, nSamples = 1000, nSamplesV = 1000, polygonBasedCV = FALSE, trainPartition = NULL, model = "rf", tuneLength = 3, kfold = 5, minDist = 2, mode = "classification", predict = TRUE, predType = "raw", filename = NULL, verbose, overwrite = TRUE, ... )superClass( img, trainData, valData = NULL, responseCol = NULL, nSamples = 1000, nSamplesV = 1000, polygonBasedCV = FALSE, trainPartition = NULL, model = "rf", tuneLength = 3, kfold = 5, minDist = 2, mode = "classification", predict = TRUE, predType = "raw", filename = NULL, verbose, overwrite = TRUE, ... )
img |
SpatRaster. Typically remote sensing imagery, which is to be classified. |
trainData |
sf or sp spatial vector data containing the training locations (POINTs,or POLYGONs). |
valData |
Ssf or sp spatial vector data containing the validation locations (POINTs,or POLYGONs) (optional). |
responseCol |
Character or integer giving the column in |
nSamples |
Integer. Number of samples per land cover class. If |
nSamplesV |
Integer. Number of validation samples per land cover class. If |
polygonBasedCV |
Logical. If |
trainPartition |
Numeric. Partition (polygon based) of |
model |
Character. Which model to use. See train for options. Defaults to randomForest ('rf'). In addition to the standard caret models, a maximum likelihood classification is available via |
tuneLength |
Integer. Number of levels for each tuning parameter (see train for details). |
kfold |
Integer. Number of cross-validation resamples during model tuning. |
minDist |
Numeric. Minumum distance between training and validation data,
e.g. |
mode |
Character. Model type: 'regression' or 'classification'. |
predict |
Logical. Produce a map (TRUE, default) or only fit and validate the model (FALSE). |
predType |
Character. Type of the final output raster. Either "raw" for class predictions or "prob" for class probabilities. Class probabilities are not available for all classification models (predict.train). |
filename |
Path to output file (optional). If |
verbose |
Logical. prints progress and statistics during execution |
overwrite |
logical. Overwrite spatial prediction raster if it already exists. |
... |
further arguments to be passed to |
SuperClass performs the following steps:
Ensure non-overlap between training and validation data. This is neccesary to avoid biased performance estimates.
A minimum distance (minDist) in pixels can be provided to enforce a given distance between training and validation data.
Sample training coordinates. If trainData (and valData if present) are polygons superClass will calculate the area per polygon and sample
nSamples locations per class within these polygons. The number of samples per individual polygon scales with the polygon area, i.e. the bigger the polygon, the more samples.
Split training/validation
If valData was provided (reccomended) the samples from these polygons will be held-out and not used for model fitting but only for validation.
If trainPartition is provided the trainingPolygons will be divided into training polygons and validation polygons.
Extract raster data
The predictor values on the sample pixels are extracted from img
Fit the model. Using caret::train on the sampled training data the model will be fit,
including parameter tuning (tuneLength) in kfold cross-validation. polygonBasedCV=TRUE will define cross-validation folds based on polygons (reccomended)
otherwise it will be performed on a per-pixel basis.
Predict the classes of all pixels in img based on the final model.
Validate the model with the independent validation data.
A superClass object (effectively a list) containing:
$model: the fitted model
$modelFit: model fit statistics
$training: indexes of samples used for training
$validation: list of
$performance: performance estimates based on independent validation (confusion matrix etc.)
