Correction to: Separation of the Total Magnetic Anomalies into Induced and Remanent Magnetization Based on Deep Learning

The presence of remanent magnetization brings uncertainty to the processing and interpretation of magnetic data. Therefore, separating the contributions of the remanent magnetization from the total magnetic data is always the research hotspot. In the literature, numerous methods have been introduced to handle this issue. However, most of the existing methods are complex to calculate, have strict requirements on magnetic sources, and need prior information. In this study, a new method for automatically separating the total magnetic anomalies into the components due to induced and remanent magnetization based on deep learning has been presented. The presented method designs an end-to-end network structure based on the U-Net network structure and then performs continuous training and parameter optimization to determine the optimal network structure. Afterward, the presented method is tested on synthetic examples and actual magnetic data in Yeshan Region (Eastern China). The results demonstrate that the presented method can separate anomalies by induced and remanent magnetization.