基于主成分特征向量的点云配准方法

doi:10.3969/j.issn.0255-8297.2024.06.006

摘要/Abstract

摘要： 已有点云配准算法对杂乱点云的配准精度较低，耗时较长，为此提出一种基于主成分特征向量的点云配准方法。首先，通过描述点云曲率变化情况提取点云特征点集，并利用重心法使参考点云与待配准点云的特征点集的重心重合，实现初始位姿确定，达到点云粗配准的目的；然后，在迭代最近点算法进行迭代时，利用主成分分析算法对特征点集进行主成分分析，选取前三个主成分特征向量，通过刚体变换进行对应匹配，再利用欧氏距离寻找最近点，实现点云精配准。采用公共点云和文物点云数据模型对所提的配准方法进行验证，结果表明该方法比已有方法的配准精度平均提高了约12%，配准耗时平均降低了约10%，具有良好的配准结果。表明该基于主成分特征向量的配准方法是一种有效的点云配准方法。

关键词: 点云配准, 曲率, 迭代最近点, 主成分分析, 特征向量

Abstract: To address the issues of low accuracy and long time consumption of the existing point cloud registration algorithms for cluttered point clouds, a point cloud registration method based on principal component eigenvectors is proposed. Firstly, feature point set is extracted by describing the curvature change of the point cloud, and the center of gravity method is applied to align the center of gravity of the reference point cloud with that of the feature point set, achieving an initial rough registration. Then, during the iterative closest point (ICP) algorithm, principal component analysis (PCA) is used to select the first three principal component feature vectors and perform corresponding matching through rigid body transformation. Lastly, the Euclidean distance is used to find the nearest points for fine registration. The proposed method was validated using both public point cloud and cultural relic point cloud. Experimental results show that the registration accuracy of the proposed method is improved by approximately 12% on average, while the registration time is reduced by about 10% on average. These results indicate that the proposed method based on principal component eigenvectors is an effective approach for point cloud registration.

Key words: point cloud registration, curvature, iteration closest point, principal component analysis, eigenvector

中图分类号:

TP391

赵夫群, 黄鹤, 耿国华. 基于主成分特征向量的点云配准方法[J]. 应用科学学报, 2024, 42(6): 962-976.

ZHAO Fuqun, HUANG He, GENG Guohua. Point Cloud Registration Method Based on Principal Component Eigenvectors[J]. Journal of Applied Sciences, 2024, 42(6): 962-976.

