GraphReg: Dynamical Point Cloud Registration With Geometry-Aware Graph Signal Processing

doi:10.1109/TIP.2022.3223793

CASIA OpenIR > 多模态人工智能系统全国重点实验室 > 三维可视计算

	GraphReg: Dynamical Point Cloud Registration With Geometry-Aware Graph Signal Processing
	Zhao, Mingyang 1,2; Ma, Lei1,3,4 ; Jia, Xiaohong 5,6; Yan, Dong-Ming7,8 ; Huang, Tiejun 9,10,11
发表期刊	IEEE TRANSACTIONS ON IMAGE PROCESSING
ISSN	1057-7149
	2022
卷号	31 页码:7449-7464
通讯作者	Ma, Lei(lei.ma@pku.edu.cn) ; Yan, Dong-Ming(yandongming@gmail.com)
摘要	This study presents a high-accuracy, efficient, and physically induced method for 3D point cloud registration, which is the core of many important 3D vision problems. In contrast to existing physics-based methods that merely consider spatial point information and ignore surface geometry, we explore geometry aware rigid-body dynamics to regulate the particle (point) motion, which results in more precise and robust registration. Our proposed method consists of four major modules. First, we leverage the graph signal processing (GSP) framework to define a new signature, i.e., point response intensity for each point, by which we succeed in describing the local surface variation, resampling keypoints, and distinguishing different particles. Then, to address the shortcomings of current physics based approaches that are sensitive to outliers, we accommodate the defined point response intensity to median absolute deviation (MAD) in robust statistics and adopt the X84 principle for adaptive outlier depression, ensuring a robust and stable registration. Subsequently, we propose a novel geometric invariant under rigid transformations to incorporate higher-order features of point clouds, which is further embedded for force modeling to guide the correspondence between pairwise scans credibly. Finally, we introduce an adaptive simulated annealing (ASA) method to search for the global optimum and substantially accelerate the registration process. We perform comprehensive experiments to evaluate the proposed method on various datasets captured from range scanners to LiDAR. Results demonstrate that our proposed method outperforms representative state-of-the-art approaches in terms of accuracy and is more suitable for registering large-scale point clouds. Furthermore, it is considerably faster and more robust than most competitors. Our implementation is publicly available at https://github.com/zikail/GraphReg.
关键词	Point cloud registration graph signal processing rigid dynamics robust statistics simulated annealing
DOI	10.1109/TIP.2022.3223793
收录类别	SCI
语种	英语
资助项目	National Key Research and DevelopmentProgram of China[2020AAA0105200] ; NationalNatural Science of Foundation for Outstanding Young Scholars[12022117] ; CAS Project for Young Scientists in BasicResearch[YSBR-034] ; National Natural ScienceFoundation of China[61872354] ; National Natural ScienceFoundation of China[62172415] ; Open Research Fund Program of State key Laboratory ofHydroscience and Engineering, Tsinghua University[sklhse-2022-D-04]
项目资助者	National Key Research and DevelopmentProgram of China ; NationalNatural Science of Foundation for Outstanding Young Scholars ; CAS Project for Young Scientists in BasicResearch ; National Natural ScienceFoundation of China ; Open Research Fund Program of State key Laboratory ofHydroscience and Engineering, Tsinghua University
WOS研究方向	Computer Science ; Engineering
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS记录号	WOS:000896645500002
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计	被引频次：4[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/51031
专题	多模态人工智能系统全国重点实验室_三维可视计算
通讯作者	Ma, Lei; Yan, Dong-Ming
作者单位	1.Beijing Acad Artificial Intelligence, Beijing 100190, Peoples R China 2.Chinese Acad Sci, Inst Automat, NLPR, Beijing 100190, Peoples R China 3.Inst Artificial Intelligence, Beijing 100190, Peoples R China 4.Peking Univ, Natl Biomed Imaging Ctr, Beijing 100871, Peoples R China 5.Chinese Acad Sci, Acad Math & Syst Sci, Beijing 100190, Peoples R China 6.UCAS, Sch Math Sci, Beijing 100149, Peoples R China 7.Chinese Acad Sci, Inst Automat, NLPR, Beijing 100864, Peoples R China 8.UCAS, Sch AI, Beijing 101408, Peoples R China 9.Peking Univ, Natl Engn Res Ctr Visual Technol, Sch Comp Sci, Beijing 100871, Peoples R China 10.Peking Univ, BAAI, Beijing 100871, Peoples R China 11.Peking Univ, Inst Artificial Intelligence, Beijing 100871, Peoples R China
第一作者单位	模式识别国家重点实验室
通讯作者单位	模式识别国家重点实验室
推荐引用方式 GB/T 7714	Zhao, Mingyang,Ma, Lei,Jia, Xiaohong,et al. GraphReg: Dynamical Point Cloud Registration With Geometry-Aware Graph Signal Processing[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2022,31:7449-7464.
APA	Zhao, Mingyang,Ma, Lei,Jia, Xiaohong,Yan, Dong-Ming,&Huang, Tiejun.(2022).GraphReg: Dynamical Point Cloud Registration With Geometry-Aware Graph Signal Processing.IEEE TRANSACTIONS ON IMAGE PROCESSING,31,7449-7464.
MLA	Zhao, Mingyang,et al."GraphReg: Dynamical Point Cloud Registration With Geometry-Aware Graph Signal Processing".IEEE TRANSACTIONS ON IMAGE PROCESSING 31(2022):7449-7464.