Local to global registration of multi-view range scans using spanning tree. (February 2017)

• Record Type: Journal Article
• Authors: Zhu, Jihua; Zhu, Li; Jiang, Zutao; Bai, Xiuxiu; Li, Zhongyu; Wang, Lin
• Is Part Of: Computers & electrical engineering. Volume 58(2017)
• Page Start: 477
• Page End: 488
• Publication Date: 2017-02
• Subjects: 3D reconstruction -- Registration -- Range scan -- Spanning tree -- Iterative closest point; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2016.07.003
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: An approach is proposed for multi-view registration of scans using spanning tree. It achieves multi-view registration of initially posed scans with good performances. It can be applied to 3D object and scene reconstruction. Graphical abstract: Abstract: Registration is a prerequisite in 3-dimensional reconstruction. This paper proposes an effective approach for multi-view registration of range scans using a spanning tree. Given initially posed scans, it is necessary to estimate the overlap percentage for each scan pair and construct a spanning tree. Based on this tree, pairwise registration can be sequentially used to align the root scan and its directly connected scans. Then, local multi-view registration is applied to the root scan and its connected scans, which must be merged into the root scan after the local registration. By updating the spanning tree, the new root scan with its connected scans can be recursively processed and merged until all scans are merged into one scan. Finally, global registration can be applied to all scans to reduce the accumulated errors generated from scan merging. Experimental results confirm that the proposed approach can achieve multi-view registration with excellent performances.