董惠心,任鹏,余兴瑞,王廷伟

• 1:中国石油大学(华东)信息与控制工程学院

摘要(Abstract)：

提出一种基于"客观度量"和"深度学习"共同驱动的立体匹配方法,互补"度量"和"学习"特征,提升立体匹配视差图的精度。将基于灰度差绝对和(SAD)与灰度梯度差绝对和(GRAD)两类算子的客观计算特征和基于数据驱动的深度学习特征进行加权融合,构建匹配代价模型;采用引导滤波器对匹配代价进行聚合;通过胜者全赢算法得到初始视差图;最后,运用左右一致性校验和加权中值滤波器优化视差图,去除误匹配点,得到最优视差图。在Middlebury立体匹配评估平台上的测试实验表明,所提算法能有效降低视差图平均绝对误差和均方根误差。

关键词(KeyWords)： 立体匹配;深度学习;特征融合;引导滤波器;胜者全赢算法;视差图

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金(61671481);; 青岛市自主创新计划应用基础研究项目(16-5-1-11-jch)~~

作者(Author): 董惠心,任鹏,余兴瑞,王廷伟

DOI: 10.16652/j.issn.1004-373x.2018.01.014

参考文献(References)：

• [1]ZITNICK C L,KANG S B.Stereo for image-based rendering using image over-segmentation[J].International journal of computer vision,2007,75(1):49-65.
• [2]孙文健,肖渤涛,骆春波,等.一种基于Panda Board的运动物体检测系统设计[J].现代电子技术,2015,38(2):104-107.SUN Wenjian,XIAO Botao,LUO Chunbo.Design of a moving object detection system based on Panda Board[J].Modern electronics technique,2015,38(2):104-107.
• [3]SCHARSTEIN D,SZELISKI R.A taxonomy and evaluation of dense two-frame stereo correspondence algorithms[J].International journal of computer vision,2002,47(1/3):7-42.
• [4]BOYKOV Y,VEKSLER O,ZABIH R.Fast approximate energy minimization via graph cuts[J].IEEE transactions on pattern analysis and machine intelligence,2001,23(11):1222-1239.
• [5]FELZENSZWALB P F,HUTTENLOCHER D P.Efficient belief propagation for early vision[J].International journal of computer vision,2006,70(1):41-54.
• [6]FORSTMANN S,KANOU Y,OHYA J,et al.Real-time stereo by using dynamic programming[C]//Proceedings of 2004 Conference on Computer Vision and Pattern Recognition Workshop.Washington DC:IEEE,2004:29.
• [7]KLAUS A,SORMANN M,KARNER K.Segment-based stereo matching using belief propagation and a self-adapting dissimilarity measure[C]//Proceedings of the 18th International Conference on Pattern Recognition.Hongkong,China:IEEE,2006:15-18.
• [8]KANADE T,KANO H,KIMURA S,et al.Development of a video-rate stereo machine[C]//Proceedings of 1995 IEEE/RSJ International Conference on Intelligent Robots and Systems.Pittsburgh:IEEE,1995:3095-4100.
• [9]何人杰.双目立体视觉区域局部匹配算法的改进及其实现[J].现代电子技术,2009,32(12):68-70.HE Renjie.Improvement of regional related match algorithm for binocular stereo vision and its implementation[J].Modern electronics technique,2009,32(12):68-70.
• [10]SCHARSTEIN D.Matching images by comparing their gradient fields[C]//Proceedings of 1994 the 12th IAPR International Conference on Pattern Recognition.Jerusalem:IEEE,1994:572-575.
• [11]ZBONTAR J,LECUN Y.Stereo matching by training a convolutional neural network to compare image patches[J].Journal of machine learning research,2016,17:1-32.
• [12]HE K,SUN J,TANG X.Guided image filtering[J].IEEE transactions on pattern analysis and machine intelligence,2013,35(6):1397-1409.
• [13]HOSNI A,RHEMANN C,BLEYER M,et al.Fast cost-volume filtering for visual correspondence and beyond[J].IEEE tran-sactions on pattern analysis and machine intelligence,2013,35(2):504-511.
• [14]张浩峰,赵春霞.一种利用动态规划和左右一致性的立体匹配算法[J].中国图象图形学报,2008(4):756-760.ZHANG Haofeng,ZHAO Chunxia.A stereo matching algorithm using dynamic programming and left-right consistency[J].Journal of image and graphics,2008(4):756-760.
• [15]SCHARSTEIN D,SZELISKI R.The Middlebury stereo vision page[EB/OL].[2014-10-23].http://vision.middlebury.edu/stereo/eval3/,2014.