基于人工鱼群粒子滤波的TLD改进算法

在面对光照变化、部分遮挡、背景杂乱和平面内外旋转等跟踪难点时, 跟踪学习检测算法(Tracking-Learning-Detection, TLD)容易产生漂移导致跟踪失败, 其跟踪性能还有待提高。在传统TLD算法的基础上, 提出一种基于人工鱼群粒子滤波的TLD改进算法。首先使用人工鱼群粒子滤波跟踪器代替金字塔光流跟踪器, 将颜色直方图特征和方向梯度直方图特征进行融合, 建立目标表观模型, 引入图像金字塔多尺度思想进行尺度匹配, 提高目标跟踪的稳健性。然后通过粒子滤波过程预测目标区域, 将TLD算法检测模块的全局扫描改进为局部扫描, 剔除大量非目标区域, 提高检测模块的检测效率。实验结果表明: 基于人工鱼群粒子滤波的TLD改进算法具有良好的跟踪性能, 与传统TLD算法相比, 其平均成功率和精准度分别提高了19.04%和28.00%, 平均跟踪速度可达33.87 FPS, 提高了38.78%。

Abstract

Faced with tracking difficulties such as illumination variation, occlusion, background clutters, in-plane and out-of-plane rotation, the Tracking-Learning-Detection (TLD) algorithm is prone to drift and cause tracking failures, and its tracking performance needs to be improved. Based on the traditional TLD algorithm, an improved TLD algorithm based on artificial fish swarm particle filtering is proposed. Firstly, the artificial fish swarm particle filter tracker is used to replace the optical flow tracker. The color histogram feature and histogram of oriented gradient feature are fused to establish the target appearance model. The image pyramid multi-scale idea is introduced for scale matching to improve the robustness of target tracking. Then, the target area is predicted through the particle filtering process. The global scanning of TLD algorithm detection module is improved to be the local scanning. A large number of non-target areas are eliminated, and the detection efficiency is improved. The experimental results show that the improved TLD algorithm based on artificial fish swarm particle filtering has good tracking performance. Compared with the traditional TLD algorithm, its average success rate and precision have improved by 19.04% and 28.00%, and the average tracking speed can reach 33.87 FPS, which has improved by 38.78%.

参考文献

[1] 尹宏鹏, 陈波, 柴毅, 等．基于视觉的目标检测与跟踪综述[J]．自动化学报, 2016, 42(10): 1466-1489．

YIN H P, CHEN B, CHAI Y, et al. Vision-based object detection and tracking: a review[J]. Acta Automatica Sinica, 2016, 42(10): 1466-1489. (in Chinese).

[2] 高文, 朱明, 贺柏根,等. 目标跟踪技术综述[J]. 中国光学, 2014, 7(3): 365-375.

GAO W, ZHU M, HE B G,et al. Overview of target tracking technology[J]. Chinese Optics, 2014, 7(3): 365-375. (in Chinese).

[3] 薛陈, 朱明, 刘春香. 遮挡情况下目标跟踪算法综述[J]. 中国光学, 2009, 2(5): 388-394.

XUE C, ZHU M, LIU C X. Review of tracking algorithms under occlusions[J].Chinese Optics, 2009, 2(5): 388-394. (in Chinese).

[4] 唐聪, 凌永顺, 杨华, 等．基于深度学习的红外与可见光决策级融合跟踪[J]．激光与光电子学进展, 2019, 56(7): 071502．

TANG C, LING Y S, YANG H, et al. Decision-level fusion tracking for infrared and visible spectra based on deep learning[J]. Laser & Optoelectronics Progress, 2019, 56(7): 071502. (in Chinese).

[5] KALAL Z, MIKOLAJCZYK K, MATAS J. Tracking-learning-detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(7): 1409-1422.

[6] 吴荻, 战凯, 肖小凤．基于改进光流法和纹理权重的视觉里程计[J]．计算机工程与设计, 2019, 40(1): 230-235．

WU D, ZHAN K, XIAO X F. Visual odometry based on modified optical flow and texture weights[J]. Computer Engineering and Design, 2019, 40(1): 230-235. (in Chinese).

[7] SUN C J, ZHU S H, LIU J W. Fusing Kalman filter with TLD algorithm for target tracking[C]//Proceedings of 2015 34th Chinese Control Conference. Hangzhou: IEEE, 2015: 3736-3741.

