Graph learning-based method and system for identifying metallogenic anomalies in shallowly covered areas

Abstract

The invention discloses a graph learning-based method and system for identifying metallogenic anomalies in shallowly covered areas. The method includes: acquiring coordinates of geochemical sampling points in the shallowly covered area, obtaining geochemical data and geophysical response values corresponding to the coordinates, and performing depth correction on the geochemical data according to the thickness of the overburden to obtain a corrected geochemical dataset; conducting spatial interpolation on the corrected geochemical dataset to construct a continuous data grid reflecting the spatial distribution pattern of elements, so as to obtain high-resolution spatial distribution data; extracting local anomaly features based on the spatial distribution data to generate an initial anomaly signal distribution map, and denoising the initial anomaly signals to obtain a denoised anomaly signal dataset. (Fig. 1)