摘要
针对立体匹配中的代价聚集问题,提出一种改进的代价聚集算法。对图像进行超像素分割并建立最小生成树,采用树形滤波器进行代价聚集,生成超像素代价。利用权重融合初始像素代价和超像素代价,得到最终的像素代价。实验结果表明,与最小生成树、线段树等算法相比,该算法的时间复杂度较低,且生成的视差图具有良好的边缘保持特性。
For the cost aggregation problem in stereo matching,an improved cost aggregation algorithm is proposed.The image is segmented by superpixel and the Minimum Spanning Tree(MST) is established.Then,use tree filter for cost aggregation to generate superpixel cost.The final pixel cost is obtained by weighting the initial pixel cost and superpixel cost.Experimental results show that compared with MST,Segmcat-Tree(ST) and other methods,the proposed algorithm has lower time complexity,and can get disparity map with good edge preserve features.
引文
[1] SUN Jian,SHUM Heung Yeung,ZHENG Nanning.Stereo matching using belief propagation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2003,25(7):787-800.
[2] HIRSCHMULLER H.Accurate and efficient stereo processing by semi-global matching and mutual information[C]//Proceedings of 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Computer Society,2005:807-814.
[3] 王新艳,潘巍,王月莲,等.基于全局差错能量函数的立体匹配算法[J].计算机工程,2017,43(7):244-249.
[4] 王年,范益政,鲍文霞,等.基于图割的图像匹配算法[J].电子学报,2006,34(2):232-236.
[5] 王召月,陈丽芳.基于mean-shift全局立体匹配方法[J].计算机工程与科学,2017,39(7):1333-1337.
[6] HONG Li,CHEN George.Segment-based stereo matching using graph cuts[C]//Proceedings of 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Computer Society,2004:74-81.
[7] SCHARSTEIN D,SZELISKI R.A taxonomy and evaluation of dense two-frame stereo correspondence algorithms[J].International Journal of Computer Vision,2002,47(1/2/3):7-42.
[8] PORIKLI F.Constant time O(1) bilateral filtering[C]//Proceedings of International Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2008:1-8.
[9] HE Kaiming,SUN Jian,TANG Xiao’ou.Guided image filtering[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2013,35(6):1397-1409.
[10] LU Jiangbo,SHI Keyang,MIN Dongbo,et al.Cross-based local multipoint filtering[C]//Proceedings of International Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2012:430-437.
[11] YANG Qingxiong.A non-local cost aggregation method for stereo matching[C]//Proceedings of International Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2012:1402-1409.
[12] MEI Xing,SUN Xun,DONG Weiming,et al.Segment-tree based cost aggregation for stereo matching[C]//Proceedings of International Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2013:313-320.
[13] 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):2287-2318.
[14] LI Lincheng,YU Xin,ZHANG Shunli,et al.3D cost aggregation with multiple minimum spanning trees for stereo matching[J].Applied Optics,2017,56(12):3411-3420.
[15] LEE S,LEE J H,LIM J,et al.Robust stereo matching using adaptive random walk with restart algorithm[J].Image and Vision Computing,2015,37:1-11.
[16] ACHANTA R,SHAJI A,SMITH K,Et al.SLIC superpixels compared to state-of-the-art superpixel methods[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(11):2274-2282.
[17] MEI Xing,SUN Xun,ZHOU Mingcai,et al.On building an accurate stereo matching system on graphics hardware[C]//Proceedings of International Conference on Computer Vision Workshops.Washington D.C.,USA:IEEE Press,2011:467-474.