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

基于多特征自适应融合的抗遮挡目标跟踪算法

张方方 曹家晖 王海静 赵鹏博

张方方, 曹家晖, 王海静, 赵鹏博. 基于多特征自适应融合的抗遮挡目标跟踪算法[J]. 红外技术, 2023, 45(2): 150-160.
引用本文: 张方方, 曹家晖, 王海静, 赵鹏博. 基于多特征自适应融合的抗遮挡目标跟踪算法[J]. 红外技术, 2023, 45(2): 150-160.
ZHANG Fangfang, CAO Jiahui, WANG Haijing, ZHAO Pengbo. Anti-Occlusion Moving Target Tracking Algorithm Based on Multifeature Self-Adaptive Fusion[J]. Infrared Technology , 2023, 45(2): 150-160.
Citation: ZHANG Fangfang, CAO Jiahui, WANG Haijing, ZHAO Pengbo. Anti-Occlusion Moving Target Tracking Algorithm Based on Multifeature Self-Adaptive Fusion[J]. Infrared Technology , 2023, 45(2): 150-160.

基于多特征自适应融合的抗遮挡目标跟踪算法

基金项目: 

国家自然科学基金项目 62273311

国家自然科学基金项目 62173311

河南省青年人才托举工程项目 2020HYTP006

详细信息
    作者简介:

    张方方(1986-),男,河南人,副教授,博士,从事自主机器人、多机器人控制、多智能体系统等研究,E-mail:zhangfangfang@zzu.edu.cn

  • 中图分类号: TP391

Anti-Occlusion Moving Target Tracking Algorithm Based on Multifeature Self-Adaptive Fusion

  • 摘要: 针对目前的目标跟踪算法在目标发生运动模糊或被遮挡等情况下跟踪效果较差,容易出现跟踪失败等情况,本文提出了一种多特征自适应融合的抗遮挡相关滤波跟踪算法。算法首先提取梯度方向直方图特征HOG和颜色直方图特征,以最大化跟踪质量为目标自适应融合两种特征的相关滤波响应;在跟踪的过程中根据响应图的质量存储高质量滤波模板,采用高质量模板和正常更新模板检测响应图的质量差值来检测目标的遮挡情况,当目标遮挡消失的时候,跟踪器的模板回溯到高质量模板来重新跟踪目标。根据在OTB100、UAV123的实验结果,本文算法相对于其他同类型的相关滤波在跟踪精度和成功率方面表现更好,在发生目标遮挡时仍能很好地跟踪。
  • 图  1  响应图所有的峰值

    Figure  1.  All peaks in response graph

    图  2  自适应调整融合系数

    Figure  2.  Adaptive adjustment of fusion coefficient

    图  3  完整的算法流程

    Figure  3.  Complete algorithm flow

    图  4  OTB100、UAV123数据集中精度排名

    Figure  4.  Accuracy ranking of OTB100 and UAV123 data sets

    图  5  OTB100、UAV123数据集中成功率排名

    Figure  5.  Ranking of success rate of OTB100 and UAV123 data sets

    图  6  OTB100数据中各个模块消融实验的跟踪结果

    Figure  6.  Tracking results of ablation experiment of each module in OTB100 data

    图  7  7种算法在不同视频的关键帧中对比结果

    Figure  7.  Comparison results of seven algorithms in key frames of different videos

    表  1  OTB100的精度

    Table  1.   Accuracy of OTB100

    OUR SRDCFdecon LMCF SRDCF Staple MCCT-H AutoTrack
    IV 0.757 0.785 0.765 0.723 0.686 0.717 0.697
    OPR 0.765 0.712 0.674 0.666 0.674 0.754 0.697
    SV 0.742 0.678 0.664 0.62 0.628 0.693 0.662
    OCC 0.782 0.693 0.708 0.689 0.727 0.736 0.739
    DEF 0.798 0.744 0.753 0.732 0.768 0.793 0.770
    MB 0.719 0.678 0.672 0.708 0.676 0.688 0.731
    FM 0.694 0.714 0.665 0.698 0.627 0.651 0.689
    IPR 0.706 0.627 0.644 0.559 0.519 0.688 0.633
    OV 0.713 0.522 0.614 0.536 0.666 0.632 0.711
    BC 0.766 0.805 0.753 0.684 0.604 0.742 0.650
    LR 0.595 0.543 0.605 0.557 0.588 0.535 0.708
    下载: 导出CSV

