引入样本删除机制的TLD粒子群目标跟踪

郭巳秋; 张涛; 苗锡奎

doi:10.3788/OPE.20192705.1206

您当前的位置：

首页 >

文章列表页 >

引入样本删除机制的TLD粒子群目标跟踪

信息科学 | 更新时间：2020-08-13

- 引入样本删除机制的TLD粒子群目标跟踪
- TLD particle swarm optimization target tracking using a sample deletion mechanism
- 光学精密工程 2019年27卷第5期页码：1206-1217
- 作者机构：
  
  1.中国科学院长春光学精密机械与物理研究所，吉林长春 130033
  2.中国科学院大学，北京 100049
  3.中国洛阳电子装备试验中心光电对抗测试评估技术重点实验室，河南洛阳 471000
- 作者简介：
  
  [ "郭巳秋 (1989-)，女，吉林长春人，博士研究生，助理研究员，2012年、2015年于北京理工大学分别获得学士、硕士学位，主要从事目标检测、跟踪算法方面的研究工作。E-mail:guo_qiuqiu@163.com" ]
  [ "张涛 (1964-)，男，吉林长春人，博士，研究员，1987年、1990年于浙江大学分别获得学士、硕士学位，1993年于长春光学精密机械与物理研究所获得博士学位，主要从事探测与成像技术，稳定控制技术方法的研究。E-mail: ZhangT@ciomp.ac.cn" ]
- 基金信息：
  
  中国洛阳电子装备试验中心光电对抗测试评估技术重点实验室开放课题资助项目(GKCP2017001);中国洛阳电子装备试验中心光电对抗测试评估技术重点实验室开放课题资助项目(GKCP2017002)
- DOI：10.3788/OPE.20192705.1206
  中图分类号： TP394.1
- 收稿日期：2018-12-05，
  
  录用日期：2019-1-3，
  
  纸质出版日期：2019-05-15
- 稿件说明：
移动端阅览
郭巳秋, 张涛, 苗锡奎. 引入样本删除机制的TLD粒子群目标跟踪[J]. 光学精密工程, 2019,27(5):1206-1217.

Si-qiu GUO, Tao ZHANG, Xi-kui MIAO. TLD particle swarm optimization target tracking using a sample deletion mechanism[J]. Optics and precision engineering, 2019, 27(5): 1206-1217.
郭巳秋, 张涛, 苗锡奎. 引入样本删除机制的TLD粒子群目标跟踪[J]. 光学精密工程, 2019,27(5):1206-1217. DOI： 10.3788/OPE.20192705.1206.

Si-qiu GUO, Tao ZHANG, Xi-kui MIAO. TLD particle swarm optimization target tracking using a sample deletion mechanism[J]. Optics and precision engineering, 2019, 27(5): 1206-1217. DOI： 10.3788/OPE.20192705.1206.

摘要

为提高TLD算法在广泛场景下跟踪鲁棒性和实时性的问题

本文从跟踪模块和学习模块两个方面对TLD算法进行了改进，提出引入样本删除机制的TLD粒子群目标跟踪算法。首先，用基于颜色特征的粒子群目标跟踪算法替代TLD算法中原来的跟踪模块，增强TLD算法在应对目标出现非刚性形变、尺度变化、旋转、遮挡等情况下的跟踪鲁棒性。接着，针对TLD算法的学习模块引入样本删除机制，在跟踪过程中为样本库中正负样本数量分别设定一个阈值，当正负样本数都达到各自阈值时，便会启动样本删除机制。然后，对待分类进入样本库的图像块进行等级评价，删除对正负样本表征能力都较弱图像块。最后，将样本库中的正负样本与当前目标进行相似度匹配，删除对当前目标表征能力低的样本。通过对OTB2013和OTB2015数据集中相关视频序列的实验结果证明，本文算法的OPE精确度达到0.687，算法的OPE成功率为0.488，算法运算效率平均提高了25.71%。基本满足广泛场景下目标跟踪的鲁棒性，并显著了提高算法运算效率。

