抗遮挡自适应的粒子滤波算法

doi:10.16182/j.issn1004731x.joss.201809041

系统仿真学报 ›› 2018, Vol. 30 ›› Issue (9): 3552-3557.doi: 10.16182/j.issn1004731x.joss.201809041

抗遮挡自适应的粒子滤波算法

李菊^1,2, 曹明伟³, 余烨³, 夏瑜¹, 周立凡¹

1.常熟理工学院计算机科学与工程学院, 江苏常熟 215500;
2.苏州大学江苏省计算机信息处理技术重点实验室, 江苏苏州 215006;
3.合肥工业大学vcc实验室, 安徽合肥 230009

收稿日期:2016-03-06 出版日期:2018-09-10 发布日期:2019-01-08
作者简介:李菊(1981-),女,江苏常熟,博士,副教授,研究方向为运动目标检测与跟踪。
基金资助:
国家自然科学基金(61300118,61602062),江苏省高校重点实验室开放课题(KJS1521,KJS1522),江苏省高校自然科学研究项目(16KJD520001)

An Anti-occlusion Adaptive Particle Filtering Algorithm

Li Ju^1,2, Cao Mingwei³, Yu Ye³, XiaYu¹, Zhou Lifan¹

1. School of Computer Science and Engineering, Changshu Institute of Technology, Changshu 215500, China;
2. Provincial Key Laboratory for Computer Information Processing Technology, Soochow University, Suzhou 215006, China;
3.VCC Division, School of Computer and Information, Hefei University of Technology, Hefei 230009, China

Received:2016-03-06 Online:2018-09-10 Published:2019-01-08

摘要/Abstract

摘要： 针对局部遮挡对粒子滤波跟踪算法的影响,提出了抗遮挡自适应的粒子滤波算法。该算法采用矩形作为跟踪窗,在重要性重采样之后,引入K均值聚类(KM聚类)算法,将空间特征与权重分布近似的粒子进行聚类,得到粒子子群,并通过粒子子群估计目标的最终状态。当目标外形变化超过5%时,说明存在遮挡或同色干扰现象,则跟踪窗保持上一帧尺寸不变,否则跟踪窗随着运动目标尺度变化而自适应变化。实验结果表明该算法对局部遮挡和运动目标尺度变化具有较强的鲁棒性。

关键词: 抗遮挡, K均值聚类, 粒子滤波, 运动目标跟踪

Abstract: Considering the influence of local occlusion on particle filter tracking algorithm, an anti occlusion adaptive particle filtering algorithm is proposed. It adopts a rectangle as the tracking window, and uses the K mean clustering algorithm to complete particle clustering in resampling, and then obtains the particle subgroup. It estimates the final state according to particles subgroups, and modifies the tracking window. When the area changes more than 5%, the tracking window maintains the same as the one in last frame. Otherwise, the tracking window will change according to the size of moving object, which is a self-adaptation process. At the same time it solves the degeneration problem of particle filter. This algorithm strengthens the robustness of tracking algorithm in case of local occlusion and moving object scale changing. The method performs better than the traditional particle filter tracking algorithm.

Key words: anti-occlusion, KM clustering, particle filter, moving object tracking

中图分类号:

TP391.41

李菊, 曹明伟, 余烨, 夏瑜, 周立凡. 抗遮挡自适应的粒子滤波算法[J]. 系统仿真学报, 2018, 30(9): 3552-3557.

Li Ju, Cao Mingwei, Yu Ye, XiaYu, Zhou Lifan. An Anti-occlusion Adaptive Particle Filtering Algorithm[J]. Journal of System Simulation, 2018, 30(9): 3552-3557.

