Research Papers

Improvements to the Iterative Closest Point Algorithm for Shape Registration in Manufacturing

[+] Author and Article Information
Tsz-Ho Kwok

Department of Mechanical and
Aerospace Engineering,
The Hong Kong University of
Science and Technology,
Clear Water Bay, Hong Kong
e-mail: tom.thkwok@gmail.com

Kai Tang

Department of Mechanical and
Aerospace Engineering,
The Hong Kong University of
Science and Technology,
Clear Water Bay, Hong Kong

1Corresponding author.

Contributed by the Manufacturing Engineering Division of ASME for publication in the JOURNAL OF MANUFACTURING SCIENCE AND ENGINEERING. Manuscript received November 1, 2014; final manuscript received August 7, 2015; published online September 9, 2015. Assoc. Editor: Dragan Djurdjanovic.

J. Manuf. Sci. Eng 138(1), 011014 (Sep 09, 2015) (7 pages) Paper No: MANU-14-1575; doi: 10.1115/1.4031335 History: Received November 01, 2014; Revised August 07, 2015

Iterative closest point (ICP) is a popular algorithm used for shape registration while conducting inspection during a production process. A crucial key to the success of the ICP is the choice of point selection method. While point selection can be customized for a particular application using its prior knowledge, normal-space sampling (NSS) is commonly used when normal vectors are available. Normal-based approach can be further improved by stability analysis—called covariance sampling. The stability analysis should be accurate to ensure the correctness of covariance sampling. In this paper, we go deep into the details of covariance sampling, and propose a few improvements for stability analysis. We theoretically and experimentally show that these improvements are necessary for further success in covariance sampling. Experimental results show that the proposed method is more efficient and robust for the ICP algorithm.

Copyright © 2016 by ASME
Your Session has timed out. Please sign back in to continue.


