Parcel-level features from S1+S2 time series

Anyone doing per-parcel aggregates that fuse Sentinel‑1 VV/VH backscatter with Sentinel‑2 NDVI in QGIS or PostGIS? I’m pulling 2019–2024 L2A via STAC and GRD via ASF, using s2cloudless and RTC, but I’m stuck on preserving acquisition timestamps and incidence angle through zonal stats so the final joined table doesn’t explode into one row per pixel-date…

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍‌⁠‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‌​⁠‌‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​‍‌‌‍‍‍​⁠‌‍‌‌​⁠‌‍⁠​‌‍‌‌‌‌‍‌​⁠‍​​⁠‍‌‌‌‍‍‌⁠‌‌‌⁠​‍‌​‌⁠‌‍‌‍​⁠‌​‌⁠‌‍​‍​‍‌⁠⁠‌​​

In PostGIS, ST_SummaryStatsAgg(., parcel.geom) grouped by acq_ts+incidence_angle; lateral nearest-date S1↔S2 join avoids the ‘pixel‑date’ explosion. Okay to bin angle?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍‌⁠‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‌‍​⁠​‍​⁠‌‍​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​‍‍‌⁠‌‌‌‌‍‍‌​​‍‌​‌‍‌‌‌‍‌‍​‌‌‍‌⁠‌‍‍​‌‌‌‌‌‌‍‍‌‍‌​‌‌​‍​⁠‌⁠‌‍‍⁠​⁠‌‍​‍​‍‌⁠⁠‌​