参考文献

[1] 王龙生, 顾浩, 余云智. 基于微分进化的组合重采样粒子滤波算法[J]. 电光与控制, 2012, 19(11): 43-46.
Longsheng Wang, Hao Gu, Yunzhi Yu.Combined resampling particle filter algorithm based on differential evolution[J]. Electro-optic and control, 2012, 19(11): 43-46.
[2] 骆荣剑, 李颖, 钱广华, 等. 一种改进的自适应重采样粒子滤波算法[J]. 光电工程, 2014, 41(4): 35-40.
Rongjian Luo, Ying Li, Guanghua Qian, et al.An improved adaptive resampling particle filter algorithm[J]. Photoelectric Engineering, 2014, 41(4): 35-40.
[3] 陈龙, 郭宝龙, 孙伟. 基于FCM 聚类的粒子滤波多目标跟踪算法[J]. 仪器仪表学报, 2011, 32(11): 16-21.
Long Chen, Baolong Guo, Wei Sun.Particle filter multi-target tracking algorithm based on FCM clustering[J]. Journal of Instruments and Instruments, 2011, 32(11): 16-21.
[4] 颜佳, 吴敏渊, 陈淑珍, 等. 应用Mean Shift 和分块的抗遮挡跟踪[J]. 光学精密工程, 2010, 18(6): 1413-1419.
Jia Yan, Minyuan Wu, Shuzhen Chen, et al.Applied Mean Shift and block anti occlusion tracking[J]. optical precision engineering, 2010, 18(6): 1413-1419.
[5] Tao Yang, Stan Z Li, Quan Pan, et al.Real-time Multiple Objects Tracking with Occlusion Handling in Dynamic Scenes[J]. Computer Vision and Pattern Recognition, 2005, 1(1): 20-25.
[6] Comaniciu D, Ramesh V, Meer P.Kernel-based objecttrackin[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2003, 25(5): 564-577.
[7] 彭宁嵩, 杨杰, 刘志, 等. Mean Shift跟踪算法中核函数窗宽的自动选取[J]. 软件学报, 2005, 16(9): 1542-1550.
Ningsong Peng, Jie Yang, Zhi Liu, et al.Automatic selection of kernel window width in Mean Shift tracking algorithm[J]. Journal of Software, 2005, 16(9): 1542-1550.
[8] Nebehay, Georg, Roman Pflugfelder. Consensus-based matching and tracking of keypoints for object tracking[C]. Applications of Computer Vision, 2014 IEEE Winter Conference on. IEEE, 2014: 40-51.
[9] Danelljan, Martin.Adaptive color attributes for real-time visual tracking[C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2014: 332-340.
[10] Li Hanxi, Chunhua Shen, Qinfeng Shi.Real-time visual tracking using compressive sensing[C]. Computer Vision and Pattern Recognition, 2011 IEEE Conference on. IEEE, 2011: 226-235.
[11] Zhou S K, Moghaddam B.Appearance tracking using adaptive models in a particle filter[C]. Asian Conference on Computer Vision, 2015: 1321-1329.
[12] Yi Wu, Jongwoo Lim, Yang Minghsuan.Online Object Tracking: A Benchmark[C]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2015: 2411-2418.

抗遮挡自适应的粒子滤波算法

An Anti-occlusion Adaptive Particle Filtering Algorithm

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 9

编辑推荐

Metrics

本文评价

[1]	李元, 耿泽伟. 基于LLE与K均值聚类算法的工业过程故障诊断[J]. 系统仿真学报, 2021, 33(9): 2066-2073.
[2]	李维刚, 李阳, 赵云涛, 严保康. 基于改进灰狼算法的粒子滤波算法研究[J]. 系统仿真学报, 2021, 33(1): 37-45.
[3]	赵勇, 陈彬, 王小东, 朱正秋, 王戎骁, 邱晓刚. 化工园区有害气体泄漏寻源算法研究[J]. 系统仿真学报, 2019, 31(12): 2810-2815.
[4]	王戎骁, 陈彬, 邱思航, 朱正秋, 邱晓刚. 基于源项估计的污染气体扩散数据驱动仿真[J]. 系统仿真学报, 2017, 29(9): 2100-2108.
[5]	张建春, 康凤举, 梁洪涛, 徐皓. 基于鸡群优化的粒子滤波算法研究[J]. 系统仿真学报, 2017, 29(2): 295-300.
[6]	祁鹏程, 赵忠盖, 刘飞. 一种融合时滞值的状态估计方法及其应用[J]. 系统仿真学报, 2016, 28(8): 1818-1823.
[7]	陈莹, 程雪雪. 基于粒子滤波节点选择的协同跟踪[J]. 系统仿真学报, 2016, 28(6): 1281-1288.
[8]	夏飞, 郝硕涛, 张浩, 彭道刚. 改进粒子滤波和小波包在汽轮机振动诊断中的应用[J]. 系统仿真学报, 2016, 28(11): 2823-2831.
[9]	张子迎, 周明全, 税午阳, 武仲科, 郑霞. 基于K均值聚类和区域匹配的颜色迁移技术[J]. 系统仿真学报, 2015, 27(10): 2359-2364.