| Title: | Calculate Textures from Grey-Level Co-Occurrence Matrices (GLCMs) |
|---|---|
| Description: | Enables calculation of image textures (Haralick 1973) <doi:10.1109/TSMC.1973.4309314> from grey-level co-occurrence matrices (GLCMs). Supports processing images that cannot fit in memory. |
| Authors: | Alex Zvoleff [aut, cre] |
| Maintainer: | Alex Zvoleff <[email protected]> |
| License: | GPL (>= 3) |
| Version: | 1.6.5 |
| Built: | 2024-11-18 06:52:36 UTC |
| Source: | CRAN |
Enables calculation of image textures derived from grey-level co-occurrence matrices (GLCMs) in R. The texture calculation is coded in C++ to optimize computation time.
Alex Zvoleff, [email protected]
This function is called by the glcm function. It is
not intended to be used directly.
calc_texture(d, n_grey, window_dims, shift, statistics, na_opt, na_val)calc_texture(d, n_grey, window_dims, shift, statistics, na_opt, na_val)
d |
a matrix containing the pixels to be used in the texture calculation |
n_grey |
number of grey levels to use in texture calculation |
window_dims |
2 element list with row and column dimensions of the texture window |
shift |
a matrix where each row gives an (x, y) shift to use when computing co-occurrency matrices. Textures will be calculated for each shift, and the average over all shifts will be returned. |
statistics |
a list of strings naming the texture statistics to calculate |
na_opt |
one of "ignore", "center", or "any" |
na_val |
what value to use to fill missing values on edges or where necessary due to chosen na_opt value |
a list of length equal to the length of the statistics input
parameter, containing the selected textures measures
# Calculate GLCM textures on a matrix using low-level calc_texture function d <- matrix(seq(1:25), nrow=5, ncol=5, byrow=TRUE) calc_texture(d, n_grey=25, window_dims=c(3,3), shift=matrix(c(1,1), nrow=1), statistics=c('variance'), na_opt="any", na_val=NA)# Calculate GLCM textures on a matrix using low-level calc_texture function d <- matrix(seq(1:25), nrow=5, ncol=5, byrow=TRUE) calc_texture(d, n_grey=25, window_dims=c(3,3), shift=matrix(c(1,1), nrow=1), statistics=c('variance'), na_opt="any", na_val=NA)
This is the output from running a "co-occurrence measures" calculation to
calculate GLCM textures in EXELIS ENVI from the test_raster included
in the glcm package. The following settings were used: window size
3x3; co-occurrence shift 1 row (y in ENVI), 1 column (x in ENVI); greyscale
textures to compute: mean, variance, homogeneity, contrast, dissimilarity,
entropy, second moment, correlation.
expected_textures_5x7_2x3
expected_textures_5x3_n1xn2
This is the output from running a "co-occurrence measures" calculation to
calculate GLCM textures in EXELIS ENVI from the test_raster included
in the glcm package. The following settings were used: window size
5x3; co-occurrence shift -1 row (y in ENVI), -2 columns (x in ENVI);
greyscale quantization levels 32; textures to compute: mean, variance,
homogeneity, contrast, dissimilarity, entropy, second moment, correlation.
expected_textures_3x3_1x1
expected_textures_5x7_2x3
This is the output from running a "co-occurrence measures" calculation to
calculate GLCM textures in EXELIS ENVI from the test_raster included
in the glcm package. The following settings were used: window size
5x7; co-occurrence shift 2 rows (y in ENVI), 3 columns (x in ENVI);
greyscale textures to compute: mean, variance, homogeneity, contrast,
dissimilarity, entropy, second moment, correlation.
expected_textures_3x3_1x1
expected_textures_5x3_n1xn2
This function supports calculating texture statistics derived from grey-level co-occurrence matrices (GLCMs). The default textures are calculated using a 45 degree shift. See Details for other options.
glcm(x, n_grey = 32, window = c(3, 3), shift = c(1, 1), statistics = c("mean", "variance", "homogeneity", "contrast", "dissimilarity", "entropy", "second_moment", "correlation"), min_x=NULL, max_x=NULL, na_opt="any", na_val=NA, scale_factor=1, asinteger=FALSE)glcm(x, n_grey = 32, window = c(3, 3), shift = c(1, 1), statistics = c("mean", "variance", "homogeneity", "contrast", "dissimilarity", "entropy", "second_moment", "correlation"), min_x=NULL, max_x=NULL, na_opt="any", na_val=NA, scale_factor=1, asinteger=FALSE)
x |
a |
n_grey |
number of grey levels to use in texture calculation |
window |
the window size to consider for texture calculation as a two element integer vector (number of rows, number of columns) |
shift |
a list or matrix specifying the shift to use. See Details. |
statistics |
A list of GLCM texture measures to calculate (see Details). |
min_x |
minimum value of input |
max_x |
maximum value of input |
na_opt |
How to handle NA values in |
na_val |
the value to use to fill NA values on edges of |
scale_factor |
factor by which to multiply results. Useful if rounding
results to integers (see |
asinteger |
whether to round results to nearest integer. Can be used to
save space by saving results as, for example, an 'INT2S' |
The statistics parameter should be a list, and can include any (one
or more) of the following: 'mean', 'mean_ENVI', 'variance', 'variance_ENVI',
'homogeneity', 'contrast', 'dissimilarity', 'entropy', 'second_moment',
and/or 'correlation'. By default all of the statistics except for
"mean_ENVI" and "variance_ENVI" will be returned .