$validationSamples: actual pixel coordinates plus reference and predicted values used for validation
$validationGeometry: validation polygpns (clipped with mindist to training geometries)
$map: the predicted raster
$classMapping: a data.frame containing an integer <-> label mapping
library(RStoolbox) library(caret) library(randomForest) library(e1071) library(terra) train <- readRDS(system.file("external/trainingPoints_rlogo.rds", package="RStoolbox")) ## Plot training data olpar <- par(no.readonly = TRUE) # back-up par par(mfrow=c(1,2)) colors <- c("yellow", "green", "deeppink") plotRGB(rlogo) plot(train, add = TRUE, col = colors[train$class], pch = 19) ## Fit classifier (splitting training into 70\% training data, 30\% validation data) SC <- superClass(rlogo, trainData = train, responseCol = "class", model = "rf", tuneLength = 1, trainPartition = 0.7) SC ## Plots plot(SC$map, col = colors, legend = FALSE, axes = FALSE, box = FALSE) legend(1,1, legend = levels(train$class), fill = colors , title = "Classes", horiz = TRUE, bty = "n") par(olpar) # reset parlibrary(RStoolbox) library(caret) library(randomForest) library(e1071) library(terra) train <- readRDS(system.file("external/trainingPoints_rlogo.rds", package="RStoolbox")) ## Plot training data olpar <- par(no.readonly = TRUE) # back-up par par(mfrow=c(1,2)) colors <- c("yellow", "green", "deeppink") plotRGB(rlogo) plot(train, add = TRUE, col = colors[train$class], pch = 19) ## Fit classifier (splitting training into 70\% training data, 30\% validation data) SC <- superClass(rlogo, trainData = train, responseCol = "class", model = "rf", tuneLength = 1, trainPartition = 0.7) SC ## Plots plot(SC$map, col = colors, legend = FALSE, axes = FALSE, box = FALSE) legend(1,1, legend = levels(train$class), fill = colors , title = "Classes", horiz = TRUE, bty = "n") par(olpar) # reset par
Calculates brightness, greenness and wetness from multispectral imagery. Currently implemented Landsat 4 TM, Landsat 5 TM, Landsat 7ETM+, Landsat 8 OLI, MODIS, QuickBird, Spot5 and RapidEye.
tasseledCap(img, sat, ...)tasseledCap(img, sat, ...)
img |
SpatRaster. Input image. Band order must correspond to sensor specifications (see Details and Examples) |
sat |
Character. Sensor; one of: c("Landsat4TM", "Landsat5TM", "Landsat7ETM", "Landsat8OLI", "MODIS", "QuickBird", "Spot5", "RapidEye"). Case is irrelevant. |
... |
Further arguments passed to writeRaster. |
Currently implemented: Landsat 4 TM, Landsat 5 TM, Landsat 7ETM+, Landsat 8 OLI, MODIS, QuickBird, Spot5, RapdiEye. Input data must be in top of atmosphere reflectance. Moreover, bands must be provided in ascending order as listed in the table below. Irrelevant bands, such as Landsat Thermal Bands or QuickBird/Spot5 Panchromatic Bands must be omitted. Required bands are:
| sat | bands | coefficients | data unit |
| Landsat4TM | 1,2,3,4,5,7 | Crist 1985 | reflectance |
| Landsat5TM | 1,2,3,4,5,7 | Crist 1985 | reflectance |
| Landsat7ETM | 1,2,3,4,5,7 | Huang 2002 | reflectance |
| Landsat8OLI | 2,3,4,5,6,7 | Baig 2014 | reflectance |
| MODIS | 1,2,3,4,5,6,7 | Lobser 2007 | reflectance |
| QuickBird | 2,3,4,5 | Yarbrough 2005 | reflectance |
| Spot5 | 2,3,4,5 | Ivtis 2008 | reflectance |
| RapidEye | 1,2,3,4,5 | Schoenert 2014 | reflectance |
Returns a SpatRaster with the thee bands: brigthness, greenness, and (soil) wetness.
Crist (1985) "A TM Tasseled Cap Equivalent Transformation for Reflectance Factor Data." Remote Sensing of Environment 17 (3): 301-306
Huang et al. (2002) "Derivation of a Tasselled Cap Transformation Based on Landsat 7 At-Satellite Reflectance." International Journal of Remote Sensing 23 (8): 1741-1748
Baig et al. (2014) "Derivation of a Tasselled Cap Transformation Based on Landsat 8 At-Satellite Reflectance." Remote Sensing Letters 5 (5): 423-431.
