Is the quality of terrain correction (TC) using Range-Doppler dependent on subset size?
I ask because I get slightly different pixel values after TC on subsets of different size, but with a common area of interest to me.
In the example below, you can see the size of the subimages and that the smaller subset for a specific pixel yields a value of -21.1 dB, while the same pixel in the larger subset yields -20.6 dB after TC.
Before TC, both subsets have exactly the same pixel values for a given pixel.
Also, I already checked that this isn’t a simple case of the image being shifted one pixel over in one case.
Can someone help me understand what is going on?