I am new to SNAP–StaMPS processing. First, I am planning to complete the preprocessing steps in SNAP for a stack of 68 images (for 3 burst with both ASC and DSC)) by following these steps: Back-Geocoding, ESD(optionel), Deburst, Subset, Interferogram, TopoPhaseRemove, and Export to StaMPS.
However, in ascending images processing, I encountered a “Java heap space” and '' cannot construct databuffer'' errors during the Back-Geocoding. And then, in descending images processing, Back-Geocoding is ok (no problem). But, ESD not worked.(it takes hours but no results)
I assume this is related to the limitations of my computer’s hardware.
As a workaround, I am considering splitting the dataset into two separate stacks (34+34 etc), processing them independently, and then combining the two StaMPS export outputs at the end.
If this approach is feasible, should I copy only the diff0 and rslc folders for both stacks, and keep the DEM and geo files from only one of the stacks? Below are the SNAP performance metrics along with my computer specifications. (CPU: AMD Ryzen 7 5700U with Radeon G, CPU:16, Memory: 5420MiB / 38918MiB)