Volume XLII-4/W18
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 759–763, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-759-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 759–763, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-759-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

  18 Oct 2019

18 Oct 2019

SIMULTANEOUS REGISTRATION AND INTEGRATION OF TWO SEQUENTIAL VELODYNE POINT CLOUDS USING VOXEL-BASED LEAST SQUARE ADJUSTMENT

L. Moradi and M. Saadatseresht L. Moradi and M. Saadatseresht
  • School of Surveying and Geospatial Engineering, University of Tehran, Iran

Keywords: Scan Matching, Velodyne Point Clouds, Voxel-Based Registration, Integration, Indoor mapping

Abstract. In this paper, a model for simultaneous registration and 3D modelling of Velodyne VLP 32e laser scanner point clouds based on least square adjustment methods was developed. Considering that the most of proposed methods for registration of point clouds which obtained by mobile mapping systems have applications in navigation and visualization. They usually do not pay enough attention to geometric accuracy, error propagation, and weights analysis. In addition, in these methods, some point correspondence solutions are used which increase the computation time and decrease the accuracy. Therefore, the purpose of this paper is to develop a model based on the least square adjustment and focus on the weight of the plane parameters which created by a robust least square fitting algorithm. It also simultaneously creates a 3D environmental model and registers point clouds. To do this, it utilizes both point cloud voxelization and differential planes techniques. The result illustrates the high capability of the proposed solution with the optimum weight of plane parameters to 100, and average distance between two scans can reach to below 10 mm.In addition, the best voxel size was 10 cm which is twice of point cloud resolutions.