Underground coal mine auxiliary operation robot equipment can reduce labour intensity, improve safety production efficiency, by more and more researchers' attention, three-dimensional scene real-time perception and spatial target detection and positioning is the basis of accurate control and autonomous operation of underground auxiliary operation robot. The research is centred on target detection and control system for underground assisted operation robots in coal mines, firstly, it studies the fusion analysis method based on 3D LiDAR and video image data, and realises the spatial and temporal synchronization of laser point cloud data and image data, as well as the matching analysis of the position. Then, the improved YOLOv8s image target detection algorithm model based on Slim-neck feature fusion network and the 3D point cloud target detection algorithm model based on PointPillars are designed, which reduces the complexity of the model while maintaining the recognition accuracy, on the basis of which, the improved DS evidence theory based on the Lange's distance is proposed, and the fusion target detection model YOPilaNet for the video image and 3D point cloud data is established. Experimental validation is carried out using the KITTI dataset, and the experimental results show that the proposed YOPilaNet fusion model significantly outperforms the target detection performance under single modal data. Finally, combining the embedded GPU processor and real-time operating system, a cooperative control system for target detection and multi-axis robotic arm is designed for underground coal mine robots, which can meet the demand for efficient automated operation in underground handling, pipeline assistance and other application scenarios.