Affine Registration Based on Chord Height Point and Genetic Algorithm
-
摘要: 针对复杂场景中目标由于成像畸变、部分遮挡和局部缺失难于识别的难题, 提出了一种新的特征点——弦高点, 将其和遗传算法相结合用于图像的仿射配准. 算法首先给出了弦高点的定义, 并证明了其仿射不变性; 然后,应用遗传算法搜索模型和目标轮廓上两对对应点, 以弦高点作为第三对对应点, 求解最优的仿射变换矩阵; 最后,对遗传算法搜索的结果再进行线性搜索, 提高配准的精度. 本文利用 LTS Hausdorff距离(Least trimmed square Hausdorff distance, LTS-HD) 进行度量, 能有效克服部分遮挡或局部缺失的影响. 由于采用遗传算法, 并只需搜索两对对应点, 配准的速度得到提高. 理论分析和实验结果均表明, 该算法能有效地进行仿射配准, 并能处理部分遮挡或局部缺失.
-
关键词:
- 仿射配准 /
- LTS Hausdorff距离 /
- 遗传算法 /
- 弦高点
Abstract: It is difficult to recognize objects when they are distorted and partially occluded or broken. In order to solve the problems, a new registration algorithm combining chord height point with genetic algorithm is proposed in this paper. Firstly, we define the chord height point and prove it is affine invariant. Then, the global optimal affine transformation matrix is calculated by using three pairs of corresponding points, where two corresponding points in the contours of model and target are searched by the genetic algorithm, and the chord height point is used as the third pair of corresponding points. Finally, linear search is introduced to search the local optimal affine transform matrix, and the accuracy of registration is improved. We use least trimmed square Hausdorff distance (LTS-HD) to measure the similarity between the model and target, so our method can deal with object partially occlued and broken. Furthermore, the registration speed is improved by using genetic algorithm to search corresponding points and only searching two pairs of corresponding points. The theory analysis and experimental results show that our algorithm can be effectively used for affine registration, and can deal with objects partially occlued and broken. -
[1] Brown L G. A survey of image registration techniques. ACM Computing Surveys, 1992, 24(4): 325-376 [2] Zitová B, Flusser J. Image registration methods: a survey. Image and Vision Computing, 2003, 21(11): 977-1000 [3] Lu X S, Zhang S, Su H, Chen Y Z. Mutual information-based multimodal image registration using a novel joint histogram estimation. Computerized Medical Imaging and Graphics, 2008, 32(3): 202-209 [4] Lee J H, Kim Y S, Lee D, Kang D G, Ra J B. Robust CCD and IR image registration using gradient-based statistical information. IEEE Signal Processing Letters, 2010, 17(4): 347-350 [5] Zhang X Q, Men T, Liu C, Yang J. Infrared and visible images registration using BEMD and MI. In: Proceedings of the 3rd IEEE International Conference on Computer Science and Information Technology. Chengdu, China: IEEE, 2010. 644-647 [6] Besl P J, McKay H D. A method for registration of 3-D shapes. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(2): 239-256 [7] Sharp G C, Lee S W, Wehe D K. ICP registration using invariant features. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(1): 90-102 [8] Liu Y H. Improving ICP with easy implementation for free-form surface matching. Pattern Recognition, 2004, 37(2): 211-226 [9] Hrkać T, Kalafatić Z, Krapac J. Infrared-visual image registration based on corners and Hausdorff distance. Lecture Notes in Computer Science. Berlin: Springer-Verlag, 2007. 383-392 [10] Gao Feng, Wen Gong-Jian, Lv Jin-Jian. An optimal algorithm for IR/visual image registration based on main-line-pairs. Chinese Journal of Computers, 2007, 30(6): 1014-1021(高峰, 文贡坚, 吕金建. 基于干线对的红外与可见光最优图像配准算法. 计算机学报, 2007, 30(6): 1014-1021) [11] Su Juan, Lin Xing-Gang, Liu Dai-Zhi. A multi-sensor image registration algorithm based on structure feature edges. Acta Automatica Sinica, 2009, 35(3): 251-257(苏娟, 林行刚, 刘代志. 一种基于结构特征边缘的多传感器图像配准方法. 自动化学报, 2009, 35(3): 251-257) [12] Zhang Xiu-Wei, Zhang Yan-Ning, Yang Tao, Zhang Xin-Gong, Shao Da-Pei. Automatic visual-thermal image sequence registration based on co-motion. Acta Automatica Sinica, 2010, 36(9): 1220-1231(张秀伟, 张艳宁, 杨涛, 张新功, 邵大培. 基于Co-motion的可见光—热红外图像序列自动配准算法. 自动化学报, 2010, 36(9): 1220-1231) [13] Bilodeau G A, Torabi A, Morin F. Visible and infrared image registration using trajectories and composite foreground images. Image and Vision Computing, 2011, 29(1): 41-50 [14] Lian Lin, Li Guo-Hui, Zhang Jun, Tu Dan. An automatic registration algorithm of infrared and visible images based on optimal mapping of edges. Acta Automatica Sinica, 2012, 38(4): 570-581(廉蔺, 李国辉, 张军, 涂丹. 基于边缘最优映射的红外和可见光图像自动配准算法. 自动化学报, 2012, 38(4): 570-581) [15] Tsang P W M. A genetic algorithm for aligning object shapes. Image and Vision Computing, 1997, 15(11): 819-831 [16] Tsang P W M, Yuen T Y F. Affine invariant matching of broken boundaries based on an enhanced genetic algorithm and distance transform. IET Computer Vision, 2008, 2(3): 142-149 [17] Tsang P W M, Situ W C. Affine invariant matching of broken boundaries based on simple genetic algorithm and contour reconstruction. Pattern Recognition Letters, 2010, 31(9): 771-780 [18] Sim D G, Kwon O K, Park R H. Object matching algorithms using robust Hausdorff distance measures. IEEE Transactions on Image Processing, 1999, 8(3): 425-429 [19] Holland J H. Adaptation in Natural and Artificial Systems. Ann Arbor: University of Michigan Press, 1975 [20] Mei Xiang-Ming, Liu Zeng-Xian, Wang Hui-Chun, Wang Zhi-Qiu. Higher Geometry (Third edition). Beijing: Higher Education Press, 2008. 13 (梅向明, 刘增贤, 王汇淳, 王智秋. 高等几何 (第3版). 北京: 高等教育出版社, 2008. 13)
点击查看大图
计量
- 文章访问数: 1488
- HTML全文浏览量: 80
- PDF下载量: 1310
- 被引次数: 0