Anomalous Areas Detection in Rocks Using Time-Difference Adjoint Tomography

Detecting anomalous areas (such as caves, faults, and weathered layers) in rocks is essential for the safety of facilities and personnel in subsurface engineering. Seismic tomography has been proved to be an effective exploration technology in engineering geophysics. However, the complexity, anisotropy, and uncertainty in rock environments pose challenges to the resolution and robustness of tomography methods. Traditional tomography methods have difficulty balancing reliability and efficiency. Therefore, we developed a time-difference adjoint tomography method combining the arrival-time difference and the adjoint state method. The effectiveness was verified by numerical experiments and a laboratory-scale acoustic experiment. The effectiveness of the proposed method was demonstrated by the experimental results. The adjoint scheme avoids additional ray tracing and improves the efficiency of the inversion, which allows the use of finer forward grids in practice. By considering the differential arrivals of receiver pairs, the proposed method is robust in the face of systematic errors and relatively stable against large random noises. Moreover, the velocity contrast obtained by the proposed method is sharper than for first-arrival tomography in the areas where the rays are not dense, resulting in a clearer indication of the anomalous areas in the tomographic image.