Coal mine geological hazardous body detection using surface ground penetrating radar velocity tomography

Ground penetrating radar (GPR) velocity tomography has shown much promise in detection of geological hazardous body because of its ability of providing highly detailed images of wave velocity. In this paper, velocity tomography using surface GPR is presented tor detection harmful geology structure in underground coal mine. Compared with borehole radar, it avoids drilling and can be used to implement nondestructive detection with no influence on coal mining. Damping least square orthogonal triangle (LSQR) is employed to solve the large sparse velocity tomography equations. It is compared to the classical algebraic reconstruction technique (ART) and simultaneous iterative reconstruction technique (SIRT) using three synthetic geological models with representative disaster areas, namely, fault, cavity and water zone, considering both high and low velocity anomaly body. In each case, damping LSQR provides comparable or better results than ART and SIRT. The shape of the anomaly body has great influence on inversion accuracy. The simulation results verify the proposed approach can identify the location, scope and characters of disaster areas and faults parallel to transmission direction are better inverted.