《计算机应用研究》|Application Research of Computers

基于相关滤波器和最小二乘估计的目标跟踪方案

Target tracking scheme based on correlation filter and least squares estimation

免费全文下载 (已被下载 次)  
获取PDF全文
作者 赖益强,魏二有,彭绍湖
机构 1.广东外语外贸大学 南国商学院 信息科学技术学院,广州 510545;2.鲁东大学 信息与电气工程学院,山东 烟台 264025;3.广州大学 机械与电气工程学院,广州 510006
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2018)08-2520-04
DOI 10.3969/j.issn.1001-3695.2018.08.070
摘要 针对复杂背景下基于视觉图像的目标跟踪问题,提出一种结合相关滤波和最小二乘估计(LSE)的目标跟踪方案。首先,对已有视频帧进行初始化处理,并构建目标的几何失真图像模板,以此设计一个相关滤波器。然后,当采集到新视频帧时,根据之前的目标位置序列,利用LSE预测出该帧中的目标位置。接着,提取目标图像,并计算其与目标模板之间的相关性。最后,通过阈值来判断是否检测到真正的目标,并进行反馈用来更新目标模板。在GPU上实现该方案,几个视频序列上的实验结果表明,该方案能够准确跟踪目标,且对噪声和光照等环境因素具有很好的鲁棒性。
关键词 目标跟踪;相关滤波;最小二乘估计;目标模板;图形处理单元(GPU)
基金项目 国家自然科学基金资助项目(61175120)
广东省青年创新人才项目(2015KQNCX199)
本文URL http://www.arocmag.com/article/01-2018-08-070.html
英文标题 Target tracking scheme based on correlation filter and least squares estimation
作者英文名 Lai Yiqiang, Wei Eryou, Peng Shaohu
机构英文名 1.SchoolofInformationScience&Technology,SouthChinaBusinessCollege,GuangdongUniversityofForeignStudies,Guangzhou510545,China;2.CollegeofInformation&ElectricalEngineering,LudongUniversity,YantaiShandong264025,China;3.SchoolofMechanical&ElectricalEngineering,GuangzhouUniversity,Guangzhou510006,China
英文摘要 For the issues that the target tracking based on visual image in complex background, this paper proposed a target tracking scheme combining correlation filtering and least squares estimation (LSE). First, it initialized some existing video frames and constructed a geometric distortion image template to design a correlation filter. Then, when the new video frame was acquired, it predicted the target position in the frame by the LSE according to the previous target position sequence. Next, it extracted the target image and calculated the correlation between the target image and the target template. Finally, it used the threshold to determine whether a real target was detected and feedback was used to update the target template. Experiments on the GPU and the experimental results on several video sequences show that the scheme can accurately track the target and has good robustness to environmental factors such as noise and light.
英文关键词 target tracking; correlation filtering; least squares estimation; target template; graphics processing unit(GPU)
参考文献 查看稿件参考文献
  [1] 张焕龙, 胡士强, 杨国胜. 基于外观模型学习的视频目标跟踪方法综述[J] . 计算机研究与发展, 2015, 52(1):177-190.
[2] 徐玉龙, 王家宝, 李阳, 等. 基于相关滤波的尺度自适应目标跟踪[J] . 计算机应用研究, 2016, 33(11):3513-3516, 3520.
[3] Wu Yi, Lim J, Yang M H. Online object tracking:a benchmark[C] //Proc of Computer Vision and Pattern Recognition. Piscataway, NJ:IEEE Press, 2013:2411-2418.
[4] Babenko B, Yang M H, Belongie S. Robust object tracking with online multiple instance learning[J] . IEEE Trans on Pattern Analysis & Machine Intelligence, 2011, 33(8):1619-1632.
[5] Hare S, Golodetz S, Saffari A, et al. Struck:structured output tracking with kernels[J] . IEEE Trans on Pattern Analysis & Machine Intelligence, 2015, 38(10):2096-2109.
[6] Kalal Z, Mikolajczyk K, Matas J. Tracking-learning-detection[J] . IEEE Trans on Pattern Analysis & Machine Intelligence, 2012, 34(7):1409-1422.
[7] 杨德东, 蔡玉柱, 毛宁, 等. 采用核相关滤波器的长期目标跟踪[J] . 光学精密工程, 2016, 24(8):2037-2049.
[8] 刘维建, 谢文冲, 王永良. 基于对角加载的自适应匹配滤波器和自适应相干估计器[J] . 系统工程与电子技术, 2013, 35(3):463-468.
[9] He Yujie, Li Min, Zhang Jinli, et al. Infrared target tracking via weighted correlation filter[J] . Infrared Physics & Technology, 2015, 7(3):103-114.
[10] Diaz-Ramirez V H, Picos K, Kober V. Target tracking in nonuniform illumination conditions using locally adaptive correlation filters[J] . Optics Communications, 2014, 323(7):32-43.
[11] 孟军英, 刘教民, 韩明. 基于联合特征的边缘粒子滤波目标跟踪算法研究[J] . 计算机应用研究, 2015, 32(6):1906-1911, 1916.
[12] 彭鑫, 李仁发, 杨柳, 等. 基于最小二乘优化的车辆位置估计算法[J] . 通信学报, 2010, 31(8):9-15.
[13] Xuan Zhiwu, Nie Kai. Constrained least square estimation algorithm for multisensor bearings-only passive target tracking[J] . Applied Mechanics & Materials, 2014, 60(5):1842-1846.
[14] Long Chengzhi, Luo Jianping, Xiang Mantian, et al. Study on target tracking technology based on the improved least square filter[J] . Journal of Convergence Information Technology, 2012, 7(18):418-424.
[15] 张平. 基于CUDA的TLD视觉跟踪算法研究[D] . 北京:北京交通大学, 2014:16-17.
收稿日期 2017/4/27
修回日期 2017/6/8
页码 2520-2523
中图分类号 TP391
文献标志码 A