This is possible, with caveats (full disclosure, I am a Descartes Labs employee).
Geovisual Search is not part of the Descartes Labs Platform and any results you collect via the site are subject to the Descartes Labs website Terms of Service.
Geovisual Search is an index of visual similarity - not a model trained to find specific objects or patterns. For the complete methodology, see this publication.
That said - each search of the index returns 1000 results ordered by similarity within the index. This is served by a JSON API at https://search.descarteslabs.com/search?key={key}&{layer}&limit={limit}
where key
is a UTM DLTile identifier, layer
is an identifier for the dataset indexed, and limit
is the number of results to return.
If all you need is a the list of tiles returned on the left - you can get that JSON response by looking in your browser's developer console to find the request made by the API. The result will be a list of GeoJSON features, with a distance
property that corresponds to the visual similarity relative to the key tile (lower is more similar). This will include 1,000 results by default:
"features":[
{"type":"Feature",
"properties":{"distance":0,"key":"64:32:1.0:13:-3650:56342"},
"geometry":{"coordinates":[[[-107.48853126199128,32.56553667302686],[-107.48720082953817,32.56553667302686],[-107.48720082953817,32.56671724913455],[-107.48853126199128,32.56671724913455],[-107.48853126199128,32.56553667302686]]],"type":"Polygon"}},{"type":"Feature","properties":{"distance":0,"key":"64:32:1.0:13:-3649:56353"},"geometry":{"coordinates":[[[-107.48802537553277,32.571894941909456],[-107.48669484774975,32.571894941909456],[-107.48669484774975,32.57307551646963],[-107.48802537553277,32.57307551646963],[-107.48802537553277,32.571894941909456]]],"type":"Polygon"}},{"type":"Feature","properties":{"distance":3,"key":"64:32:1.0:13:-3643:56348"},"geometry":{"coordinates":[[[-107.4838586894654,32.56909183457887],[-107.48252814297841,32.56909183457887],[-107.48252814297841,32.57027236609133],[-107.4838586894654,32.57027236609133],[-107.4838586894654,32.56909183457887]]],"type":"Polygon"}},{"type":"Feature","properties":{"distance":3,"key":"64:32:1.0:13:-3645:56345"},"geometry":{"coordinates":[[[-107.48517328714188,32.56733449119291],[-107.48384278325798,32.56733449119291],[-107.48384278325798,32.568515034838654],[-107.48517328714188,32.568515034838654],[-107.48517328714188,32.56733449119291]]],"type":"Polygon"}},{"type":"Feature","properties":{"distance":3,"key":"64:32:1.0:13:-3651:56345"},"geometry":{"coordinates":[[[-107.48926022103753,32.56725357076538],[-107.48792977482196,32.56725357076538],[-107.48792977482196,32.56843415507681],[-107.48926022103753,32.56843415507681],[-107.48926022103753,32.56725357076538]]],"type":"Polygon"}},{"type":"Feature","properties":{"distance":3,"key":"64:32:1.0:13:-3651:56347"},"geometry":{"coordinates":[[[-107.48929210170445,32.568407165664674],[-107.48796163990536,32.568407165664674],[-107.48796163990536,32.56958775092752],[-107.48929210170445,32.56958775092752],[-107.48929210170445,32.568407165664674]]],"type":"Polygon"}},{"type":"Feature","properties":{"distance":4,"key":"64:32:1.0:13:-3650:56357"},"geometry":{"coordinates":[[[-107.48877032562693,32.574188636091414],[-107.4874397762816,32.574188636091414],[-107.4874397762816,32.57536921933257],[-107.48877032562693,32.57536921933257],[-107.48877032562693,32.574188636091414]]],"type":"Polygon"}},
...