Seismic surface wave analysis has been widely adopted for characterizing shallow subsurface shear wave velocity (Vs) structures. The application of surface-wave analysis is now expanding from seismic imaging to time-lapse monitoring of the dynamic changes in subsurface properties. However, conventional workflows of surface-wave data processing and inversion face challenges in labor cost, subjectivity, and computational efficiency. The rapid rise of deep learning offers a transformative new paradigm in surface-wave analysis. This review examines the integration of deep learning in the key steps of surface-wave analysis: data acquisition optimization, automated dispersion curve extraction, intelligent Vs inversion, and time-lapse analysis. We analyze the emerging deep learning applications and challenges in surface-wave analysis and provide a perspective on future studies toward fully automated, intelligent, and generalized workflows for subsurface imaging and monitoring.
Mi et al. (Sun,) studied this question.