PROCESSING TECHNIQUES FOR GENERATING, TRACKING, AND VISUALIZING ENVIRONMENTAL INSIGHTS

Abstract: various embodiments of the present disclosure provide environmental modeling techniques for generating and presenting diverse sets of geographic insights for a geographic region. the geographic insights include landcover insights derived from hyperspectral data using machine learning techniques. the landcover insights include landcover predictions that correspond to geographic polygons and the machine learning techniques leverage models that are trained using hyperspectral image frames that correspond to labeled geographic polygons. the geographic insights include carbon stock insights derived from multiple static and dynamic rasters for a geographic region. the rasters include carbon zone, a landcover, a burned area, a soil carbon stock rasters, and the like, that may be aggregated according to a prediction scheme to generate time-dependent insights tailored to a particular point in time. these insights and others may be presented through specially designed overlays for a user interface.

Inventor(s): Leonardo Arturo Prieto Williamson, Pablo Francisco Gonzalez Ordoñez, Alexandra Sthephania Castañeda Sepúlveda

CPC Classification: G06V20/194 (IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING)

Search for rejections for patent application number 20250166375