一种基于贝叶斯理论的高效立体匹配方法

李娇; 钱惟贤; 陈钱; 顾国华; 任建乐

doi:doi:10.3788/lop51.101001

激光与光电子学进展, 2014, 51 (10): 101001, 网络出版: 2014-09-22

一种基于贝叶斯理论的高效立体匹配方法下载： 523次

An Efficient Stereo Matching Method Based on Bayesian Theory

论文大纲

李娇 ^*钱惟贤陈钱顾国华任建乐

作者单位

南京理工大学电光学院, 江苏南京 210094

图像处理稠密立体匹配贝叶斯理论支撑点三角剖分分割 image processing dense stereo matching Bayesian theory support points triangulation segmentation

AI 词云图 AI语音精读 AI语音超短摘要

注：本部分内容由 AI 自动生成，请您知悉。

摘要

在立体匹配中,同时保证算法的精度和速度是一个技术难题。提出了一种基于贝叶斯理论的快速稠密立体匹配算法。将贝叶斯概率分布理论应用到立体匹配问题上，利用MSERDoG 算子提取支撑点,像素灰度值作为匹配代价、固定窗口作为代价聚合对其进行匹配，对匹配好的支撑点进行三角剖分，将支撑点的视差、梯度、三角剖分的线性系数及分割作为计算视差的先验概率条件，从而保证了有效的视差搜索空间，提高了匹配效率。最终通过最小化能量函数获得稠密的视差图。在国际标准Middlebury 平台进行实验，结果表明提出的算法匹配精度高，速度快，误匹配率低，匹配效率高。

Abstract

Ensuring both of the algorithm accuracy and speed is a key technical problem of stereo matching algorithms. A fast dense stereo matching algorithm based on Bayesian is presented. Bayesian probability theory is applied to the stereo matching problem, the support points are extracted with the MSERDoG operator with the pixel gray value as the matching cost and the fixed window as the cost aggregation matched, the matched support points are triangulated, the disparity and gradient of the support points, the formation of the linear coefficient triangulation and segmentation are selected as priori probability conditions, thus ensuring efficient disparity search space and improving the matching efficiency. The dense disparity map is obtained by minimizing the energy function. In experiments with the international standard Middlebury platform, the results show that the proposed algorithm gets matching with high precision, high speed, low mismatch, and high matching efficiency.

参考文献

[1] M Weber, M Humenberger, W Kubinger. A very fast census- based stereo matching implementation on a graphics processing unit [C]. IEEE Computer Vision Workshop, 2009. 786-793.

[2] Kuk- Jin Yoon, In So Kweon. Adaptive support- weight approach for correspondence search [J].IEEE Transaction on Pattern Analysis and Machine Intelligence, 2006,28(4): 650-656.

[3] F Tombari, S Mattoccia, L Di Stefano. Segmentation- based adaptive support for accurate stereo correspondence [C]. IEEE Pacific-Rim Symposium on Image and Video Technology, 2007. 427-438.

[4] S Roy, I Cox. A maximum- flow formulation of the camera stereo correspondence problem [C]. 6th IEEE International Conference on Computer Vision, 1998. 492-499.

[5] Jian Sun, Nanning Zheng, Y H Shum. Stereo matching using belief propagation [C].IEEE Transaction on Pattern Analysis and Machine Intelligence, 2003, 25(7): 787-800.

[6] Baker H. Depth from edge and intensity based stereo[C]. International Joint Conference on Artificial Intelligence, 1981. 631-636.

[7] U Grenander. Lectures on Pattern Theory [M]. New York: Springer-Veflag, 1981.

[8] S Geman, D Geman. Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1984,PAMI-6 (6): 721-741.

[9] P N Belhumeur. A bayesian-approach to binocular stereopsis [J]. International Journal of Computer vision,1996, 19(3): 237-260.

[10] I J Cox, S L Hingorani, S B Rao. A maximum- likelihood stereo algorithm [J]. Computer Vision and Image Understanding, 1996, 63(3): 542-567.

[11] D Comaniciu, P Meer. Robust analysis of feature spaces: color image segmentation [C]. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1997. 750-755.

[12] 李鹤喜,孙玲云.基于彩色图像分割的置信传播快速立体匹配算法[J].数字技术与应用, 2013, (1): 115-117

[13] P Cignonni, C Montain, R Scopigno. A Fast Divide&Conquer Delaunay Triangulation Algorithm in Ed [J]. IEEE,1997 ,30(5): 333-341.

[14] P Su, R L Scot, Drysdale. A comparison of sequential delaunay traingulation algorithms [C]. 11th ACM Computational Geometry Conference, 1995. 61-70.

[15] Y Boykov, O Veksler. Markov random field with efficient approximation [C]. IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1998. 648-655.

[16] A Geiger, M Roser, R Urtasun. Efficient large- scale stereo matching [C]. 10th Asian conference on Computer Vision, 2011. 25-38.

[17] 杨奎, 赵剡, 苏庆华.基于递推自适应权重的快速稠密立体匹配[J].北京航空航天大学学报, 2013, 39(7): 963-967.

Yang Kui, Zhao Yan, Su Qinghua. Matching based on recursive adaptive weight fast dense stereo [J]. Journal of Beijing University of Aeronautics and Astronautics, 2013, 39(7): 963-967.

[18] 郭大波, 卢朝阳, 何华君.一种新的立体视差估计算法[J].西安电子科技大学学报, 2007, 34(3): 337-341.

Guo Dabo, Lu Zhaoyang, He Huajun. A new disparity estimation algorithm [J]. Journal of Xidian University, 2007, 34(3): 337-341.

[19] L Cheng, T Caelli. Bayesian stereo matching [J]. Computer Vision and Image Understanding, 2006, 106(1): 85-96.

[20] 顾骋, 钱惟贤, 陈钱, 等. 基于双目立体视觉的快速人头检测方法[J].中国激光, 2014, 41(1): 0108001.

Gu Cheng, Qian Weixian, Chen Qian, et al.. Rapid head detection methods based on binocular stereo vision [J]. Chinese J Lasers, 2014, 41(1): 0108001.

[21] 高俊钗, 雷志勇, 王泽民.线阵相机的图像校正[J].激光与光电学进展, 2010, 47(9): 091501.

Gao Junchai, Lei Zhiyong, Wang Zemin. Image correction for linear array camera [J]. Laser & Optoelectronics Progress, 2010, 47(9): 091501.

李娇, 钱惟贤, 陈钱, 顾国华, 任建乐. 一种基于贝叶斯理论的高效立体匹配方法[J]. 激光与光电子学进展, 2014, 51(10): 101001. Li Jiao, Qian Weixian, Chen Qian, Gu Guohua, Ren Jianle. An Efficient Stereo Matching Method Based on Bayesian Theory[J]. Laser & Optoelectronics Progress, 2014, 51(10): 101001.

一种基于贝叶斯理论的高效立体匹配方法下载： 523次

关于本站 Cookie 的使用提示

全站搜索

一种基于贝叶斯理论的高效立体匹配方法 下载： 523次

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索

一种基于贝叶斯理论的高效立体匹配方法下载： 523次