Xu, K. , and Chen, Y. , 2015, “ Mask Image Planning for Deformation Control in Projection-Based Stereolithography Process,” ASME J. Manuf. Sci. Eng, 137(3), p. 031014. [CrossRef]
Song, X. , Pan, Y. , and Chen, Y. , 2015, “ Development of a Low-Cost Parallel Kinematic Machine for Multidirectional Additive Manufacturing,” ASME J. Manuf. Sci. Eng, 137(2), p. 021005. [CrossRef]
Besl, P. J. , and McKay, N. D. , 1992, “ A Method for Registration of 3-D Shapes,” IEEE Trans. Pattern Anal. Mach. Intell., 14(2), pp. 239–256. [CrossRef]
Rusinkiewicz, S. , and Levoy, M. , 2001, “ Efficient Variants of the ICP Algorithm,” Third International Conference on 3D Digital Imaging and Modeling (3DIM), Quebec City, Canada, pp. 145–152.
Turk, G. , and Levoy, M. , 1994, “ Zippered Polygon Meshes From Range Images,” 21st Annual Conference on Computer Graphics and Interactive Techniques, SIGGRAPH’94, pp. 311–318.
Masuda, T. , Sakaue, K. , and Yokoya, N. , 1996, “ Registration and Integration of Multiple Range Images for 3-D Model Construction,” 13th International Conference on Pattern Recognition, Vienna, Austria, Aug. 25–29, Vol. 1, pp. 879–883.
Weik, S. , 1997, “ Registration of 3-D Partial Surface Models Using Luminance and Depth Information,” International Conference on Recent Advances in 3-D Digital Imaging and Modeling, Ottawa, Canada, May 12–15, pp. 93–100.
Gelfand, N. , Ikemoto, L. , Rusinkiewicz, S. , and Levoy, M. , 2003, “ Geometrically Stable Sampling for the ICP Algorithm,” Fourth International Conference on 3-D Digital Imaging and Modeling (3DIM), Oct. 6–10, pp. 260–267.
Gelfand, N. , Mitra, N. J. , Guibas, L. J. , and Pottmann, H. , 2005, “ Robust Global Registration,” Third Eurographics Symposium on Geometry Processing, SGP’05, Article 197.
Kwok, T.-H. , Yeung, K.-Y. , and Wang, C. C. L. , 2014, “ Volumetric Template Fitting for Human Body Reconstruction From Incomplete Data,” J. Manuf. Syst., 33(4), pp. 678–689. [CrossRef]
Salvi, J. , Matabosch, C. , Fofi, D. , and Forest, J. , 2007, “ A Review of Recent Range Image Registration Methods With Accuracy Evaluation,” Image Vision Comput., 25(5), pp. 578–596. [CrossRef]
Pottmann, H. , and Leopoldseder, S. , 2003, “ A Concept for Parametric Surface Fitting Which Avoids the Parametrization Problem,” Comput. Aided Geom. Design, 20(6), pp. 343–362. [CrossRef]
Chen, Y. , and Medioni, G. , 1992, “ Object Modeling by Registration of Multiple Range Images,” Image Vision Comput., 10(3), pp. 145–155. [CrossRef]
Mitra, N. J. , Gelfand, N. , Pottmann, H. , and Guibas, L. , 2004, “ Registration of Point Cloud Data From a Geometric Optimization Perspective,” Eurographics/ACM SIGGRAPH Symposium on Geometry Processing, SGP’04, pp. 22–31.
Wang, W. , Pottmann, H. , and Liu, Y. , 2006, “ Fitting B-Spline Curves to Point Clouds by Curvature-Based Squared Distance Minimization,” ACM Trans. Graph., 25(2), pp. 214–238. [CrossRef]
Gelfand, N. , and Guibas, L. J. , 2004, “ Shape Segmentation Using Local Slippage Analysis,” Eurographics/ACM SIGGRAPH Symposium on Geometry Processing, SGP’04, pp. 214–223.
Bokeloh, M. , Berner, A. , Wand, M. , Seidel, H.-P. , and Schilling, A. , 2008, Slippage Features, Technical Report WSI-2008-03, University of Tübingen, WSI/GRIS.
Gatzke, T. , Grimm, C. , Garland, M. , and Zelinka, S. , 2005, “ Curvature Maps for Local Shape Comparison,” International Conference on Shape Modeling and Applications, SMI’05, June 13–17, pp. 246–255.
Tevs, A. , Berner, A. , Wand, M. , Ihrke, I. , and Seidel, H.-P. , 2011, “ Intrinsic Shape Matching by Planned Landmark Sampling,” Comput. Graphics Forum, 30(2), pp. 543–552. [CrossRef]
Sun, J. , Ovsjanikov, M. , and Guibas, L. , 2009, “ A Concise and Provably Informative Multi-Scale Signature Based on Heat Diffusion,” Symposium on Geometry Processing, pp. 1383–1392.
Godil, A. , and Wagan, A. I. , 2011, “ Salient Local 3D Features for 3D Shape Retrieval,” Proc. SPIE, 7864, pp. 78640S–78640S–8.
Heider, P. , Pierre-Pierre, A. , Li, R. , and Grimm, C. , 2011, “ Local Shape Descriptors, a Survey and Evaluation,” 4th Eurographics Conference on 3D Object Retrieval, EG 3DOR’11, pp. 49–5 6 .
Song, R. , Liu, Y. , Martin, R. R. , and Rosin, P. L. , 2012, “ Saliency-Guided Integration of Multiple Scans,” IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1474–1481.
Zheng, Y. , Lin, M. C. , and Manocha, D. , 2011, “ Efficient Simplex Computation for Fixture Layout Design,” Comput. Aided Des., 43(10), pp. 1307–1318. [CrossRef]


Grahic Jump Location
Fig. 1

The contact surfaces of these mechanical parts are not completely constrained, i.e., they have certain DOF

Grahic Jump Location
Fig. 2

These two parts are mechanically constrained (left), where their contact surface (right) is analyzed. (Top-right) Before optimizing the rotation center position (located at theintersection of the three rotation axes), the stability of the y-axis is reported as high. (Bottom-right) After updating the rotation center, the analysis gives a much lower stability for they-axis.

Grahic Jump Location
Fig. 3

Two examples of point extraction: a cube (top) and a blade (bottom). From left to right: the input with the candidate points, and the extracted points in three cases with differentstabilities of 40.0, 20.0, and 10.0 for cube; 8.0, 4.0, 2.0 for blade.

Grahic Jump Location
Fig. 4

Comparison between the original covariance sampling and the improved one proposed in this paper. Three examples are shown in this figure, and each of them shows a chart of RMS alignment error against the iteration steps. From the rabbit and rocker arm examples, we can see that the improved version converges more than two times faster than the original one. In the chair example, even though both methods have the same minimum stability, the original method fails to converge while the improved one does.

Grahic Jump Location
Fig. 5

Comparisons between our method and the HKS method, the NSS method, and the 3D-SIFT method. In the tests on both objects, our method performs much better than the rest.

Grahic Jump Location
Fig. 6

Three range surfaces capture different portions of a teapot, and their data are overlapped only partially

Grahic Jump Location
Fig. 7

Comparison between different levels of Gaussian noise is added in terms of transformation error. The amount of noise is related to the percentage of the model's bounding ball radius.




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In