Seismic structure interpretation method and device based on multi-source multi-task deep learning

The invention relates to the technical field of seismic data interpretation, in particular to a seismic structure interpretation method and device based on multi-source multi-task deep learning, a processor and a storage medium. The seismic structure interpretation method comprises the following steps: acquiring original seismic wave data; processing the original seismic wave data by adopting a plurality of seismic data interpretation methods to form a multi-modal data volume, and forming a multi-source input data set by the multi-modal data volume and the original seismic wave data; carrying out fault labeling on the original seismic wave data and calculating a relative geological time body to form a fault-relative geological time body label data set; combining the fault-relative geological time body label data set with the multi-source input data set to form a training sample; and training a multi-source multi-task seismic structure interpretation network model by using the training sample. According to the