Abstract: In order to solve the problem of difficulty in fully characterizing the hidden dangers of sand excavation holes in the South to North Water Diversion Middle Route Project using a single geophysical method, a drone collaborative multi-source geophysical fine detection scheme is designed, integrating drone aerial survey (DOM/DSM), ground penetrating radar (GPR), high-density electrical method (HDERT), and transient surface wave method (SWM) to carry out fine identification of hidden dangers in typical canal sections of sand excavation holes. Through multi-source data fusion, precise positioning of the spatial distribution of sand excavation holes is achieved (with a lateral error of ± 0.3 m and a vertical error of ± 0.5 m). The system reveals the structural characteristics of "surface tension cracks middle cavity bottom loose filling" in its early disaster manifestations, analyzes the causal mechanism, and establishes a geological geophysical response model based on drilling verification. The results indicate that unmanned aerial vehicle (UAV) aerial survey can quickly capture the surface characteristics of early diseases in sand excavation tunnels. Multi source collaboration of shallow geophysical methods such as ground penetrating radar and transient surface wave method can significantly improve the accuracy of hazard identification (comprehensive accuracy>92%). The limitations and complementarity of various geophysical techniques in detecting hidden dangers in sand dredging tunnels were discussed, and the optimal detection schemes for different types of sand dredging tunnels were analyzed. The results showed that the overall stability of the canal section was good.