A Deep Learning Approach to Automate and Enhance Seismic First Break Picking

Summary A novel methodology for first break picking (FBP) using deep learning algorithms is introduced in this paper. The goal of this study is to automate the FBP process by applying a CNN model trained on synthetic seismic data that comprehensively mimics and describes the target real data. The proposed approach is evaluated and validated using publicly available real seismic data. A detailed description of the obtained results is provided. The results of this study open up promising potential for improving the accuracy of FBP outcomes and significantly reducing processing time.