Seismic imaging enhancement of sparse ocean-bottom node data using deep learning

Ocean bottom node (OBN) surveys are a type of geophysical survey that utilizes sensors placed on the seafloor to collect seismic data. These surveys provide high-quality four-component (4C) data, which include converted shear waves, and thus, allows us to utilize the elastic assumption in imaging and inversion. However, OBN surveys can be expensive due to the difficulty in deploying the necessary sensors on the seafloor, resulting in often sparse node spacing to reduce acquisition time and cost. The sparse data result in poor illumination and imaging challenges. In order to address these issues in the context of 4C elastic imaging, we present a deep learning-based method using a multi-scale convolution neural network (Ms-CNN) to improve the imaging quality of OBN surveys with sparse data acquisition. The Ms-CNN is trained in a supervised fashion to map from sparse data images of PP and PS sections produced by 4C Gaussian beam migration to the equivalent dense data images, allowing for the direct processing of sparse data to improve the imaging quality. The effectiveness of the proposed method is demonstrated on synthetic and field data, enhancing the images to improve event continuity and reduce migration noise from sparse OBN acquisitions.