International Journal of Earth Sciences and Engineering

Open Access Open Access  Restricted Access Subscription Access

Improved ICP Point Cloud Registration Based on KDTree


Affiliations
1 School of Civil Engineering, Anhui Jianzhu University, Hefei, China
2 Department of Library, Anhui Jianzhu University, Hefei, China
 

Based on the analysis of the principle of ICP and KDTree algorithms, the paper summarized the defects of ICP and the advantages of KDTree, and with KDTree characteristics proposed fast traversal search nearest neighbor point set to form same name points, then the registration is realized to calculate transformation parameters R and T by Four Element method. Through verification of measured data and simulated data, it has been significantly improved in time and efficiency for the improved ICP algorithm based on KDTree algorithm which related to traditional ICP algorithm. This method enriches Terrestrial 3D laser scanning technology and registration technology. It will play an important role in the field of Geographic Information Science, Remote Sensing image processing, Reverse Engineering.

Keywords

Point Cloud Data Registration, ICP, KDTree, Four Element Method.
User
Notifications
Font Size

Abstract Views: 142

PDF Views: 267




  • Improved ICP Point Cloud Registration Based on KDTree

Abstract Views: 142  |  PDF Views: 267

Authors

Guigang Shi
School of Civil Engineering, Anhui Jianzhu University, Hefei, China
Xuguang Gao
School of Civil Engineering, Anhui Jianzhu University, Hefei, China
Xinghua Dang
Department of Library, Anhui Jianzhu University, Hefei, China

Abstract


Based on the analysis of the principle of ICP and KDTree algorithms, the paper summarized the defects of ICP and the advantages of KDTree, and with KDTree characteristics proposed fast traversal search nearest neighbor point set to form same name points, then the registration is realized to calculate transformation parameters R and T by Four Element method. Through verification of measured data and simulated data, it has been significantly improved in time and efficiency for the improved ICP algorithm based on KDTree algorithm which related to traditional ICP algorithm. This method enriches Terrestrial 3D laser scanning technology and registration technology. It will play an important role in the field of Geographic Information Science, Remote Sensing image processing, Reverse Engineering.

Keywords


Point Cloud Data Registration, ICP, KDTree, Four Element Method.