Use Case 1:

Apply ML models on datacubes - any region, any time.

The Service:

Run RSVQA on rasdaman datacubes for an interactive selection of times and locations, using english language for conversation.

The basic structure of the WCPS inference query, relying on Sentinel-1 and Sentinel-2, is:

for $S1 in (S1_GRDH_IW), $S2 in (S2_L2A)
let $patch := [ {space-time selection of 120x120 patch} ],
return rsvqa.predict2( $S1[subs2], $S2[subs2], "rsvqa_trained_model.pt", "Are there any airports?")

Use Case 2:

Zero-coding interaction with rasdaman datacubes.

The Service:

AI Agent Intelligence on Datacubes