Abstract: In order to solve the key problem that the existence of the mine car point sets in open pit decreases the accuracy of the UAV point cloud data， an automatic extraction method of the mine car point sets in open pit stope was proposed. Using the UAV point cloud of the Yabaling open pit as the data source， the ground points and non-ground points were distinguished using the progressive morphological filtering algorithm， then the set of mine car points were extracted from the non-ground points by clustering them with an improved Euclidean clustering algorithm， and finally the mine car extraction results were evaluated and analyzed based on the error criterion proposed by the International Society for Photogrammetry and Remote Sensing (ISPRS). The results show that the method can effectively extract the set of mine car points in the open pit， and can provide an important technical support for fast and efficient check and acceptance of the open pit.

Key words: UAV 3D point cloud； acceptance measurement of open pit； point cloud segmentation and extraction； progressive morphological filtering algorithm； Euclidean clustering algorithm