Abstract: The total station for point measurement has been mainly adopted for the acceptance measurement of open-pit mines at present，which has the disadvantages of low sampling density，large volume calculation error and long measurement period.The use of 3D laser scanning for open-pit mine acceptance has become a trend.However， the existing point cloud registration algorithm has the problems of low registration efficiency and poor precision.Aming at the current iterative closet point （ICP） point cloud fine registration requires higher initial input point cloud，and the original SIFT-ICP （Scale invariant feature transform and iterative closest point） algorithm needs to use image data as the coarse registration data source of SIFT algorithm，the original SIFT-ICP algorithm is improved，and a new SIFT-ICP algorithm is proposed.The new algorithm uses only one data source of point cloud， and combines the SIFT algorithm to quickly identify feature points and ICP precise registration，which improves the initial point cloud input in the fine registration process，thus achieving fast and accurate registration.The actual application verification in Anqian Mine in Anshan City，Liaoning Province show that the new SIFT-ICP algorithm can quickly complete the terrestrial 3D laser scanning point cloud registration，and the relative error between the calculated stripping volume and the reference results is only for 0.72%，which can replace the widely-used point measurement method such as total station，besides that，the efficiency of acceptance surveys in open-pit mines can be improved.

Key words: Terrestrial 3D laser scanning, Point cloud, SIFT, ICP, Stripping volume