polygon outside bounds of sentinel-1 scene it is recorded for

[rbavery]

https://cerulean-ui-5qkjkyomta-uc.a.run.app/?zoom=8.98925&lat=40.49821200535761&lng=17.070769165779147&startDate=2023-01-01&endDate=2023-01-02&slick=372&showImagery=1&machineConfidence=0_0.4&imageGamma=2

[rbavery]

Either this is a georeferencing problem, and the slick should be elsewhere, or a metadata problem, and the slick should be recorded against a different S1 scene ID and shouldn't be rendered in the above UI for this particular scene.

[rbavery]

I manually went to ASF Vertex and found the coordinates for the slick above Barletta that shouldn't be associated with the scene in a different location.

https://search.asf.alaska.edu/#/?polygon=POLYGON((16.1715%2041.2509,17.2581%2040.928,16.5854%2041.468,16.1715%2041.468,16.1715%2041.2509))&zoom=8.850&center=17.300,40.691&start=2023-01-01T08:00:00Z&end=2023-01-02T07:59:59Z&resultsLoaded=true&granule=S1A_IW_GRDH_1SDV_20230101T164937_20230101T165002_046593_059574_E870-GRD_HD

That slick doesn't show up in prediction or the base image when I run local inference or the predict endpoint, only with orchestrate. I tested the scene that does show up under that slick, S1A_IW_GRDH_1SDV_20230101T164937_20230101T165002_046593_059574_E870

This just tells us that this isn't a metadata issue for an actual false positive or true positive. I think this still feels like a georeferencing problem, something buggy.