    表  2  OTB100成功率

    Table  2.   Success of OTB100

    OUR SRDCFdecon LMCF SRDCF Staple MCCT-H AutoTrack
    IV 0.697 0.721 0.715 0.649 0.627 0.652 0.653
    OPR 0.679 0.654 0.634 0.605 0.593 0.691 0.632
    SV 0.658 0.655 0.601 0.588 0.554 0.640 0.607
    OCC 0.698 0.654 0.649 0.622 0.622 0.664 0.668
    DEF 0.763 0.691 0.729 0.666 0.712 0.745 0.726
    MB 0.656 0.636 0.611 0.619 0.577 0.621 0.680
    FM 0.619 0.704 0.603 0.420 0.555 0.602 0.649
    IPR 0.620 0.596 0.578 0.532 0.536 0.646 0.587
    OV 0.600 0.522 0.563 0.526 0.505 0.536 0.659
    BC 0.717 0.748 0.706 0.579 0.568 0.544 0.684
    LR 0.526 0.519 0.503 0.456 0.460 0.479 0.633
    下载: 导出CSV

    表  3  UAV123的精度

    Table  3.   Accuracy of UAV123

    OUR SRDCFdecon STRCF SRDCF Staple MCCT-H AutoTrack
    VC 0.594 0.477 0.537 0.474 0.485 0.474 0.588
    ARC 0.602 0.476 0.524 0.472 0.459 0.482 0.598
    CM 0.645 0.536 0.602 0.527 0.499 0.519 0.647
    BC 0.539 0.427 0.477 0.389 0.409 0.443 0.502
    FM 0.537 0.403 0.488 0.427 0.356 0.335 0.525
    FOC 0.512 0.427 0.426 0.418 0.388 0.397 0.444
    IV 0.563 0.423 0.493 0.436 0.438 0.458 0.550
    LR 0.535 0.436 0.509 0.431 0.408 0.447 0.532
    OV 0.585 0.483 0.523 0.592 0.441 0.459 0.554
    POC 0.616 0.514 0.559 0.504 0.507 0.530 0.584
    SV 0.632 0.535 0.580 0.531 0.519 0.538 0.629
    SOB 0.699 0.621 0.630 0.585 0.612 0.618 0.664
    下载: 导出CSV

    表  4  UAV123成功率

    Table  4.   Success of UAV123

    OUR SRDCFdecon STRCF SRDCF Staple MCCT-H AutoTrack
    VC 0.514 0.404 0.438 0.398 0.431 0.398 0.480
    ARC 0.505 0.391 0.413 0.387 0.403 0.387 0.476
    CM 0.586 0.483 0.526 0.476 0.460 0.471 0.564
    BC 0.426 0.352 0.374 0.333 0.351 0.372 0.415
    FM 0.443 0.309 0.377 0.354 0.288 0.259 0.407
    FOC 0.347 0.273 0.270 0.261 0.259 0.272 0.291
    IV 0.482 0.357 0.406 0.403 0.383 0.388 0.472
    LR 0.359 0.302 0.334 0.275 0.263 0.299 0.372
    OV 0.516 0.429 0.450 0.432 0.404 0.422 0.490
    POC 0.513 0.440 0.454 0.453 0.432 0.453 0.496
    SV 0.549 0.471 0.494 0.465 0.455 0.470 0.535
    SOB 0.627 0.535 0.555 0.509 0.574 0.570 0.569
    下载: 导出CSV
  • [1] 陈万敏, 振宏, 刘辉. 结合时空上下文信息的相关滤波目标跟踪方法[J]. 红外技术, 2019, 41(9): 866-873. http://hwjs.nvir.cn/article/id/hwjs201909011

