Abstract: At present, my country's thick coal seams mainly adopt fully mechanized caving mining technology, and problems such as over-caving or under-caving often occur during the caving process. Therefore, the three-dimensional reconstruction of the upper part of the rear scraper during the caving process is the core to realize the fully mechanized caving mining of extra-thick coal seams. One of the technical issues. Aiming at this problem, a laser scanning intelligent monitoring system for coal discharge is designed. The laser scanning method is used to obtain the point cloud data of the coal flow profile on the scraper conveyor near the coal caving port. According to the characteristic information of the coal flow on the scraper conveyor Real-time construction of a high-precision three-dimensional model of coal stockpile; proposed a three-dimensional reconstruction method of coal caving capacity suitable for fully mechanized caving face, combined with the latest iteration point algorithm (ICP) to solve the transformation relationship between each frame of laser data, design high Sub-regression filtering eliminates splicing errors and environmental noise, making the model effect more suitable for the actual working conditions of the working face; using the peak signal-to-noise ratio and the average structural similarity of the harmonic number evaluation method, the denoising performance of different algorithms is reasonably scored And the comparison proves that the algorithm in this paper has stronger applicability.