面向机械零件三角网格模型自动配准中增强特征的分割方法

巫志辉; 王立忠; 梁晋; 龚春园; 朱峰; 常志文; 徐建宁

doi:10.37188/CO.2023-0225

面向机械零件三角网格模型自动配准中增强特征的分割方法

doi: 10.37188/CO.2023-0225

西安交通大学机械工程学院精密微纳制造技术全国重点实验室, 陕西西安 710049

基金项目: 国家重点研发计划项目（No. 2022YFB4601802）；国家自然科学基金资助项目（No. 52275543）

详细信息

作者简介:
巫志辉(1998—)，男，江西抚州人，硕士研究生，2020年于太原理工大学获得学士学位，主要从事机器视觉及点云三角网格的研究。E-mail：wuzhihui@stu.xjtu.edu.cn

王立忠(1968—)，男，山东梁山人，博士，教授，博士生导师，2004 年于西安交通大学获得博士学位，主要从事三维光学测量技术的研究。E-mail：wanglz@mail.xjtu.edu.cn

中图分类号: TP394.1;TH691.9
计量
- 文章访问数: 75
- HTML全文浏览量: 35
- PDF下载量: 3
- 被引次数: 0
出版历程
- 收稿日期: 2023-12-18
- 录用日期: 2024-02-28
- 网络出版日期: 2024-05-17

Segmentation method for enhanced features in automatic registration of triangular mesh model of mechanical parts

State Key Laboratory for Manufacturing Systems Engineering, College of Mechanical Engineering, Xi’an Jiaotong University, Xi’an 710049, China

Funds: Supported by the National Key R&D Program of China (No. 2022YFB4601802); National Natural Science Foundation of China (No. 52275543)

More Information

Corresponding author: wanglz@mail.xjtu.edu.cn

摘要

摘要:
三角网格模型配准是工业自动化检测软件中重要的一环，配准精度对检测机械零件的形位公差有重要影响。针对三角网格模型的自动配准的精度低、鲁棒性差的问题，本文提出一种面向机械零件三角网格模型自动配准中增强特征的分割方法。首先，确定三角网格模型特征分割的K值，通过拉普拉斯矩阵确定种子点进行迭代初始化。其次本文采用合适的区域形状代理和代价函数加速该过程，并进行多源迭代聚类得到特征分割结果。最终，在三角网格模型特征分割结果的基础上进行基于奇异值分解法的粗配准，根据EM-ICP进行精配准。与传统的特征描述子粗配准再ICP精配准的方法进行对比，实验结果表明本文方法的配准误差下降了25.2%，自动配准时间缩短了62.6%，有效地提高了三角网格模型自动配准的精度和效率。
- 三角网格模型 /
- 特征分割 /
- 种子点选取 /
- 分割后处理 /
- 自动配准
Abstract:
Triangular mesh model registration is an important part of industrial automation detection software. The registration accuracy has an important influence on the shape and position tolerance of mechanical parts. Aiming at the problems of low accuracy and poor robustness of automatic registration of triangular mesh models, this paper proposes a segmentation method for enhanced features in automatic registration of triangular mesh models for mechanical parts. Firstly, the K value of the feature segmentation of the triangular mesh model is determined, and the seed points are determined by the Laplacian matrix for iterative initialization. Secondly, this paper uses the appropriate region shape agent and cost function to accelerate the process, and performs multi-source iterative clustering to obtain the feature segmentation results. Finally, based on the feature segmentation results of the triangular mesh model, the coarse registration based on the singular value decomposition method is performed, and the fine registration is performed according to the EM-ICP. Compared with the traditional feature descriptor coarse registration and ICP fine registration method, the experimental results show that the registration error of the proposed method is reduced by 25.2 %, and the automatic registration time is shortened by 62.6 %, which effectively improves the accuracy and efficiency of the automatic registration of the triangular mesh model.
- triangular mesh model /
- feature segmentation /
- seed point selection /
- segmentation post-processing /
- automatic registration

HTML全文

图 1 (a)fandisk模型与(b)管道工件模型

Figure 1. (a) Fandisk model and (b) Piping workpiece model

下载: 全尺寸图片幻灯片

图 2 种子与H距离关系示意图

Figure 2. The schematic diagram of the relationship between seed and H distance

下载: 全尺寸图片幻灯片

图 3 (a)局部三角网格和(b)抽象的图结构

Figure 3. (a) Local triangular mesh and (b) Abstract graph structure

下载: 全尺寸图片幻灯片

图 4 种子点结果示意图

Figure 4. Seed point result diagram

下载: 全尺寸图片幻灯片

图 5 (a)三角面片点与投影点和(b)点至代理平面距离

Figure 5. (a) Triangular patch points and projection points and (b) Distance from point to agent plane.

