Phase 6 · Look for results agreement¶
Now hunt for zones where your layers light up in the same place but treat spatial overlap as a starting point rather than the confirmation itself.
Be careful what you count as agreement. Layers from the same provider run on the same pixel of the same image, so when SFF, SAM and GSI agree, that's three methods agreeing about one observation rather than three observations. Useful, and the easiest thing to spot on the map, but it can't be your final confirmation. Different providers are another matter, because EMIT at 60 m, Sentinel-2 at 10 to 20 m and pre-2008 ASTER are different footprints from different years, so check the provider and date before you call it agreement.
An overlap tells you where to look, not that you're right. Every one of these methods can return a false positive, so three of them overlapping can be three false positives. What settles it is the curve. Agreement that isn't spectral at all is Phase 8. It is a good place to recheck if your masks are right.
Ignacy can do part of this: ask where three or more minerals occur together and he'll find the hotspots and save them as a layer if you confirm. Use that to decide which minerals to look at first, not as results to believe.
If nothing stands out¶
That's a result, and worth recording properly: nothing above detection, in the fraction of your area that survived masking, for the minerals you loaded. It isn't evidence that the alteration isn't there, because a mineral can sit below what a mixed 60 m pixel shows. And check Phase 8 first, since magnetics sees under cover.
Done when
You have candidates. Most will be zones where two or more results overlap; a single strong result also qualifies, so carry those forward too. The curve check is what decides, not the count.