shift can be one of:
a two element integer vector giving the shift (Q in Gonzalez and Woods, 2008), as (number of rows, number of columns).
a list of integer vectors of length 2 specifying multiple (row, col)
shifts over which to calculate the GLCM textures. For example:
shift=list(c(1,1), c(-1,-1))
a matrix with two columns specifying, in rows, multiple (row, col)
shifts over which to calculate the GLCM textures. For example:
shift=matrix(c(1,1,-1,-1), byrow=TRUE, ncol=2)
If multiple shifts are supplied, glcm will calculate each texture
statistic using all the specified shifts, and return the mean value of the
texture for each pixel. To calculate GLCM textures over "all directions" (in
the terminology of commonly used remote sensing software), use:
shift=list(c(0,1), c(1,1), c(1,0), c(1,-1)). This will calculate the
average GLCM texture using shifts of 0 degrees, 45 degrees,
90 degrees, and 135 degrees.
A RasterLayer or RasterStack with the requested GLCM
texture measures.
Lu, D., and M. Batistella. 2005. Exploring TM image texture and its relationships with biomass estimation in Rondônia, Brazilian Amazon. Acta Amazonica 35:249–257.
Gonzalez, R. C. 2008. Digital image processing. 3rd ed. Prentice Hall, Upper Saddle River, N.J, pages 830–836.
Haralick, R. M., K. Shanmugam, and I. Dinstein. 1973. Textural features for image classification. IEEE Transactions on Systems, Man and Cybernetics SMC-3:610–621.
Pratt, W. K. 2007. Digital image processing: PIKS Scientific inside. 4th ed. Wiley-Interscience, Hoboken, N.J pages 540–541, 563–566.
# Calculate GLCM textures on a matrix d <- matrix(seq(1:25), nrow=5, ncol=5, byrow=TRUE) # Calculate using default 90 degree shift glcm(d, statistics=c('variance')) # Calculate over all directions glcm(d, shift=list(c(0,1), c(1,1), c(1,0), c(1,-1)), statistics=c('variance')) ## Not run: # Calculate GLCM textures on a raster require(raster) # Calculate using default 90 degree shift textures_shift1 <- glcm(raster(L5TSR_1986, layer=1)) plot(textures_shift1) # Calculate over all directions textures_all_dir <- glcm(raster(L5TSR_1986, layer=1), shift=list(c(0,1), c(1,1), c(1,0), c(1,-1))) plot(textures_all_dir) ## End(Not run)# Calculate GLCM textures on a matrix d <- matrix(seq(1:25), nrow=5, ncol=5, byrow=TRUE) # Calculate using default 90 degree shift glcm(d, statistics=c('variance')) # Calculate over all directions glcm(d, shift=list(c(0,1), c(1,1), c(1,0), c(1,-1)), statistics=c('variance')) ## Not run: # Calculate GLCM textures on a raster require(raster) # Calculate using default 90 degree shift textures_shift1 <- glcm(raster(L5TSR_1986, layer=1)) plot(textures_shift1) # Calculate over all directions textures_all_dir <- glcm(raster(L5TSR_1986, layer=1), shift=list(c(0,1), c(1,1), c(1,0), c(1,-1))) plot(textures_all_dir) ## End(Not run)
Portion of Landsat 5 Surface Reflectance image from the Landsat Climate Data Record archive. This subset of the image includes only bands 1-4.
Used in testing the output from the GLCM texture statistics C++ code.
# The image was generated with the following code: require(raster) set.seed(0) test_matrix <- matrix(runif(100)*32, nrow=10) test_raster <- raster(test_matrix, crs='+init=epsg:4326') test_raster <- cut(test_raster, seq(0, 32))# The image was generated with the following code: require(raster) set.seed(0) test_matrix <- matrix(runif(100)*32, nrow=10) test_raster <- raster(test_matrix, crs='+init=epsg:4326') test_raster <- cut(test_raster, seq(0, 32))