I am attempting to processing (apply orbit file, thermal noise removal, radiometric calibration, speckle subduction, terrain correction, linear to dB conversion) an entire S1 IW GRDH image without subsetting in SNAP but it seems too slow. I am following the graph approach where I have begun with a single image for now but the painful part is the SNAP window that doesn’t show any progress bar (although I can see the destination folder getting heavier). I would like to know if there is a more efficient way to process the full swath since I would ideally want to combine several full swaths for my investigation. Furthermore, I would like to know the possibility of accelerating the preprocessing part by employing GPUs.
In general the pace of the processing steps depend on your machine properties the processor and the RAM, Also put in mind that you’re processing the whole S1 IW GRDH,
I think no, please take a look at this thread, Source of the thread
Thanks for the clarification.
I would like to know if there is a comparitive analysis on the performance improvement of SNAP wrt the machine properties and RAM. For ex, how much is the processing time reduction if one moves from an i5 to an i7 and to i9.
In my personal experience with i7, it works fine, however resampling sigma0 GLCM, takes 9 hours, the RAM is 16, I do believe it should be increased to raise the processing pace.