Sentinel 1 classification to detect deforestation

I am doing a job with sentinel 1 and I work with sentinel images 1 GRD IW and polarization VV + VH.
I need to classify the image and be able to detect deforestation.
there is some way to combine VV polarization with VH and then classify or alone with a band can detect deforestation.

Do you work with the snap-toolbox? There is a whole module with classification tools, where you can combine different parts of products.

I am working with the SNAP version 5.0 program.
To do the classification I take one of the two polarizations and with that I can work or is there a way to combine the two polarizations?

You can combine them as you want - for example calculate VV/VH-ratio. Or feed both polarizations to the classifier.

1 Like

To combine the two bands, you can use Band Maths operator. Once the new band is available, you can perform all kinds of classifications with it.


Rather than using the dual pol bands, have a look at the difference in coherence between two pairs of images. However, this may only work with clear cuts. Low coherence with trees can become high coherence after the trees have been cleared and no further disruptive activity exists between images.