Seismic data reconstruction method based on mask guidance attention convolutional neural network

The invention discloses a seismic data reconstruction method based on a mask guidance attention convolutional neural network. According to the method, a mask updating mechanism and attention are combined, a context feature relationship is established through a mask guiding attention mechanism, and a relationship between local features is established by calculating cross relevancy between different local features, so that global information can be obtained by the local features; and the learning ability of local features is enhanced by introducing more context information. Therefore, mask guidance attention can establish a long-distance feature relationship without significantly increasing the depth of the network, and local features and global features are effectively fused. The iterative updating operation of the mask guides the reconstruction process of the network, so that each local feature serves as input to obtain more global information, and the problem of insufficient utilization of the global feature