Delineation and Analysis of Regional Geochemical Anomaly Using the Object-Oriented Paradigm and Deep Graph Learning—A Case Study in Southeastern Inner Mongolia, North China

This research describes an advanced workflow of object-based geochemical graph learning approach, termed OGE, which includes five key steps: (1) conduct the mean removal operation on multi-elemental data and then normalize them; (2) gridding multiresolution segmentation; (3) calculate Moran’s I value construct topology graph; (4) unsupervised deep learning; (5) within-object statistical analysis. The final product OGE is anomaly score map. performance was demonstrated by a case study involving eighteen ore-forming elements (Cu, Pb, Zn, W, Sn, Mo, F, Au, Fe2O3, etc.) in stream sediment samples Bayantala-Mingantu district, North China. results showed that analysis performed at lower levels scale greatly improved quality recognition: more than 80% known ore spots, no matter what their scales mineral species, were predicted less 45% area, most spots falling outside delineated anomalous regions occur nearby them. can extract both spatial features compositional relationships variables collected irregularly distributed centroids shaped image objects, it outperforms other convolutional autoencoder models such as GAUGE detection.