I am processing a S1 SLC image as follows: Read>Apply orbit file>S1 TOPS Split>Radiometric Calibration (complex)>S1 TOPS Deburst>Polarimetric Speckle filter>Terrain correction (Range Doppler)
I got a scene with thick bands, one with no data and several others with different colors. Apparently they are introduced or at least highlighted during the S1 TOPS Deburst process. Tried to replicate the process with the full scene, however it crashes due to insufficient RAM.
I am using the data from the IW2 subswath and only 3 bursts (small area of interest)
I am not clear what is causing this problem and also how to correct it, the problem is that when I derive the polarimetric products those strips are quite evident.
Scene: S1A_IW_SLC__1SDV_20211112T192510_20211112T192538_040542_04CEEB_7AAB
Update: I have replicated the same procedure in an older laptop (8 times less RAM, slower processor, half of video memory, no SSD) and for an unknown reason, the procedure works and produces no error. I am still investigating for possible causes, yet I have no clue about this issue. I used the same data, same parameters, yet the newer laptop produces errors on the final product (even with the new SNAP 9 installed). Has someone experienced something similar?
Is the input product OK in the newer laptop? Any missing bursts, lines etc.?
I used the same SLC file in both cases, indeed I copied the same file to the older laptop and run the same steps one by one with exactly the same parameters yet the newer laptop generates products with strips (as showed in the pictures) while the older does not, even though it takes more time to process.
It is almost impossible to comment since the issue cannot be reproduced. Your graph is complex which reduces performance compared with splitting the the graph and saving intermediate products. Splitting would allow you to see which step causes the processing to fail. I would imagine the steps work one-by-one?
I have modified the process, and first I split the data, before applying the orbit file (I only need 3 bursts from IW2), the remaining steps remain the same. Then under these circumstances, all the steps work, indeed aside from some eventual RAM problems, the steps perform normally. The only difference is that so far, I have products with errors (bands) when using a newer laptop and correctly processed products in an older laptop. (Similar processors, even brands)
