I have around 100 sentinel-1 SLC images. I want to compute coherence between the successive images (1->2, 2->3… and so on). For this, the conventional method is to read images 1 and 2, select swath, apply orbit files, coregister them using the back-geocoding operator, apply ESD and then compute the coherence and the same procedure is also with images 2 and 3, and so on.
I want to discuss another approach here
- Read image 1 and coregister all other images (2,3,4…100) with respect to 1.
- Now, for images 1 and 2, we can compute the coherence as the above-mentioned method.
- For image 2 slv and image 3 slv, we stack them together considering them already coregistered and compute coherence.
SNAP does not allow to compute the coherence using this method.
I want to know why I am not able to estimate coherence with the later method.
What should be the optimal method to perform successive time series analysis?