参考文献

[1] Wang J, Wang P, Li B, et al. Discriminative optimization algorithm with global-local feature for LIDAR point cloud registration [J]. International Journal of Remote Sensing, 2021, 42(23): 8994-9014.
[2] Park B S, Kim W, Kim J K, et al. 3D static point cloud registration by estimating temporal human pose at multiview [J]. Sensors, 2022, 22(3): 1097-1107.
[3] Markovetskii A, Voronin S, Kober V, et al. Point cloud registration based on multiparameter functional [J]. Mathematics, 2021, 9(20): 2589-2599.
[4] 孙艺新, 柳占杰, 刘哲, 等. 融合北斗及LiDAR移动测量的电力工程道路横断面自动获取方法[J]. 应用科学学报, 2022, 40(6): 953-963. Sun Y X, Liu Z J, Liu Z, et al. Automatic acquisition method of electric power engineering road cross section integrating BeiDou and LiDAR mobile measurement [J]. Journal of Applied Sciences, 2022, 40(6): 953-963. (in Chinese)
[5] Besl P J, Mckay N D. A method for registration of 3-D shapes [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 239-256.
[6] 莫悠, 钟若飞, 张振鑫. 移动与定点扫描结合的室内点云数据获取方法[J]. 应用科学学报, 2018, 36(5): 756-764. Mo Y, Zhong R F, Zhang Z X. Acquisition of indoor laser point clouds based on mobile and terrestrial scanning [J]. Journal of Applied Sciences, 2018, 36(5): 756-764. (in Chinese)
[7] Risinkiewicz S. A symmetric objective function for ICP [J]. ACM Transactions on Graphics, 2019, 38(4): 1-7.
[8] Ying W J, Sun S Y. An improved Monte Carlo localization using optimized iterative closest point for mobile robots [J]. Cognitive Computation and Systems, 2022, 4(1): 20-30.
[9] 张金艺, 梁滨, 唐笛恺, 等. 粗匹配和局部尺度压缩搜索下的快速ICP-SLAM [J]. 智能系统学报, 2017, 12(3): 413-421. Zhang J Y, Liang B, Tang D K, et al. Fast ICP-SLAM with rough alignment and local scale-compressed searching [J]. CAAI Transactions on Intelligent Systems, 2017, 12(3): 413-421. (in Chinese)
[10] Jin Y T, Zhang Y H, Cui H H, et al. An aircraft skin registration method based on contour feature constraint [J]. Acta Optica Sinica, 2021, 41(3): 0312001.
[11] Li C X, Yang S S, Shi L, et al. PTRNet: global feature and local feature encoding for point cloud registration [J]. Applied Sciences, 2022, 12(3): 1741-1750.
[12] Liu D Z, Zhang Y, Luo L, et al. PDC-Net: robust point cloud registration using deep cyclic neural network combined with PCA [J]. Applied Optics, 2021, 60(11): 2990-2997.
[13] Fontana S, Sorrenti D G. A termination criterion for probabilistic point clouds registration [J]. Signals, 2021, 2(2): 159-173.
[14] Charpentier A, Mussard S, Ouraga T. Principal component analysis: a generalized Gini approach [J]. European Journal of Operational Research, 2021, 294(1): 236-249.
[15] 张银波, 李思宁, 姜鹏, 等. PCA特征提取和弹性BP神经网络的水下气泡识别[J]. 红外与激光工程, 2021, 50(6): 209-215. Zhang Y B, Li S N, Jiang P, et al. Underwater bubbles recognition based on PCA feature extraction and elastic BP neural network [J]. Infrared and Laser Engineering, 2021, 50(6): 209- 215. (in Chinese)
[16] Kim S, Kim I. Simplicial volume, barycenter method, and bounded cohomology [J]. Mathematische Annalen, 2020, 377(1): 555-616.
[17] Hasegawa K, Amabile C, Nesme M, et al. Gravity center estimation for evaluation of standing whole body compensation using virtual barycentremetry based on biplanar slot-scanning stereoradiography-validation by simultaneous force plate measurement [J]. BMC Musculoskeletal Disorders, 2022, 23(1): 22-33.
[18] Vu T, Chunikhina E, Raich R. Perturbation expansions and error bounds for the truncated singular value decomposition [J]. Linear Algebra and Its Applications, 2021, 627: 94-139.
[19] Jiang J K, Zhang Q, Xin X J, et al. Blind modulation format identification based on principal component analysis and singular value decomposition [J]. Electronics, 2022, 11(4): 612-627.
[20] Rajani D, Kumar P R. An optimized hybrid algorithm for blind watermarking scheme using singular value decomposition in RDWT-DCT domain [J]. Journal of Applied Security Research, 2022, 17(1): 103-122.
[21] Kuçak R A, Erol S, Erol B. An experimental study of a new keypoint matching algorithm for automatic point cloud registration [J]. ISPRS International Journal of Geo-Information, 2021, 10(4): 204-216.
[22] 杨稳, 周明全, 郭宝, 等. 基于曲率图的颅骨点云配准方法[J]. 光学学报, 2020, 40(16): 1610002. Yang W, Zhou M Q, Guo B, et al. Skull point cloud registration method based on curvature maps [J]. Acta Optica Sinica, 2020, 40(16): 1610002. (in Chinese)
[23] Wang Y B, Xiao J, Liu L P, et al. Efficient rock mass point cloud registration based on local invariants [J]. Remote Sensing, 2021, 13(8): 1540-1548.
[24] Song Y N, Shen W M, Peng K K. A novel partial point cloud registration method based on graph attention network [J]. The Visual Computer, 2023, 39(3): 1109-1120.
[25] Yi R B, Li J L, Luo L, et al. DOPNet: achieving accurate and efficient point cloud registration based on deep learning and multi-level features [J]. Sensors, 2022, 22(21): 8217-8229.
[26] 赵夫群. 基于多特征的兵马俑断裂面匹配方法研究[D]. 西安: 西北大学, 2019.