Descartes Labs, an emerging specialist in geospatial analysis has just launched GeoVisual Search. It is a visual search technology based on machine learning whereby users can search for similar elements available around the world, similar to those in the area they determine.
From the company they emphasize that the identification of the elements is similar to the one that Facebook uses through the images. It is being able to recognize those images visually similar.
It should be noted that GeoVisual Search is not a product as such but rather a technological demonstration of the extent to which technology is able to recognize patterns through satellite images.
For their development they have had to go through two important stages. The first being the training of the neural network created so that it is able to recognize patterns similar to that indicated through other satellite images, and the second the possibility to perform the searches in Real time. It aims to become a necessary infrastructure that makes it possible to perform quick searches and show only the best results.
GeoVisual Search uses three maps from three different image sources from the United States, China and the world. GeoVisual Search is available for free, but there is a possibility that at some point it offers payment features. It may be aiming at the moment to launch a paid version for analysts that incorporates additional functions.