Adaptive bilateral filtering point cloud smoothing and IMLS evaluation method considering normal outliers

doi:10.11996/JG.j.2095-302X.2023010131

Abstract

Abstract:

In order to address the problems that the unreasonable parameters of bilateral filtering lead to poor smoothing effect of point cloud, volume shrinkage, and the limitation of existing quality evaluation methods, a bilateral filtering algorithm with adaptive parameters and a quality evaluation method based on implicit moving least squares (IMLS) were proposed. Firstly, the KD-tree data structure was constructed for point cloud topology, then the neighborhood of each point was searched to calculate the normal of each point using the SVD decomposition method, and the normal outlier factor was introduced into bilateral filtering to remove outliers in the neighborhood. Additionally, the space and normal characteristic parameters were calculated according to the Gaussian kernel function extended by the neighborhood norm. Finally, the constructed bilateral filtering model was applied to the smoothing of the point cloud, and the implicit moving least squares method was introduced to evaluate the quality of smoothing. The experimental results of the point cloud with noise show that the adaptive bilateral filtering point cloud smoothing algorithm considering normal outliers could attain a good effect and result in smaller volume shrinkage compared with other algorithms, and that the IMLS evaluation method could be objective and effective.

Key words: point cloud smoothing, bilateral filtering, normal outlier, evaluation of denoising quality

CLC Number:

TP391

CHEN Ya-chao, FAN Yan-guo, YU Ding-feng, FAN Bo-wen. Adaptive bilateral filtering point cloud smoothing and IMLS evaluation method considering normal outliers[J]. Journal of Graphics, 2023, 44(1): 131-138.

Figures/Tables 10

Fig. 1 Smooth movement of measuring point in different neighborhood

Fig. 2 Smooth movement of measuring point containing floating points

Fig. 3 Discrimination of normal outlier factor

Fig. 4 Implicit surface of IMLS

Fig. 5 Comparison of effect between different algorithms for elephant model ((A) Comparison of over smoothing effect of different algorithms; (B) Comparison of detail effect of different algorithms)

Fig. 6 Comparison of effect between different algorithms for gargo model ((A) Comparison of over smoothing effect of different algorithms; (B) Comparison of detail effect of different algorithms)

Table 1 Quantitative evaluation table of different denoising algorithms for elephant model

方法	空间超参数	法向超参数	拟高斯噪声(%)	运行时间(s)	P-Point ^[18]评分	IMLS评分
核函数 MeanShift	10	0.1	0	0.121	1.323	0.325
	50	0.1	0	0.121	2.412	0.314
	10	0.1	10	0.124	3.377	0.522
	50	0.1	10	0.123	4.327	0.386
文献[9]	10	0.1	0	0.239	2.075	0.453
	50	0.1	0	0.233	2.287	0.466
	10	0.1	10	0.239	8.103	1.103
	50	0.1	10	0.237	8.348	1.103
文献[11]	10	0.1	0	0.252	0.861	0.370
	50	0.1	0	0.246	1.343	0.377
	10	0.1	10	0.250	2.443	0.699
	50	0.1	10	0.250	3.090	0.675
本文算法	-	-	0	0.395	0.462	0.319
本文算法	-	-	10	0.390	1.717	0.579

Table 2 Quantitative evaluation table of different denoising algorithms for gargo model

方法	空间超参数	法向超参数	拟高斯噪声(%)	运行时间(s)	P-Point ^[18]评分	IMLS评分
核函数 MeanShift	10	0.1	0	0.122	8.846	0.979
	50	0.1	0	0.120	15.677	0.866
	10	0.1	10	0.125	8.603	1.228
	50	0.1	10	0.123	17.238	0.912
文献[9]	10	0.1	0	0.234	20.852	1.440
	50	0.1	0	0.235	23.995	1.429
	10	0.1	10	0.242	32.969	1.962
	50	0.1	10	0.236	36.104	1.913
文献[11]	10	0.1	0	0.247	2.801	1.134
	50	0.1	0	0.251	6.373	1.074
	10	0.1	10	0.248	6.529	1.388
	50	0.1	10	0.246	9.984	1.281
本文算法	-	-	0	0.391	3.548	0.887
本文算法	-	-	10	0.391	6.407	1.036

Fig. 7 Iteration result of bilateral filtering class algorithm for elephant model ((a) Quasi Gaussian noise 0%; (b) Quasi Gaussian noise 10%)

Fig. 8 Iteration result of bilateral filtering class algorithm for Gargo model ((a) Quasi Gaussian noise 0%; (b) Quasi Gaussian noise 10%)