(a) Triangular patch points and projection points (b) Distance from point to agent plane

下载: 全尺寸图片幻灯片

图 6 (a)边界平滑前和(b)边界平滑后

Figure 6. (a) Before boundary smoothing and (b) After boundary smoothing

下载: 全尺寸图片幻灯片

图 7 孤立三角面片示意图

Figure 7. Isolated triangular patch diagram

下载: 全尺寸图片幻灯片

图 8 边界提取结果图

Figure 8. Boundary extraction result diagram

下载: 全尺寸图片幻灯片

图 9 (a)源三角网格模型和(b)低分辨率三角网格模型

Figure 9. (a) source triangular mesh model and (b) low resolution triangular mesh model

下载: 全尺寸图片幻灯片

图 10 fandisk模型两个视角的分割结果

Figure 10. Segmentation results of two views in fandisk model

下载: 全尺寸图片幻灯片

图 11 回转体忽略特征示意图

Figure 11. Revolving body ignore feature diagram

下载: 全尺寸图片幻灯片

图 12 分割面片的中心示意图

Figure 12. The center diagram of the split patch

下载: 全尺寸图片幻灯片

图 13 粗配准偏差色谱图

Figure 13. Coarse registration deviation chromatogram

下载: 全尺寸图片幻灯片

图 14 整体流程示意图

Figure 14. The overall process diagram

下载: 全尺寸图片幻灯片

图 15 扫描硬件设备图

Figure 15. The schematic diagram of hardware equipment

下载: 全尺寸图片幻灯片

图 16 多个模型分割结果示意图

Figure 16. Schematic diagram of multiple model segmentation results

下载: 全尺寸图片幻灯片

图 17 (a) 兰德指数和 (b) 汉明距离

Figure 17. (a) Rand index and (b) Hamming distance

下载: 全尺寸图片幻灯片

图 18 配准所用模型示意图

Figure 18. The model diagram used for registration

下载: 全尺寸图片幻灯片

表 1 自动配准RMSE和耗时结果

Table 1. Automatic registration RMSE and time-consuming results

模型序号	模型面片数量	特征描述子方法 RMSE（mm）	本文方法 RMSE（mm）	特征描述子方法耗时/s	本文方法耗时/s
1	7401766	0.695	0.249	83.43	29.76
2	34946	0.002	0.002	6.73	0.13
3	14764242	0.387	0.386	197.46	50.48
4	29686649	\	0.296	\	113.42
5	13262362	0.473	0.473	208.32	47.32
6	1920009	0.254	0.253	26.54	7.75
7	1794486	0.086	0.085	30.66	8.47
8	4418964	0.858	0.375	50.37	15.93
9	5081628	0.572	0.572	63.20	18.62
10	13136277	0.418	0.418	147.81	46.68
均值	9150133	0.416	0.311	90.50	33.86

下载: 导出CSV

参考文献(29)

