Material pile edge detection method

The invention discloses a material pile edge detection method, and relates to the technical field of material pile detection. The method comprises the following steps: carrying out reference point calibration on a material pile in advance, and obtaining monocular multi-view edge image information; carrying out feature point extraction on the monocular multi-view edge image information based on a scale-invariant feature transform (SIFT) algorithm, obtaining material pile section curve point clouds, and acquiring global static point cloud data; generating sparse point clouds from the acquired global static point cloud data based on a structure from motion (SFM) algorithm; generating dense point clouds based on a clustering views for multi-view stereo (CMVS) algorithm and a patch-based multi-view stereo (PMVS) algorithm; and acquiring a real-time three-dimensional point cloud model of the material pile. According to the invention, the engineering construction inventory difficulty is reduced, the cost is reduced,