Lobser et al. (2007) "MODIS Tasselled Cap: Land Cover Characteristics Expressed through Transformed MODIS Data." International Journal of Remote Sensing 28 (22): 5079-5101.
Yarbrough et al. (2005) "QuickBird 2 tasseled cap transform coefficients: a comparison of derivation methods." Pecora 16 Global Priorities in Land Remote Sensing: 23-27.
Ivits et al. (2008) "Orthogonal transformation of segmented SPOT5 images." Photogrammetric Engineering & Remote Sensing 74 (11): 1351-1364.
Schoenert et al. (2014) "Derivation of tasseled cap coefficients for RapidEye data." Earth Resources and Environmental Remote Sensing/GIS Applications V (9245): 92450Qs.
library(terra) ## Run tasseled cap (exclude thermal band 6) lsat_tc <- tasseledCap(lsat[[c(1:5,7)]], sat = "Landsat5TM") lsat_tc plot(lsat_tc)library(terra) ## Run tasseled cap (exclude thermal band 6) lsat_tc <- tasseledCap(lsat[[c(1:5,7)]], sat = "Landsat5TM") lsat_tc plot(lsat_tc)
account and correct for changes in illumination due to terrain elevation.
topCor( img, dem, metaData, solarAngles = c(), method = "C", stratImg = NULL, nStrat = 5, illu, ... )topCor( img, dem, metaData, solarAngles = c(), method = "C", stratImg = NULL, nStrat = 5, illu, ... )
img |
SpatRaster. Imagery to correct |
dem |
SpatRaster. Either a digital elevation model as a RasterLayer or a RasterStack/Brick with pre-calculated slope and aspect (see terrain) in which case the layers must be named 'slope' and 'aspect'.
Must have the same dimensions as |
metaData |
Character, ImageMetaData. Either a path to a Landsat meta-data file (MTL) or an ImageMetaData object (see readMeta) |
solarAngles |
Numeric vector containing sun azimuth and sun zenith (in radians and in that order). Not needed if metaData is provided |
method |
Character. One of c("cos", "avgcos", "minnaert", "C", "stat", "illu"). Choosing 'illu' will return only the local illumination map. |
stratImg |
RasterLayer or SpatRaster to define strata, e.g. NDVI. Or the string 'slope' in which case stratification will be on |
nStrat |
Integer. Number of bins or quantiles to stratify by. If a bin has less than 50 samples it will be merged with the next bin. Only relevant if |
illu |
SpatRaster. Optional pre-calculated ilumination map. Run topCor with method="illu" to calculate an ilumination map |
... |
arguments passed to |
For detailed discussion of the various approaches please see Riano et al. (2003).
The minnaert correction can be stratified for different landcover characteristics. If stratImg = 'slope' the analysis is stratified by the slope,
i.e. the slope values are divided into nStrat classes and the correction coefficient k is calculated and applied separately for each slope class.
An alternative could be to stratify by a vegetation index in which case an additional raster layer has to be provided via the stratImg argument.
SpatRaster
Riano et al. (2003) Assessment of different topographic correction in Landsat-TM data for mapping vegetation types. IEEE Transactions on Geoscience and Remote Sensing.