    CHEN Wanmin, ZHEN Hong, LIU Hui. Correlation filtering target tracking method based on spatiotemporal context information[J]. Infrared Technology, 2019, 41(9): 866-873. http://hwjs.nvir.cn/article/id/hwjs201909011
    [2] 陈志旺, 王航, 刘旺, 等. 抗遮挡与尺度自适应的改进KCF跟踪算法[J]. 控制与决策, 2021, 36(2): 457-462. https://www.cnki.com.cn/Article/CJFDTOTAL-KZYC202102024.htm

    CHEN Zhiwang, WANG Hang, LIU Wang, et al. Improved KCF tracking algorithm based on anti-occlusion and scale adaptation[J]. Control and Decision, 2021, 36(2): 457-462 https://www.cnki.com.cn/Article/CJFDTOTAL-KZYC202102024.htm
    [3] 陈婧, 孙玉娟, 周万军. 融合运动模型与联合置信度量的改进核相关跟踪算法[J]. 红外技术, 2018, 40(11): 1106-1111. http://hwjs.nvir.cn/article/id/hwjs201811015

    CHEN Jing, SUN Yujuan, ZHOU Wangjun. An improved kernel correlation tracking algorithm combining motion model and joint confidence measure[J]. Infrared Technology, 2018, 40(11): 1106-1111. http://hwjs.nvir.cn/article/id/hwjs201811015
    [4] Bertinetto L, Valmadre J, Henriques J F, et al. Fully-convolutional siamese networks for object tracking[C]//European Conference on Computer Vision, 2016: 850-865.
    [5] LI B, YAN J, WU W, et al. High performance visual tracking with siamese region proposal network[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 8971-8980.
    [6] Valmadre J, Bertinetto L, Henriques J, et al. End-to-end representation learning for correlation filter based tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 2805-2813.
    [7] Bolme D S, Beveridge J R, Draper B A, et al. Visual object tracking using adaptive correlation filters[C]//2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2010: 2544-2550.
    [8] Henriques J F, Caseiro R, Martins P, et al. Exploiting the circulant structure of tracking-by-detection with kernels[C]//European Conference on Computer Vision, 2012: 702-715.
    [9] Henriques J F, Caseiro R, Martins P, et al. High-speed tracking with kernelized correlation filters[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 37(3): 583-596.
    [10] Danelljan M, Hager G, Shahbaz Khan F, et al. Learning spatially regularized correlation filters for visual tracking[C]//Proceedings of the IEEE International Conference on Computer Vision, 2015: 4310-4318.
    [11] Bertinetto L, Valmadre J, Golodetz S, et al. Staple: complementary learners for real-time tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 1401-1409.
    [12] LI F, TIAN C, ZUO W, et al. Learning spatial-temporal regularized correlation filters for visual tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 4904-4913.
    [13] LI Y, FU C, DING F, et al. Autotrack: towards high-performance visual tracking for UAV with automatic spatio-temporal regularization [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020: 11923-11932.
    [14] Bhat G, Johnander J, Danelljan M, et al. Unveiling the power of deep tracking[C]//Proceedings of the European Conference on Computer Vision, 2018: 483-498.
    [15] WANG M, LIU Y, HUANG Z. Large margin object tracking with circulant feature maps[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 4021-4029.
    [16] WU Y, Lim J, YANG M H. Object tracking benchmark[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 37(9): 1834-1848.
    [17] Mueller M, Smith N, Ghanem B. A benchmark and simulator for UAV tracking[C]//European Conference on Computer Vision, 2016: 445-461.
    [18] WANG N, ZHOU W, TIAN Q, et al. Multi-cue correlation filters for robust visual tracking[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 4844-4853.
  • 加载中
图(7) / 表(4)
计量
  • 文章访问数:  118
  • HTML全文浏览量:  28
  • PDF下载量:  25
  • 被引次数: 0
出版历程
  • 收稿日期:  2022-03-24
  • 修回日期:  2022-05-08
  • 刊出日期:  2023-02-20

目录

    /

    返回文章
    返回