Laser & Optoelectronics Progress, Volume. 57, Issue 20, 201015(2020)

Point Cloud Adaptive Registration Algorithm Based on Color Information and Geometric Information

Yong Wang1 and Chun Li2、*
Author Affiliations
  • 1Liangjiang College of Artificial Intelligence, Chongqing University of Technology, Chongqing 401135, China
  • 2School of Computer Science and Engineering, Chongqing University of Technology, Chongqing 400054, China
  • show less
    Figures & Tables(8)
    Basic flow of algorithm
    Facial Makeup point cloud registration results. (a1) Initial pose of two point clouds; registration results of (b1) two point clouds under the classical ICP algorithm, (c1) 4D-ICP(Hue) algorithm, (d1) 4D-ICP (IAICP) algorithm,and (e1) proposed algorithm; (a2) initial pose of two point clouds; registration results of (b2) two point clouds under the classical ICP algorithm, (c2) 4D-ICP(Hue) algorithm, (d2) 4D-ICP (IAICP) algorithm, and (e2) proposed algorithm
    Kettle point cloud registration results. (a1) Initial pose of two point clouds; registration results of (b1) two point clouds under the classical ICP algorithm, (c1) 4D-ICP(Hue) algorithm, (d1) 4D-ICP (IAICP) algorithm,and (e1) proposed algorithm; (a2) initial pose of two point clouds; registration results of (b2) two point clouds under the classical ICP algorithm, (c2) 4D-ICP(Hue) algorithm, (d2) 4D-ICP (IAICP) algorithm, and (e2) proposed algorithm
    Plaster statue point cloud registration results. (a1) Initial pose of two point clouds; registration results of (b1) two point clouds under the classical ICP algorithm, (c1) 4D-ICP(Hue) algorithm, (d1) 4D-ICP (IAICP) algorithm, and (e1) proposed algorithm; (a2) initial pose of two point clouds; registration results of (b2) two point clouds under the classical ICP algorithm, (c2) 4D-ICP(Hue) algorithm, (d2) 4D-ICP (IAICP) algorithm, and (e2) proposed algorithm
    Time comparison line chart of registration results. (a) Time of each method on different models without noise; (b) time of each method on different models with Gaussian noise
    Error comparison line chart of registration results. (a) Error of each method on different models without noise; (b) error of each method on different models with Gaussian noise
    • Table 1. Comparison of registration results of four algorithms without noise

      View table

      Table 1. Comparison of registration results of four algorithms without noise

      AlgorithmFacial MakeupKettlePlaster Statue
      Time /sRMSETime /sRMSETime /sRMSE
      Classic ICP4DICP(Hue)4DICP(IAICP)Proposed41.95036.97447.46063.67310.75930.75930.73430.6285161.083824.584625.05617.29461.03081.03081.01650.5526403.943348.889252.532511.55771.77771.77771.77121.7538
    • Table 2. Comparison of registration results of four algorithms with Gaussian noise

      View table

      Table 2. Comparison of registration results of four algorithms with Gaussian noise

      AlgorithmFacial MakeupKettlePlaster Statue
      Time /sRMSETime /sRMSETime /sRMSE
      Classic ICP4DICP(Hue)4DICP(IAICP)Proposed39.39518.25937.80494.13390.96981.21830.93160.6293158.291434.657828.19508.33301.05841.10071.05820.5521454.779659.026256.959612.00351.88171.82651.88111.7543
    Tools

    Get Citation

    Copy Citation Text

    Yong Wang, Chun Li. Point Cloud Adaptive Registration Algorithm Based on Color Information and Geometric Information[J]. Laser & Optoelectronics Progress, 2020, 57(20): 201015

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Image Processing

    Received: Dec. 10, 2019

    Accepted: Mar. 9, 2020

    Published Online: Oct. 17, 2020

    The Author Email: Li Chun (2351583604@qq.com)

    DOI:10.3788/LOP57.201015

    Topics