Abstract

This study improved the tracking robustness and real-time performance of a tracking-learning-detection (TLD) algorithm for a wide range of scenarios by considering two important aspects

namely

the tracking and learning modules. The study proposed a TLD particle swarm optimization (PSO) target-tracking algorithm using a sample deletion mechanism. First

the original tracking module of a TLD algorithm was replaced by a color-feature-based PSO target-tracking algorithm

which enhanced the tracking performance of the TLD algorithm in terms of target non-rigid deformation

scale variation

rotation

and occlusion. Second

a sample deletion mechanism for the learning module of the TLD algorithm was introduced. During the tracking process

a threshold was set for the positive and negative samples. When both the positive and negative samples reach their respective thresholds

the sample deletion mechanism was initiated. The image blocks to be classified into the sample library were then graded

and those with a weak representation ability for both positive and negative samples were deleted. Finally

we matched the positive and negative samples in the sample library with the current target and delete the samples with low representational ability to the current target. Experiments on OTB2013 and OTB2015 datasets show that the one-pass evaluation (OPE) accuracy of the proposed algorithm reaches 0.687

the OPE success rate of the algorithm is 0.488

and the operation efficiency is improved by 25.64% on average. This essentially satisfies the robustness of target tracking in a wide range of scenarios and significantly improves the computational efficiency of the algorithm.

关键词

Keywords

references

王春平, 王暐, 刘江义, 等.基于色度饱和度-角度梯度直方图特征的尺度自适应核相关滤波跟踪[J].光学精密工程, 2016, 24(9): 2293-2301.

WANG CH P, WANG W, LIU J Y, et al .. Scale adaptive kernelized correlation filter tracking based on HHS-OG feature[J]. Opt. Precision Eng. , 2016, 24(9): 2293-2301. (in Chinese)

程帅, 孙俊喜, 曹永刚, 等.基于色度饱和度-角度梯度直方图特征的尺度自适应核相关滤波跟踪[J].光学精密工程, 2015, 23(4): 1161-1170.

CEHNG SH, SUN J X, CAO Y G, et al .. Target tracking based on incremental deep learning[J]. Opt. Precision Eng. , 2015, 23(4): 1161-1170. (in Chinese)

LUO L S, FAN X S. Immune particle filter algorithm for target tracking based on histograms of color and oriented gradient[J]. Optical Sensing and Imaging Technology and Applications , 2017, 10462.

李泽原, 孙晓刚, 林云.基于卷积神经网络和核相关滤波的物体跟踪算法[J].计算机应用, 2017, 37(S2):107-111.

LI Z Y, SUN X G, LIN Y. Object tracking algorithm based on convolutional neural network and kernel correlation filtering[J]. Journal of Computer Applications , 2017, 37(S2):107-111. (in Chinese)

ZHANG D, ZHANG J, XIA C. Multi-Complementary model for long-term tracking[J]. Sensors , 2018, 18(2):527.

刘教民, 郭剑威, 师硕.自适应模板更新和目标重定位的相关滤波器跟踪[J].光学精密工程, 2018, 26(6):1533-1541.

LIU J M, GUO J W, SHI SH. Correlation filter tracking based on adaptive learning rate and location refiner[J]. Opt. Precision Eng. , 2018, 26(6):1533-1541. (in Chinese)

BOLME D S, BEVERIDGE J R, DRAPER B A, et al .. Visual object tracking using adaptive correlation filters[C]. Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition , San Francisco, CA, USA, 2010: 2544-2550.

DANELLJAN M, HAGER G, KHAN F S, et al .. Accurate scale estimation for robust visual tracking[J]. British Machine Vision Conference , 2014:65.1-65.11.

HENRIQUES J F, CASEIRO R, MARTINS P, et al .. High-Speed tracking with kernelized correlation filters[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence , 2015, 37(3):583-596.

MA C, HUANG J B, YANG X, et al .. Hierarchical Convolutional Features for Visual Tracking[C]. Proceedings of the IEEE International Conference on Computer Vision , Santiago, Chile, 2016: 3074-3082.

KALAL Z, MATAS J, MIKOLAJCZYK K. Online learning of robust object detectors during unstable tracking[C]. Proceedings of the IEEE International Conference on Computer Vision Workshops ( ICCV Workshops ), Kyoto, Japan, 2009: 1417-1424.

胡欣, 高佳丽.基于改进的TLD目标跟踪算法[J].计算机应用研究, 2019, 36(11):1-5.

