Large skin edge defect detection method based on multi-scale neighborhood

The invention discloses a large skin edge defect detection method based on a multi-scale neighborhood. According to the method, a multi-scale space search method is constructed, space information of corresponding key points is better judged, the search direction is optimized and three-dimensional point cloud edge points are recognized on the basis of the space information, the operation efficiency of three-dimensional image edge recognition is greatly improved, and the recognition success rate is increased; and meanwhile, ordering the extracted boundary points, and detecting defects by using the change of a mask characteristic value. According to the method, edge defect detection of the point cloud image can be more accurately realized in a shorter time, and the method has extremely high algorithm robustness and can be applied to the field of precision detection with large scale, complex structure and high efficiency requirement. 本发明公开了一种基于多尺度邻域的大型蒙皮边缘缺陷检测方法。该方法通过构建一种多尺度空间搜索方法，更好的判断对应关键点的空间信息，并以此为依据优化搜索方向以及识别