// the find
johntruckenbrodt/pyroSAR
framework for large-scale SAR satellite data processing
pyroSAR is a Python framework for automating SAR satellite data processing pipelines, wrapping SNAP and GAMMA Remote Sensing tools with a common interface. It handles metadata, archive management, and batch processing across missions like Sentinel-1, ERS, and ALOS. The target user is a remote sensing researcher or geospatial engineer who needs to process large SAR datasets without hand-wiring shell calls to proprietary toolchains.
SQLite/SpatiaLite archive for tracking processed scenes is a practical design — you can query what you have by footprint, date, or sensor without reinventing a catalog. The SNAP XML graph abstraction lets you compose processing chains programmatically instead of clicking through the SNAP GUI. Support for multiple missions and drivers (ERS, ALOS, Sentinel-1, ENVISAT) from a single API is genuinely useful for multi-temporal or multi-sensor studies. Conda-forge packaging means the GDAL/GEOS dependency hell is mostly handled for you.
The hard dependency on SNAP or GAMMA means this is effectively useless without installing external proprietary or semi-proprietary software first — no fallback, no lightweight path for exploration. The GAMMA module wraps a commercial product, so a large chunk of the codebase is only useful if you have a license. Documentation is readthedocs RST that reads like an academic paper, not a quick-start guide; onboarding friction is high. Cloud-native execution (S3, GCS, STAC APIs) is absent — everything assumes local disk, which is a real problem for anyone processing Sentinel-1 at scale today.