In previous articles Google Earth Engine (GEE) as a public supercomputer and Google Earth Engine (GEE) as a public large geodata catalog, we got acquainted with the ways of convenient and quick access to the catalog of space images and their processing. Now we can search for drinking water, various minerals, and in general, a lot. We can also arm ourselves with machine learning (ML) methods and make our own treasure map - a forecast for finding gold deposits anywhere in the world. As always, check out the code and raw data (synthetic, of course, because real data is literally worth its weight in gold!) On GitHub: AU Prediction (ML)
On the island of West Sumbawa , using the constructed classifier, the predicted gold-bearing areas have been identified.
() () . : , , , ( , ) . , . , . , , , , . , , . , , . , , , , ( , , ).
( , ) , . , . , , (, , ) Google Earth Engine.
( ) , Google Earth Engine (GEE) (Landsat 8, Sentinel 2, ...) .
, , - . , — . (--) : Compare Spectrograms of Hyperspectral and Multispectral Satellite Missions:
, , . (, ) ( , ) , .
, . , «», , «», . — , , , . , , .
samples_synthetic_3class.geojson , . , .
, GEE_export.ipynb Google Earth Engine, West Sumbawa AU 3-Class Prediction Synthetic.ipynb . GEE, , , . , .
, — , . , (surface reflectance) Landsat-8 TOA (top of atmosphere), SR TOA . Sentinel-2 .
( ) . , , , . , , , .
P.S. !