References 19

[1]	BERGER M, TAGLIASACCHI A, SEVERSKY L M, et al. A survey of surface reconstruction from point clouds[J]. Computer Graphics Forum, 2017, 36(1): 301-329. DOI URL
[2]	宋时德, 李淼, 张健, 等. 面向以多视角立体匹配获取的植株三维点云的去噪方法[J]. 计算机应用, 2017, 37(S2): 141-145.
	SONG S D, LI M, ZHANG J, et al. Denoising method of 3D point clouds based on multi-angle stereo matching[J]. Journal of Computer Applications, 2017, 37(S2): 141-145 (in Chinese).
[3]	刘立恒, 赵夫群, 汤慧, 等. 几何特征保持的文物点云去噪算法[J]. 数据采集与处理, 2020, 35(2): 373-380.
	LIU L H, ZHAO F Q, TANG H, et al. A denoising method for point cloud of cultural relics with geometric feature preservation[J]. Journal of Data Acquisition and Processing, 2020, 35(2): 373-380 (in Chinese).
[4]	陈世海, 李俊明, 王雯. 基于3D点云的电缆自适应多尺度去噪算法研究[J]. 国外电子测量技术, 2020, 39(10): 115-118.
	CHEN S H, LI J M, WANG W. Research on cable adaptive multi-scale denoising algorithm based on 3D point cloud[J]. Foreign Electronic Measurement Technology, 2020, 39(10): 115-118 (in Chinese).
[5]	程效军, 贾东峰, 程小龙. 海量点云数据处理理论与技术[M]. 上海: 同济大学出版社, 2014: 13-14.
	CHENG X J, JIA D F, CHENG X L. Theory and technology of massive point cloud data processing[M]. Shanghai: Tongji University Press, 2014: 13-14 (in Chinese).
[6]	RUSU R B, COUSINS S. 3D is here: point cloud library (PCL)[C]//2011 IEEE International Conference on Robotics and Automation. New York: IEEE Press, 2011: 1-4.
[7]	刘巍, 逯永康, 赵海洋, 等. 一种基于区域分割的大尺度点云噪声去噪方法: CN108876744A[P]. 2018-11-23.
	LIU W, LU Y K, ZHAO H Y, et al. Method for large-scale point cloud noise reduction based on region segmentation: CN108876744A[P]. 2018-11-23 (in Chinese).
[8]	毕松, 王宇豪. 基于自适应半径滤波的农业导航激光点云去噪方法研究[J]. 农业机械学报, 2021, 52(11): 234-243.
	BI S, WANG Y H. LiDAR point cloud denoising method based on adaptive radius filter[J]. Transactions of the Chinese Society for Agricultural Machinery, 2021, 52(11): 234-243 (in Chinese).
[9]	FLEISHMAN S, DRORI I, COHEN-OR D. Bilateral mesh denoising[J]. ACM Transactions on Graphics, 2003, 22(3): 950-953. DOI URL
[10]	ZHENG Y Y, FU H B, AU O K C, et al. Bilateral normal filtering for mesh denoising[J]. IEEE Transactions on Visualization and Computer Graphics, 2011, 17(10): 1521-1530. DOI URL
[11]	袁华, 庞建铿, 莫建文. 基于噪声分类的双边滤波点云去噪算法[J]. 计算机应用, 2015, 35(8): 2305-2310. DOI
	YUAN H, PANG J K, MO J W. Denoising algorithm for bilateral filtered point cloud based on noise classification[J]. Journal of Computer Applications, 2015, 35(8): 2305-2310 (in Chinese). DOI
[12]	朱广堂, 叶珉吕. 基于曲率特征的点云去噪及定量评价方法研究[J]. 测绘通报, 2019(6): 105-108. DOI
	ZHU G T, YE M l. Research on the method of point cloud denoising based on curvature characteristics and quantitative evaluation[J]. Bulletin of Surveying and Mapping, 2019(6): 105-108 (in Chinese). DOI
[13]	ZOU B C, QIU H D, LU Y F. Point cloud reduction and denoising based on optimized downsampling and bilateral filtering[J]. IEEE Access, 2020(8): 136316-136326.
[14]	LEAL E, SANCHEZ-TORRES G, BRANCH J W. Sparse regularization-based approach for point cloud denoising and sharp features enhancement[J]. Sensors, 2020, 20(11): 3206. DOI URL
[15]	杨洋, 李明. 基于混合滤波和DBSCAN的点云去噪算法[J]. 计量与测试技术, 2020, 47(4): 24-27.
	YANG Y, LI M. A point cloud denoising method based on a hybrid filtering and density clustering algorithm[J]. Metrology & Measurement Technique, 2020, 47(4): 24-27 (in Chinese).
[16]	蒋通, 邹进贵. 基于区域生长的体素滤波点云去噪算法[J]. 测绘地理信息, 2021, 46(2): 63-66.
	JIANG T, ZOU J G. Point cloud denoising algorithm of voxel filtering based on region growing[J]. Journal of Geomatics, 2021, 46(2): 63-66 (in Chinese).
[17]	唐瑞尹, 范乃德, 史涛, 等. 似然函数和随机滤波的点云去噪方法研究[J]. 机械设计与制造, 2022(2): 74-77.
	TANG R Y, FAN N D, SHI T, et al. Research on point cloud denoising method based on likelihood function and random filter[J]. Machinery Design ＆ Manufacture, 2022(2): 74-77 (in Chinese).
[18]	JAVAHERI A, BRITES C, PEREIRA F, et al. Subjective and objective quality evaluation of 3D point cloud denoising algorithms[C]//2017 IEEE International Conference on Multimedia & Expo Workshops. New York: IEEE Press, 2017: 1-6.
[19]	DESCHAUD J E. IMLS-SLAM: scan-to-model matching based on 3D data[C]//2018 IEEE International Conference on Robotics and Automation. New York: ACM, 2018: 2480-2485.