HU X, GAO J L. Target tracking algorithm based on improved TLD[J]. Application Research of Computers , 2019, 36(11):1-5. (in Chinese)

刘曙, 狄红卫, 姚曼虹.基于自适应尺度的TLD目标跟踪算法[J].光学技术, 2017, 43(6):542-546.

LIU SH, DI H W, YAO M H. TLD target tracking algorithm based on adaptive scale[J]. Optical Technique , 2017, 43(6):542-546. (in Chinese)

吴忠文, 费树岷.基于TLD目标跟踪算法的优化[J].工业控制计算机, 2018, 31(4):77-79.

WU ZH W, FEI SH M. Optimization based on TLD target tracking algorithm[J]. Industrial Control Computer, 2018, 31(4):77-79. (in Chinese)

KALAL Z, MIKOLAJCZYK K, MATAS J. Face-tld: Tracking-learning-detection applied to faces[C]. Proceedings of the IEEE International Conference on Image Processing ( ICIP ), Hong Kong, China , 2010: 3789-3792.

KALAL Z, MIKOLAJCZYK K, MATAS J. Forward-Backward Error: Automatic Detection of Tracking Failures[C]. Proceedings of the IEEE Computer Society International Conference on Pattern Recognition, Istanbul, Turkey , 2010: 2756-2759.

王明晓.基于改进的TLD目标跟踪算法研究[D].哈尔滨: 哈尔滨理工大学, 2017. http://cdmd.cnki.com.cn/Article/CDMD-10214-1017033998.htm

WANG M X. Target Tracking Algorithms Based on Improved TLD [D].Harbin: Harbin University of Science and Technology, 2017. (in Chinese)

KALAL Z, MATAS J, MIKOLAJCZYK K. P-N learning: Bootstrapping binary classifiers by structural constraints[J]. IEEE Computer Vision and Pattern Recognition , 2010:49-56.

GUO S Q, ZHANG T, SONG Y L, et al .. Color feature- based object tracking through particle swarm optimization with improved inertia weight[J]. Sensors, 2018, 18:1292.

KENNEDY J, EBERHART R. Particle swarm optimization[C]. Proceedings of the IEEE International Conference on Neural Networks, Perth , 1995, 4: 1942-1948.

KALAL Z, MIKOLAJCZYK K, MATA J. Tracking-Learning-Detection[J]. IEEE Transactions on Software Engineering , 2011, 34(7):1409-1422.

HARE S, SAFFARI A, TORR P H S. Struck: structured output tracking with kernels[C]. Proceedings of the IEEE International Conference on Computer Vision (ICCV), Barcelona, Spain , 2011: 263-270.

KWON J, LEE K M. Visual tracking decomposition[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , San Francisco, CA, USA, 2010: 1269-1276.

HENRIQUES J F, CASEIRO R, MARTINS P, et al .. Exploiting the circulant structure of tracking-by-detection with kernels[C]. In Proceedings of the European Conference on Computer Vision (ECCV) , Florence, Italy, 2012: 702-715.

ZHANG K H, ZHANG L, YANG M H. Real-time compressive tracking[C]. Proceedings of the European Conference on Computer Vision (ECCV) , Florence, Italy, 2012: 864-877.

SEVILLA-LARA L, LEARNED-MILLER E. Distribution fields for tracking[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , Providence, RI, USA, 2012: 1910-1917.

GRABNER H, GRABNER M, BISCHOF H. Real-time tracking via on-line boosting[C]. Proceedings of the British Machine Vision Conference (BMCV) , Edinburgh, UK, 2006: 47-56.

BABENKO B, YANG M H, BELONGIE S. Visual tracking with online multiple instances learning[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Miami, FL, USA , 2009: 983-990.

NING J, YANG J, JIANG S, et al .. Object tracking via dual linear structured svm and explicit feature map[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA , 2016: 4266-4274.

WU Y, LIM J, YANG M H. Online object tracking: a benchmark[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Portland, OR, USA , 2013: 2411-2418.

浏览量

229

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

数据驱动的多模复合制导信息融合及其试验验证

基于非机动单站的目标运动状态无偏估计

地基大口径望远镜动态目标跟踪中压电式快速反射镜迟滞效应的补偿

基于级联滑模控制的高精度光电跟踪与捕获

基于场景辅助特征的T-S目标跟踪