## Load example data metaData <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") metaData <- readMeta(metaData) ## Minnaert correction, solar angles from metaData lsat_minnaert <- topCor(lsat, dem = srtm, metaData = metaData, method = "minnaert") ## C correction, solar angles provided manually lsat_C <- topCor(lsat, dem = srtm, solarAngles = c(1.081533, 0.7023922), method = "C")## Load example data metaData <- system.file("external/landsat/LT52240631988227CUB02_MTL.txt", package="RStoolbox") metaData <- readMeta(metaData) ## Minnaert correction, solar angles from metaData lsat_minnaert <- topCor(lsat, dem = srtm, metaData = metaData, method = "minnaert") ## C correction, solar angles provided manually lsat_C <- topCor(lsat, dem = srtm, solarAngles = c(1.081533, 0.7023922), method = "C")
Unsupervised clustering of SpatRaster data using kmeans clustering
unsuperClass( img, nSamples = 10000, nClasses = 5, nStarts = 25, nIter = 100, norm = FALSE, clusterMap = TRUE, algorithm = "Hartigan-Wong", output = "classes", ... )unsuperClass( img, nSamples = 10000, nClasses = 5, nStarts = 25, nIter = 100, norm = FALSE, clusterMap = TRUE, algorithm = "Hartigan-Wong", output = "classes", ... )
img |
SpatRaster. |
nSamples |
Integer. Number of random samples to draw to fit cluster map. Only relevant if clusterMap = TRUE. |
nClasses |
Integer. Number of classes. |
nStarts |
Integer. Number of random starts for kmeans algorithm. |
nIter |
Integer. Maximal number of iterations allowed. |
norm |
Logical. If |
clusterMap |
Logical. Fit kmeans model to a random subset of the img (see Details). |
algorithm |
Character. kmeans algorithm. One of c("Hartigan-Wong", "Lloyd", "MacQueen") |
output |
Character. Either 'classes' (kmeans class; default) or 'distances' (euclidean distance to each cluster center). |
... |
further arguments to be passed to writeRaster, e.g. filename |
Clustering is done using kmeans. This can be done for all pixels of the image (clusterMap=FALSE), however this can be slow and is
not memory safe. Therefore if you have large raster data (> memory), as is typically the case with remote sensing imagery it is advisable to choose clusterMap=TRUE (the default).
This means that a kmeans cluster model is calculated based on a random subset of pixels (nSamples). Then the distance of *all* pixels to the cluster centers
is calculated in a stepwise fashion using predict. Class assignment is based on minimum euclidean distance to the cluster centers.
The solution of the kmeans algorithm often depends on the initial configuration of class centers which is chosen randomly.
Therefore, kmeans is usually run with multiple random starting configurations in order to find a convergent solution from different starting configurations.
The nStarts argument allows to specify how many random starts are conducted.
Returns an RStoolbox::unsuperClass object, which is a list containing the kmeans model ($model) and the raster map ($map). For output = "classes", $map contains a SpatRaster with discrete classes (kmeans clusters); for output = "distances" $map contains a SpatRaster, with 'nClasses' layers, where each layer maps the euclidean distance to the corresponding class centroid.
## Not run: library(terra) input <- rlogo ## Plot olpar <- par(no.readonly = TRUE) # back-up par par(mfrow=c(1,2)) plotRGB(input) ## Run classification set.seed(25) unC <- unsuperClass(input, nSamples = 100, nClasses = 5, nStarts = 5) unC ## Plots colors <- rainbow(5) plot(unC$map, col = colors, legend = FALSE, axes = FALSE, box = FALSE) legend(1,1, legend = paste0("C",1:5), fill = colors, title = "Classes", horiz = TRUE, bty = "n") ## Return the distance of each pixel to each class centroid unC <- unsuperClass(input, nSamples = 100, nClasses = 3, output = "distances") unC ggR(unC$map, 1:3, geom_raster = TRUE) par(olpar) # reset par ## End(Not run)## Not run: library(terra) input <- rlogo ## Plot olpar <- par(no.readonly = TRUE) # back-up par par(mfrow=c(1,2)) plotRGB(input) ## Run classification set.seed(25) unC <- unsuperClass(input, nSamples = 100, nClasses = 5, nStarts = 5) unC ## Plots colors <- rainbow(5) plot(unC$map, col = colors, legend = FALSE, axes = FALSE, box = FALSE) legend(1,1, legend = paste0("C",1:5), fill = colors, title = "Classes", horiz = TRUE, bty = "n") ## Return the distance of each pixel to each class centroid unC <- unsuperClass(input, nSamples = 100, nClasses = 3, output = "distances") unC ggR(unC$map, 1:3, geom_raster = TRUE) par(olpar) # reset par ## End(Not run)
validate a map from a classification or regression model. This can be useful to update the accuracy assessment after filtering, e.g. for a minimum mapping unit.
