• Laser & Optoelectronics Progress
  • Vol. 59, Issue 10, 1015010 (2022)
Wei Cai1, Dongjie Yue1、*, and Qiang Chen2
Author Affiliations
  • 1School of Earth Science and Engineering, Hohai University, Nanjing 211100, Jiangsu , China
  • 2Shanghai Institute of Surveying and Mapping, Third Branch, Shanghai 200063, China
  • show less
    DOI: 10.3788/LOP202259.1015010 Cite this Article Set citation alerts
    Wei Cai, Dongjie Yue, Qiang Chen. Point Cloud Data Registration Based on Binary Feature Descriptors[J]. Laser & Optoelectronics Progress, 2022, 59(10): 1015010 Copy Citation Text show less
    Keypoints extracted by different algorithms. (a) BSC algorithm; (b) our algorithm
    Fig. 1. Keypoints extracted by different algorithms. (a) BSC algorithm; (b) our algorithm
    Local coordinate system of keypoints and their neighborhoods
    Fig. 2. Local coordinate system of keypoints and their neighborhoods
    Point cloud projection and gridding. (a) xoy plane; (b) xoz plane; (c) yoz plane
    Fig. 3. Point cloud projection and gridding. (a) xoy plane; (b) xoz plane; (c) yoz plane
    Weighted projection density characteristics
    Fig. 4. Weighted projection density characteristics
    Weighted projection distance characteristics
    Fig. 5. Weighted projection distance characteristics
    Schematic diagram of binarization of projection distance feature
    Fig. 6. Schematic diagram of binarization of projection distance feature
    Original point cloud data. (a) Source point cloud; (b) target point cloud
    Fig. 7. Original point cloud data. (a) Source point cloud; (b) target point cloud
    PR curves of different descriptors
    Fig. 8. PR curves of different descriptors
    Matching point pairs obtained based on improved geometric consistency method
    Fig. 9. Matching point pairs obtained based on improved geometric consistency method
    Matching point pair processed based on RANSAC algorithm
    Fig. 10. Matching point pair processed based on RANSAC algorithm
    Coarse registration results for different algorithms. (a) FPFH; (b) SHOT; (c) BSC; (d) our algorithm
    Fig. 11. Coarse registration results for different algorithms. (a) FPFH; (b) SHOT; (c) BSC; (d) our algorithm
    Fine registration result of our algorithm
    Fig. 12. Fine registration result of our algorithm
    AlgorithmNumber of neighbor points
    101001000
    FPFH152.917217.471295.652
    SHOT114.543183.762228.776
    3DHoPD59.63171.76793.185
    BSC18.16320.13222.173
    Ours10.31213.53715.273
    Table 1. Computation time of different algorithms
    Wei Cai, Dongjie Yue, Qiang Chen. Point Cloud Data Registration Based on Binary Feature Descriptors[J]. Laser & Optoelectronics Progress, 2022, 59(10): 1015010
    Download Citation