Interpretation of GLCM (Texture) feature analysis for Water bodies

Microwave Toolbox
1

As we know, water backscattering values are much lower than that of non water in radarsat2 (water appeared in black on SAR).

But, when it comes to GLCM analysis, the interpretation becomes different.

Water bodies appear in White for “Homogeneity” and “Energy” images, while in black for “Dissimilarity” and “Entropy”! This indicates that the texture intensity values of water for “Homogeneity” and “Energy” is higher than that of non-water!!!

I am under impression that the GLCM Intensity values has nothing to do with the low values of sigma0 (for water) obtained from TC images! Please correct me if I am wrong!

Thanks!

1 Like
2

Though I’m not sure that I quite understand your water bodies because it is a big term, however, Please take a look at this might be it is helpful

1 Like
3

Seems quite logical to me. If your water ar areas are consistently black, they have a homogenous texture, consequently high Homogeneity values.
If you had waves in it, the values would decrease.

This doesn’t mean that only water has high Homogeneity values, but for inland water bodies chances are considerably higher than over agricultural or urban land, for example.
But this strongly depends on the spatial resolution and wavelength of your data. Other sensors or modes can respond totally different.

1 Like
4

For “homogeneity” and “Energy”, water bodies are “White” with textural intensity values > non-water regions.

Conversely, for “Entropy” and “Dissimilarity”, water bodies are “Black” with textural intensity values << non-water regions.

Thanks