Abstract: In unmanned coal loading operation, real-time detection and positioning of hatch is very important for reducing accident rates and ensuring normal loading operation. In order to solve the existing problems of manual detection based on video, a multi-sensor data fusion hatch detection method based on 5G communication is proposed. The point cloud data collected by the lidars and the encoder data of the ship loader arranged in different positions are fused to build the 3D model of the cargo ship, and the cabin recognition is based on point cloud spherical surface mapping, the complete 3D space information and mature image processing technology stored by point cloud are made full use of. The experimental results show that the algorithm has higher accuracy and recognition stability.