© 2015 Y. Fan; H. Cheng; B. Bao Xing; Z. Chao; L. Wen Jing
University School of Computer Science and Technology, Changchun, China University of Science and Technology, Changchun, China
In order to improve the efficiency of matching marked points,the accuracy and automation of point cloud stitching, a stitching algorithm of dynamic hierarchy of the Iterative Closest Point is proposed. Firstly, the dynamic distance matrix was introduced to record the distance of hierarchical searching Marked Point; to complete the coarse stitching, matching the marked point through the dynamic distance matrix and using Least Square method to resolve the transformation matrix. Secondly, at the stage of precisely stitching dynamic hierarchical search points set was taken to initialize a valid initial position for the Iterative Closest Point algorithm, then the local optimum of the Iterative Closest Point algorithm was avoided. In three dimensional stitching experiment the precise stitching distance error has reached 0.0085 mm; the method is testified to be simple, practical and characterized by high stitching precision.
Keywords: stitching, computer vision, marked point, dynamic hierarchy, Iterative Closest Point.
1. Yang Biao M., Yue Xian Z., Ling Z., Chao Long Y. Design and recognition of coded targets for 3D registration // Journal of Tsinghua University (Science & Technology). 2006. V. 46. № 2. P. 169–171.
2. Yun G., Jun Hua S., Qian Zhe L., Zhen L., Guang Jun Z. 3D matching of mark points for free-form surface vision measurement // Measurement and Control Technology Journal. 2011. V. 30. № 7. P. 105–109.
3. Yun Bo L., Wen Yi D., Xiao Ping L., Nai Guang L. Automatic registration method of multi-view 3D data based on marked points // Journal of Beijing Information Science and Technology University. 2010. V. 25. № 1. P. 30–34.
4. Xiao Dong X., Yi Z., Cong Xin L. The technique about registration of multi-view data in the optical digitizing measuring and algorithm realization // Machine Tool and Hydraulics. 2005. V. 10. P. 137–140.
5. Hu W., Sheng Lan L., Li Yan Z., Hui Z. Reference Point Based Registration Algorithm for Binocular Stereo Vision System // China Mechanical Engineering. 2009. V. 20. № 14. P. 1736–1740.
6. Bai Song C., Xue Mei Y., Li A. Minimum bounding box calculation based on nonlinear principle component analysis // Computer Integrated Manufacturing Systems. 2010. V 16. № 11. P. 2375–2378.
7. Horn B.K.P. Closed form solution of absolute orientation using unit quaternions // Journal of the Optical Society of America A. 1987. V. 4. № 4. P. 629–642.
8. Arun K.S., Huang T.S., Blostein S.D. Least square fitting of two 32D point sets // IEEE Transactions on Pattern Analysis and Machine Intelligence. 1987. V. 9. №. 5. P. 698–700.
9. Hai Ping S., Fei Peng D., Jia Yong L. Research of point-clouds registration based on least-square method // Journal of Image and Graphics. 2005. № 9. P. 1112–1116.
10. Besl P.J., Mckay N.D. A method for registration of 3D shapes // IEEE Transactions on Pattern Analysis and Machine Intelligence. 1992. V. 14. № 2. P. 239–256.
11. Bao Quan S., Jin L., Qing L., Zhen Zhong X. Precision inspection of point cloud & CAD model based on constraint search sphere // Computer Integrated Manufacturing Systems. 2010. V. 16. № 5. P. 929–934.
12. Umeyama S. Least-squares estimation of transformation parameters between two point patterns // IEEE Transactions on Pattern Analysis and Intelligence. 1991. V. 13. № 4. P. 376–380.
13. Xiang Bo O., Zhi Jian Z., Hui Yuan X. An automatic method for Point-clouds based on marked points // Journal of Image and Graphics. 2008. V. 13. № 2. P. 298–301.
Полный текст >>>