[1]	周晓东, 张雅超, 谭庆昌, 等. 基于结构光视觉技术的圆柱度测量新方法[J]. 吉林大学学报(工学版),2017,47(2):524-529. ZHOU X D, ZHANG Y CH, TAN Q CH, et al. New method of cylindricity measurement based on structured light vision technology[J]. Journal of Jilin University (Engineering and Technology Edition), 2017, 47(2): 524-529. (in Chinese).
[2]	RAFFAELI R, MENGONI M, GERMANI M, et al. Off-line view planning for the inspection of mechanical parts[J]. International Journal on Interactive Design and Manufacturing, 2013, 7(1): 1-12. doi: 10.1007/s12008-012-0160-1
[3]	PHAN N D M, QUINSAT Y, LAVERNHE S, et al. Scanner path planning with the control of overlap for part inspection with an industrial robot[J]. The International Journal of Advanced Manufacturing Technology, 2018, 98(1-4): 629-643. doi: 10.1007/s00170-018-2336-8
[4]	杨鹏程, 杨朝, 孟杰, 等. 基于法向量和面状指数特征的文物点云棱界配准方法[J]. 中国光学(中英文),2023,16(3):654-662. doi: 10.37188/CO.2022-0156 YANG P CH, YANG ZH, MENG J, et al. Aligning method for point cloud prism boundaries of cultural relics based on normal vector and faceted index features[J]. Chinese Optics, 2023, 16(3): 654-662. (in Chinese). doi: 10.37188/CO.2022-0156
[5]	JUNIOR E M O, SANTOS D R, MIOLA G A R. A new variant of the ICP algorithm for pairwise 3D point cloud registration[J]. American Scientific Research Journal for Engineering, Technology, and Sciences, 2022, 85(1): 71-88.
[6]	YANG J L, LI H D, CAMPBELL D, et al. Go-ICP: a globally optimal solution to 3D ICP point-set registration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(11): 2241-2254. doi: 10.1109/TPAMI.2015.2513405
[7]	LIAN W, ZHANG L, YANG M H. An efficient globally optimal algorithm for asymmetric point matching[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(7): 1281-1293. doi: 10.1109/TPAMI.2016.2603988
[8]	LIU Y L, WANG CH, SONG ZH J, et al. Efficient global point cloud registration by matching rotation invariant features through translation search[C]. Proceedings of the 15th European Conference on Computer Vision (ECCV), Springer, 2018: 448-463.
[9]	刘跃生, 陈新度, 吴磊, 等. 混合稀疏迭代最近点配准[J]. 光学精密工程,2021,29(9):2255-2267. doi: 10.37188/OPE.20212909.2255 LIU Y SH, CHEN X D, WU L, et al. Sparse mixture iterative closest point registration[J]. Optics and Precision Engineering, 2021, 29(9): 2255-2267. (in Chinese). doi: 10.37188/OPE.20212909.2255
[10]	林森, 张强. 应用邻域点信息描述与匹配的点云配准[J]. 光学精密工程,2022,30(8):984-997. doi: 10.37188/OPE.20223008.0984 LIN S, ZHANG Q, et al. Point cloud registration using neighborhood point information description and matching[J]. Optics and Precision Engineering, 2022, 30(8): 984-997. (in Chinese). doi: 10.37188/OPE.20223008.0984
[11]	GARLAND M, WILLMOTT A, HECKBERT P S. Hierarchical face clustering on polygonal surfaces[C]. Proceedings of the 2001 Symposium on Interactive 3D Graphics, ACM, 2001: 49-58.
[12]	THEOLOGOU P, PRATIKAKIS I, THEOHARIS T. Unsupervised spectral mesh segmentation driven by heterogeneous graphs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(2): 397-410. doi: 10.1109/TPAMI.2016.2544311
[13]	SHI J B, MALIK J. Normalized cuts and image segmentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(8): 888-905. doi: 10.1109/34.868688
[14]	CHAHHOU M, MOUMOUN L, EL FAR M, et al. Segmentation of 3D meshes usingp-spectral clustering[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(8): 1687-1693. doi: 10.1109/TPAMI.2013.2297314
[15]	DONG Q J, WANG Z X, LI M Y, et al. Laplacian2Mesh: Laplacian-based mesh understanding[J]. IEEE Transactions on Visualization and Computer Graphics, doi: 10.1109/TVCG.2023.3259044.
[16]	LEI H, AKHTAR N, SHAH M, et al. Mesh convolution with continuous filters for 3-D surface parsing[J]. IEEE Transactions on Neural Networks and Learning Systems, doi: 10.1109/TNNLS.2023.3281871.
[17]	LAVOUÉ G, DUPONT F, BASKURT A. A new CAD mesh segmentation method, based on curvature tensor analysis[J]. Computer-Aided Design, 2005, 37(10): 975-987. doi: 10.1016/j.cad.2004.09.001
[18]	BHOLOWALIA P, KUMAR A. EBK-means: A clustering technique based on elbow method and K-means in WSN[J]. International Journal of Computer Applications, 2014, 105(9): 17-24.
[19]	COHEN-STEINER D, ALLIEZ P, DESBRUN M. Variational shape approximation[C]. Proceedings of ACM SIGGRAPH 2004 Papers, ACM, 2004: 905-914.
[20]	LIANG Y Q, HE F ZH, ZENG X T. 3D mesh simplification with feature preservation based on whale optimization algorithm and differential evolution[J]. Integrated Computer-Aided Engineering, 2020, 27(4): 417-435. doi: 10.3233/ICA-200641
[21]	DU T, INALA J P, PU Y W, et al. InverseCSG: Automatic conversion of 3D models to CSG trees[J]. ACM Transactions on Graphics, 2018, 37(6): 213.
[22]	BORJI A, CHENG M M, HOU Q B, et al. Salient object detection: a survey[J]. Computational Visual Media, 2019, 5(2): 117-150. doi: 10.1007/s41095-019-0149-9
[23]	ATTENE M, FALCIDIENO B, SPAGNUOLO M. Hierarchical mesh segmentation based on fitting primitives[J]. The Visual Computer, 2006, 22(3): 181-193. doi: 10.1007/s00371-006-0375-x
[24]	SHAPIRA L, SHAMIR A, COHEN-OR D. Consistent mesh partitioning and skeletonisation using the shape diameter function[J]. The Visual Computer, 2008, 24(4): 249-259. doi: 10.1007/s00371-007-0197-5
[25]	KATZ S, TAL A. Hierarchical mesh decomposition using fuzzy clustering and cuts[J]. ACM Transactions on Graphics, 2003, 22(3): 954-961. doi: 10.1145/882262.882369
[26]	LAI Y K, HU S M, MARTIN R R, et al. Rapid and effective segmentation of 3D models using random walks[J]. Computer Aided Geometric Design, 2009, 26(6): 665-679. doi: 10.1016/j.cagd.2008.09.007
[27]	YANG H, SHI J N, CARLONE L. TEASER: fast and certifiable point cloud registration[J]. IEEE Transactions on Robotics, 2021, 37(2): 314-333. doi: 10.1109/TRO.2020.3033695
[28]	RODRIGUES R S V, MORGADO J F M, GOMES A J P. A contour-based segmentation algorithm for triangle meshes in 3D space[J]. Computers & Graphics, 2015, 49: 24-35.
[29]	CHEN X B, GOLOVINSKIY A, FUNKHOUSER T. A benchmark for 3D mesh segmentation[J]. ACM Transactions on Graphics, 2009, 28(3): 73.