Acta Geodaetica et Cartographica Sinica ›› 2023, Vol. 52 ›› Issue (6): 956-965.doi: 10.11947/j.AGCS.2023.20210463

• Photogrammetry and Remote Sensing • Previous Articles     Next Articles

Point cloud virtual datum determination method in deformation analysis

SUN Wenxiao1,2, WANG Jian1, JIN Fengxiang1,2, YANG Yikun3   

  1. 1. College of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, China;
    2. College of Surveying and Geo-Informatics, Shandong Jianzhu University, Jinan 250101, China;
    3. College of Global Change and Earth System Science, Beijing Normal University, Beijing 100875, China
  • Received:2021-08-17 Revised:2022-09-22 Published:2023-07-08
  • Supported by:
    Introduction Plan of High-end Foreign Experts (No. G2021025006L)

Abstract: Aiming at the problem that conventional fixed-point-based datum construction methods are difficult to apply to overall datum detection based on point cloud, a point cloud virtual datum determination method based on the centroid stability and distribution similarity of corresponding grids of the multitemporal laser point cloud is proposed in our study. Firstly, the point cloud distribution characteristics in the stable and deformed areas are analyzed according to the changing pattern of the traditional datum points, and the detection principle of the point cloud virtual datum is discussed. Then, the point position errors caused by distance measurement, angle measurement, and footprint scale are weighted, and the axis vectors angle corresponding to the minimum moment of inertia is calculated to determine the point cloud distribution similarity. Furthermore, the centroid stability is analyzed by using the squared Msplitsimilarity transformation, and the virtual datum of the multitemporal point cloud is detected. Finally,the tank, landslide, and regular geometry point cloud captured by terrestrial laser scanning technology are applied to verify the feasibility and application scehes of the proposed method. Results show that the point cloud virtual datum that does not require on-site layout or regular maintenance can be detected by combining the centroid stability and the distribution similarity characterized by the axis vectors angles of the minimum moment of inertia, which provides the foundation for multitemporal point cloud coordinate unification and deformation analysis.

Key words: point cloud virtual datum, centroid stability, distribution similarity, deformation analysis

CLC Number: