• 北大核心期刊(《中文核心期刊要目总览》2017版)
  • 中国科技核心期刊(中国科技论文统计源期刊)
  • JST 日本科学技术振兴机构数据库(日)收录期刊

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于改进Census变换与颜色梯度融合的立体匹配

刘建明 何晴 陈辉

刘建明, 何晴, 陈辉. 基于改进Census变换与颜色梯度融合的立体匹配[J]. 微电子学与计算机, 2021, 38(9): 38-44.
引用本文: 刘建明, 何晴, 陈辉. 基于改进Census变换与颜色梯度融合的立体匹配[J]. 微电子学与计算机, 2021, 38(9): 38-44.
LIU Jianming, HE Qing, CHEN Hui. Stereo matching based on improved Census transform and color gradient fusion[J]. Microelectronics & Computer, 2021, 38(9): 38-44.
Citation: LIU Jianming, HE Qing, CHEN Hui. Stereo matching based on improved Census transform and color gradient fusion[J]. Microelectronics & Computer, 2021, 38(9): 38-44.

基于改进Census变换与颜色梯度融合的立体匹配

基金项目: 

国家自然科学基金 61262074

详细信息
    作者简介:

    刘建明  男,(1975-),博士,教授.研究方向为计算机通信网络组网工程、综合智能系统

    何晴  女,(1994-),硕士研究生.研究方向为智能信息处理

    通讯作者:

    陈辉(通讯作者)  男,(1974-),博士,副高级研究员.研究方向为机器学习、机器视觉、智能检测与智能控制非线性控制、物联网、模式识别与图像处理、图像处理、无人驾驶技术.E-mail: 1061790644@qq.com

  • 中图分类号: TP391

Stereo matching based on improved Census transform and color gradient fusion

  • 摘要: 针对立体匹配中传统Census变换窗口中心点易受外界环境影响以及部分深度不连续区域匹配精度较低的问题,提出了一种基于改进的Census变换和传播滤波的立体匹配算法.在初始匹配代价计算中将改进的Census变换与颜色,梯度代价进行融合;同时在代价聚合阶段引入传播滤波来保持视差空间图像边缘,不受传统局部算法窗口大小的影响;然后在视差处理部分采用“胜者为王”算法进行初始视差计算;在后面的视差优化部分采用左右一致性检测和中值滤波的方法来获得最终的视差图.在Middleburry上的测试实验表明:与传统的Census算法相比,本文算法匹配精度明显提高,且具有良好的实时性和很好的稳健性.
  • 图  1  中心像素不同时的像素值

    图  2  初始匹配代价的计算

    图  3  d=5时窗口大小为9时原图和各个算法代价聚合图

    图  4  结果视差对比图

    表  1  图像密集训练下错误像素百分比为1的测试数据集比较结果

    匹配图像 MANE FC-DCNN GANetREE_RVC Proposed
    Training dense nonocc all nonocc all nonocc all nonocc all
    Tsukuba 43.3 47.7 31.3 28.0 26.7 29.0 21.6 23.1
    Venus 21.5 37.9 17.1 13.4 11.6 19.4 12.2 14.5
    Teddy 42.3 55.3 38.5 27.1 51.5 55.9 21.4 36.4
    Cones 38.7 45.1 29.8 27.7 25.1 29.5 16.2 17.8
    MotorcycleE 36.9 43.5 29.5 27.6 18.4 22.2 16.0 17.7
    Piano 44.6 48.9 36.2 28.4 23.7 27.0 25.0 27.2
    PianoL 55.4 58.8 47.5 26.9 36.4 38.8 39.0 43.2
    Pipes 32.9 44.1 25.2 21.2 19.6 27.6 13.9 16.8
    Playroom 56.6 62.5 42.8 32.5 31.6 36.8 33.1 36.2
    Playtable 69.5 72.6 48.6 32.2 25.4 27.8 39.6 41.5
    PlaytableP 38.8 45.2 31.5 26.3 18.4 22.5 17.1 18.6
    Recycle 41.6 45.8 31.8 27.6 22.7 24.5 24.9 26.5
    Shelves 68.4 70.4 59.7 43.2 35.3 37.3 47.1 51.6
    Teddy 16.5 25.1 12.5 10.3 11.0 13.7 10.0 11.2
    Vintage 62.3 65.4 51.5 34.7 41.6 44.8 35.7 41.1
    Weight Avg 41.0 48.3 32.6 25.8 25.1 29.1 24.8 26.2
    下载: 导出CSV

    表  2  密集训练下图像视差的误匹配率和时长

    Training dense Adirondack Motorcycle Piano Teddy
    time avgerr time avgerr time avgerr time avgerr
    MANE 0.09 11.6 0.02 12.4 0.02 15.1 0.09 12.5
    FC-DCNN 10.9 2.87 10.9 4.65 10.1 4.45 29.6 2.17
    GANetREE_RVC 8.87 1.09 9.94 1.41 8.86 1.33 3.70 0.92
    Proposed 0.960 1.55 1.27 1.66 1.15 2.05 0.46 1.16
    下载: 导出CSV
  • [1] DO P N B, NGUYEN Q C. A review of stereo-photogrammetry method for 3-D reconstruction in computer vision[C]//Proceedings of the 2019 19th International Symposium on Communications and Information Technologies. Ho Chi Minh City, Vietnam: IEEE, 2019: 138-143. DOI: 10.1109/ISCIT.2019.8905144.
    [2] SCHARSTEIN D, SZELISKI R. A taxonomy and evaluation of dense two-frame stereo correspondence algorithms[J]. International Journal of Computer Vision, 2002, 47(1): 7-42. DOI:  10.1023/A:1014573219977.
    [3] SHI C B, WANG G J, YIN X W, et al. High-accuracy stereo matching based on adaptive ground control points[J]. IEEE Transactions on Image Processing, 2015, 24(4): 1412-1423. DOI:  10.1109/TIP.2015.2393054.
    [4] ABREU B, PAIM G, GRELLERT M, et al. Exploiting absolute arithmetic for power-efficient sum of absolute differences[C]//Proceedings of the 2017 24th IEEE International Conference on Electronics, Circuits and Systems. Batumi, Georgia: IEEE, 2017. DOI: 10.1109/ICECS.2017.8292062.
    [5] HISHAM M B, YAAKOB S N, RAOF R A A, et al. Template matching using sum of squared difference and normalized cross correlation[C]//Proceedings of 2015 IEEE Student Conference on Research and Development. Kuala Lumpur, Malaysia: IEEE, 2015. DOI: 10.1109/SCORED.2015.7449303.
    [6] WU Q, LAW R. The forecasting model based on modified SVRM and PSO penalizing Gaussian noise[J]. Expert Systems with Applications, 2011, 38(3): 1887-1894. DOI:  10.1016/j.eswa.2010.07.120.
    [7] FIFE W S, ARCHIBALD J K. Improved census transforms for resource-optimized stereo vision[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2013, 23(1): 60-73. DOI:  10.1109/TCSVT.2012.2203197.
    [8] HIRSCHMULLER H, SCHARSTEIN D. Evaluation of cost functions for stereo matching[C]//Proceedings of 2007 IEEE Conference on Computer Vision and Pattern Recognition. Minneapolis: IEEE, 2007: 1-8. DOI: 10.1109/CVPR.2007.383248.
    [9] LIM J, KIM Y, LEE S. A census transform-based robust stereo matching under radiometric changes[C]//Proceedings of 2016Asia-Pacific Signal and Information Processing Association Annual Summit and Conference. Jeju, Korea (South): IEEE, 2016. DOI: 10.1109/APSIPA.2016.7820841.
    [10] 萧红, 田川, 张毅, 等. 基于改进Census变换与梯度融合的立体匹配算法[J]. 激光与光电子学进展, 2021, 58(2): 0215008. DOI:  10.3788/LOP202158.0215008.

    XIAO H, TIAN C, ZHANG Y, et al. Stereo matching algorithm based on improved Census transform and gradient fusion[J]. Laser & Optoelectronics Progress, 2021, 58(2): 0215008. DOI:  10.3788/LOP202158.0215008.
    [11] CHANG J H R, WANG Y C F. Propagated image filtering[C]//Proceedings of 2015 IEEE Conference on Computer Vision andPattern Recognition. Boston: IEEE, 2015. DOI: 10.1109/CVPR.2015.7298595.
    [12] FORERO M G, RESTREPO-TABORDA C A. Generalization of propagated image filtering[C]//Proceedings of the 7th Latin American Conference on Networked and Electronic Media. Valparaiso: 2017: 6-8. DOI: 10.1049/ic.2017.0033.
    [13] 程子怡, 卢荣胜, 毛翠丽. 光亮表面双目立体视觉三维形貌测量方法[J]. 激光与光电子学进展, 2020, 57(7): 071202. doi:  10.3788/LOP57.071202

    CHENG Z Y, LU R S, MAO C L. Measurement method of three-dimensional shape of brightsurface with binocularstereo vision[J]. Laser & Optoelectronics Progress, 2020, 57(7): 071202. DOI:  10.3788/LOP57.071202.
    [14] HSIAO S F, CHANG C H. Hardware design of disparity computation for stereo vision using guided image filtering[C]//Proceedings of 2018 International Symposium on VLSI Design, Automation and Test. Hsinchu, China: IEEE, 2018. DOI: 10.1109/VLSI-DAT.2018.8373266.
    [15] KORNPROBST P, TUMBLIN J, DURAND F. Bilateral filtering: theory and applications[J]. Foundations and Trends in Computer Graphics and Vision, 2009, 4(1): 1-73. DOI:  10.1561/0600000020.
    [16] 李婕, 巩朋成. 结合传播滤波的立体匹配算法研究[J]. 计算机工程与应用, 2020, 56(19): 189-196. https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG202019027.htm

    LI J, GONG P C. Cost aggregation based on propagated filtering in stereo matching[J]. Computer Engineering and Applications, 2020, 56(19): 189-196. https://www.cnki.com.cn/Article/CJFDTOTAL-JSGG202019027.htm
    [17] ZEGLAZI O, RZIZA M, AMINE A. An enhanced cross-scale adaptive cost aggregation for stereo matching[C]//Proceedings of 2017 International Conference on Wireless Networks and Mobile Communications. Rabat, Morocco: IEEE, 2017: 1-5. DOI: 10.1109/WINCOM.2017.8238214.
    [18] 欧永东, 谢小鹏. 基于改进Census变换的多特性立体匹配算法[J]. 计算机工程与科学, 2020, 42(6): 1030-1036. DOI:  10.3969/j.issn.1007-130X.2020.06.010.

    OU Y D, XIE X P. A multi-feature stereo matching algorithm based on improved Census transform[J]. Computer Engineering and Science, 2020, 42(6): 1030-1036. DOI:  10.3969/j.issn.1007-130X.2020.06.010.
    [19] SCHARSTEIN D, SZELISKI R, ZABIH R. A taxonomy and evaluation of dense two-framestereo correspondence algorithms[C]//Workshop on Stereo and Multi-Baseline Vision (in conjunction with IEEE CVPR 2001), Kluwer Academic Publishers. 2001: 131-140.
    [20] VAZQUEZ DELGADOH D, PEREZ PATRICIOM, AGUILAR GONZALEZA, et al. Real-time multi-window stereo matching algorithm with fuzzy logic[J]. IET Computer Vision, 2021, 15(3): 208-223. DOI:  10.1049/cvi2.12031.
    [21] HIRNER D, FRAUNDORFER F. FC-DCNN: a densely connected neural network for stereo estimation[C]//Proceedings of the 2020 25th International Conference on Pattern Recognition. Milan, Italy: IEEE, 2020. DOI: 10.1109/ICPR48806.2021.9413281.
    [22] ZHANG F H, PRISACARIU V, YANG R G, et al. GA-Net: guided aggregation net for end-to-end stereo matching[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach: IEEE, 2019. DOI: 10.1109/CVPR.2019.00027.
  • 加载中
图(4) / 表(2)
计量
  • 文章访问数:  38
  • HTML全文浏览量:  18
  • PDF下载量:  3
  • 被引次数: 0
出版历程
  • 收稿日期:  2020-12-20
  • 修回日期:  2021-01-19
  • 刊出日期:  2021-09-05

目录

    /

    返回文章
    返回