[8] 曲海成, 单晓晨, 孟煜, 等．检测区域动态调整的TLD目标跟踪算法[J]．计算机应用, 2015, 35(10): 2985-2989．

QU H C, SHAN X C, MENG Y, et al. Improved TLD target tracking algorithm based on automatic adjustment of surveyed areas[J]. Journal of Computer Applications, 2015, 35(10): 2985-2989. (in Chinese).

[9] 周鑫, 钱秋朦, 叶永强, 等．改进后的TLD视频目标跟踪方法[J]．中国图象图形学报, 2013, 18(9): 1115-1123．

ZHOU X, QIAN Q M, YE Y Q, et al. Improved TLD visual target tracking algorithm[J]. Journal of Image and Graphics, 2013, 18(9): 1115-1123. (in Chinese).

[10] 秦飞, 汪荣贵, 梁启香, 等．基于关键特征点的改进TLD目标跟踪算法研究[J]．计算机工程与应用, 2016, 52(4): 181-187．

QIN F, WANG R G, LIANG Q X, et al. Improved TLD target tracking algorithm based on key feature points[J]. Computer Engineering and Applications, 2016, 52(4): 181-187. (in Chinese).

[11] WU T H, WANG P, YIN S N, et al. A new human eye tracking algorithm of optimized TLD based on improved mean-shift[J]. International Journal of Pattern Recognition and Artificial Intelligence, 2017, 31(3): 1755007.

[12] 康海林, 赵婷, 周骅, 等．结合BRISK与区域预估的改进长时跟踪算法[J]．激光与光电子学进展, 2018, 55(6): 061503．

KANG H L, ZHAO T, ZHOU H, et al. Improved long time tracking algorithm by combining BRISK and region estimation[J]. Laser & Optoelectronics Progress, 2018, 55(6): 061503. (in Chinese).

[13] 王姣尧, 侯志强, 余旺盛, 等．采用核相关滤波的快速TLD视觉目标跟踪[J]．中国图象图形学报, 2018, 23(11): 1686-1696．

WANG J Y, HOU Z Q, YU W S, et al. Fast TLD visual tracking algorithm with kernel correlation filter[J]. Journal of Image and Graphics, 2018, 23(11): 1686-1696. (in Chinese).

[14] 张晶, 王旭, 范洪博．时空上下文相似性的TLD目标跟踪算法[J]．计算机科学与探索, 2018, 12(7): 1169-1181．

ZHANG J, WANG X, FAN H B. TLD object tracking algorithm based on Spatio-temporal context similarity[J]. Journal of Frontiers of Computer Science and Technology, 2018, 12(7): 1169-1181. (in Chinese).

[15] 王法胜, 鲁明羽, 赵清杰, 等．粒子滤波算法[J]．计算机学报, 2014, 37(8): 1679-1694．

WANG F S, LU M Y, ZHAO Q J, et al. Particle filtering algorithm[J]. Chinese Journal of Computers, 2014, 37(8): 1679-1694. (in Chinese).

[16] 李志, 谢强．一种基于改进粒子滤波的运动目标跟踪[J]．计算机科学, 2014, 41(2): 232-235, 252．

LI Z, XIE Q. Moving target tracking based on improved particle filter[J]. Computer Science, 2014, 41(2): 232-235, 252. (in Chinese).

[17] 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．

GUO S Q, ZHANG T, MIAO X K. TLD particle swarm optimization target tracking using a sample deletion mechanism[J]. Optics and Precision Engineering, 2019, 27(5): 1206-1217. (in Chinese).

周志峰, 涂婷, 王立端, 吴明晖. 基于人工鱼群粒子滤波的TLD改进算法[J]. 液晶与显示, 2020, 35(9): 965. ZHOU Zhi-feng, TU Ting, WANG Li-duan, WU Ming-hui. Improved TLD algorithm based on artificial fish-swarm particle filter[J]. Chinese Journal of Liquid Crystals and Displays, 2020, 35(9): 965.

基于人工鱼群粒子滤波的TLD改进算法

关于本站 Cookie 的使用提示

全站搜索

基于人工鱼群粒子滤波的TLD改进算法

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索