Radioactive geophysical processing and spatial anomalies analysis based on adaptive weighting fusion and multifractal SVM.

To improve the limitation associated with a single data set and obtain more comprehensive anomaly information, this paper adopts the adaptive weighting fusion and multifractal SVM method for radioactive exploration data processing. Firstly, the adaptive weighting fusion method was used to fuse various radioactive exploration data from the field, including ground gamma-ray spectrometry, thermoluminescence, and Po activity. Then, we compared and analyzed it against the anomalous results obtained from each method. It is observed that the adaptive fusion yielded a more comprehensive dataset, which effectively reflected the anomalous geophysical characteristics of radioactivity, thereby eliminating the need to map each method separately. Finally, the concentration-area (C-A) multifractal model was used to classify the supervised learning labels, 70% of the sampling point data were selected as the training data, and the SVM was executed to predict the favorable prospecting target area, and the prediction accuracy reached 82.7%. At the same time, a model has been established to analyze the spatial distribution characteristics of the abnormal radioactive ore body and infer the deep favorable ore-forming target area.