validateMap( map, valData, responseCol, nSamplesV = 500, mode = "classification", classMapping = NULL )validateMap( map, valData, responseCol, nSamplesV = 500, mode = "classification", classMapping = NULL )
map |
SpatRaster. The classified map. |
valData |
sf object with validation data (POLYGONs or POINTs). |
responseCol |
Character. Column containing the validation data in attribute table of |
nSamplesV |
Integer. Number of pixels to sample for validation (only applies to polygons). |
mode |
Character. Either 'classification' or 'regression'. |
classMapping |
optional data.frame with columns |
Returns a structured list includng the preformance and confusion-matrix of your then validated input data
library(caret) library(terra) ## Training data poly <- readRDS(system.file("external/trainingPolygons_lsat.rds", package="RStoolbox")) ## Split training data in training and validation set (50%-50%) splitIn <- createDataPartition(poly$class, p = .5)[[1]] train <- poly[splitIn,] val <- poly[-splitIn,] ## Classify (deliberately poorly) sc <- superClass(lsat, trainData = train, responseCol = "class", nSamples = 50, model = "mlc") ## Polish map with majority filter polishedMap <- focal(sc$map, matrix(1,3,3), fun = modal) ## Validation ## Before filtering val0 <- validateMap(sc$map, valData = val, responseCol = "class", classMapping = sc$classMapping) ## After filtering val1 <- validateMap(polishedMap, valData = val, responseCol = "class", classMapping = sc$classMapping)library(caret) library(terra) ## Training data poly <- readRDS(system.file("external/trainingPolygons_lsat.rds", package="RStoolbox")) ## Split training data in training and validation set (50%-50%) splitIn <- createDataPartition(poly$class, p = .5)[[1]] train <- poly[splitIn,] val <- poly[-splitIn,] ## Classify (deliberately poorly) sc <- superClass(lsat, trainData = train, responseCol = "class", nSamples = 50, model = "mlc") ## Polish map with majority filter polishedMap <- focal(sc$map, matrix(1,3,3), fun = modal) ## Validation ## Before filtering val0 <- validateMap(sc$map, valData = val, responseCol = "class", classMapping = sc$classMapping) ## After filtering val1 <- validateMap(polishedMap, valData = val, responseCol = "class", classMapping = sc$classMapping)
Writes binary ENVI spectral library files (sli) with accompanying header (.sli.hdr) files OR ASCII spectral library files in ENVI format.
writeSLI( x, path, wavl.units = "Micrometers", scaleF = 1, mode = "bin", endian = .Platform$endian )writeSLI( x, path, wavl.units = "Micrometers", scaleF = 1, mode = "bin", endian = .Platform$endian )
x |
data.frame with first column containing wavelengths and all other columns containing spectra. |
path |
path to spectral library file to be created. |
wavl.units |
wavelength units. Defaults to Micrometers. Nanometers is another typical option. |
scaleF |
optional reflectance scaling factor. Defaults to 1. |
mode |
character string specifying output file type. Must be one of |
endian |
character. Optional. By default the endian is determined based on the platform, but can be forced manually by setting it to either "little" or "big". |
ENVI spectral libraries with ending .sli are binary arrays with spectra saved in rows.
Does not return anything, write the SLI file directly to your drive for where your specified your path parameter
## Example data sliFile <- system.file("external/vegSpec.sli", package="RStoolbox") sliTmpFile <- paste0(tempdir(),"/vegetationSpectra.sli") ## Read spectral library sli <- readSLI(sliFile) head(sli) plot(sli[,1:2], col = "orange", type = "l") lines(sli[,c(1,3)], col = "green") ## Write to binary spectral library writeSLI(sli, path = sliTmpFile)## Example data sliFile <- system.file("external/vegSpec.sli", package="RStoolbox") sliTmpFile <- paste0(tempdir(),"/vegetationSpectra.sli") ## Read spectral library sli <- readSLI(sliFile) head(sli) plot(sli[,1:2], col = "orange", type = "l") lines(sli[,c(1,3)], col = "green") ## Write to binary spectral library writeSLI(sli, path = sliTmpFile)