Detection of the multivariate geochemical anomalies associated with mineralization using a deep convolutional neural network and a pixel-pair feature method

摘要: Abstract Machine learning (ML) algorithms are widely applied in various fields owing to their strong ability abstract high-level features from a large number of training samples. However, few supervised ML have been geochemical prospecting and mineral exploration because mineralization is rare geological event that leads an insufficient Generating samples crucial for the application exploration. In this study, novel anomaly detection framework combined with pixel-pair feature (PPF) method deep convolutional neural network (CNN) was employed identify multivariate anomalies associated mineralization. First, PPF generate sufficient by recombining pixel pairs labeled Then, multilayer CNN framework, which consists 13 layers, average pooling layer, fully connected trained these recognition. The testing procedure based on fact neighboring pixels belong same class high probability. dual-window detector detect related Fe polymetallic southwest Fujian province China. identified exhibited close spatial correlation known deposits, validates potential proposed method. Therefore, developed study can enhance