Construction of a three-dimensional geological model for mine roadways based on multi-source data fusion

Transparent characterization of mine geological conditions is a key prerequisite for rapid and intelligent roadway driving. To address the insufficient three-dimensional visualization, limited model accuracy, and inadequate dynamic representation of existing geological modeling techniques, this study takes an advancing roadway in the Huainan-Huaibei mining area as the engineering case. The sources of geological information and the characteristics of detection data acquired during roadway driving were systematically analyzed. On the basis of drilling, geophysical prospecting, and actual exposure data, a multi-source heterogeneous data fusion framework was developed, and a three-dimensional static geological model for roadway driving was established. At the same time, the static geological model is updated by using the dynamic mine seismic reflection method and electrical method detection and the advanced detection data obtained from drilling in the process of roadway driving. A dynamic reconstruction method for a three-dimensional roadway geological model based on staged progressive prediction was proposed, and the prediction accuracy of the dynamic model was evaluated. The results show that three-dimensional geological modeling based on multi-source geological data can effectively improve the accuracy of geological-condition visualization. Compared with the static model, the updated model reduced the overall modeling error by approximately 73.2% and increased relative accuracy by approximately 3.77 times, more faithfully represent the spatial characteristics of geological bodies, and support the construction of a refined three-dimensional geological model for an advancing roadway.