《电子技术应用》
您所在的位置:首页 > 其他 > 设计应用 > 关联动态特征的目标自适应跟踪算法
关联动态特征的目标自适应跟踪算法
2022年电子技术应用第11期
孙志成1,董一杰2,胡爱兰2,张瑞权2
1.63861部队,吉林 白城137000;2.华北计算机系统工程研究所,北京100083
摘要: 在复杂的靶场试验场景中,试验现场常常涉及扬尘、强光、遮挡等多变的自然环境。针对这种情况下快速运动的目标物体跟踪,提出了一种关联动态特征的单目标跟踪算法。首先使用门控循环单元(Gated Recurrent Unit,GRU)提取待跟踪目标的时序动态特征,获得候选处理目标框集合;然后利用卷积网络(Convolutional Neural Network,CNN)提取候选目标框的深度卷积特征并确定目标位置,同时分离出背景卷积特征;在跟踪过程中,使用分离出的背景卷积特征图对网络进行参数更新,增强网络的鲁棒性与自适应性。实验结果表明,所提出的算法可以对靶场图像采集系统中的被试移动目标进行自适应跟踪,并且在复杂环境背景下算法仍能保持优异的鲁棒性与适应性。
中图分类号: TP18
文献标识码: A
DOI:10.16157/j.issn.0258-7998.212358
中文引用格式: 孙志成,董一杰,胡爱兰,等. 关联动态特征的目标自适应跟踪算法[J].电子技术应用,2022,48(11):57-62.
英文引用格式: Sun Zhicheng,Dong Yijie,Hu Ailan,et al. Adaptive tracking algorithm for target based on associated dynamic features[J]. Application of Electronic Technique,2022,48(11):57-62.
Adaptive tracking algorithm for target based on associated dynamic features
Sun Zhicheng1,Dong Yijie2,Hu Ailan2,Zhang Ruiquan2
1.63861 Troop,Baicheng 137000,China; 2.National Computer System Engineering Research Institute of China,Beijing 100083,China
Abstract: In the complex scene of shooting range test, the test site often involves the changeable natural environment including dust, strong light, occlusion, etc. A single target tracking algorithm associated with dynamic features is proposed to track fast moving targets in this case. Firstly, the gated recurrent unit is used to extract the time series dynamic characteristics of the target which need to be tracked, so as to obtain a set of candidate processing target frames. Then,convolutional network is adopted to extract the depth convolution features of the candidate target frame and determine target position, as well as separating the background convolution features. In the tracking process, the separated background convolution feature map is applied to update network parameters to enhance the robustness and adaptability of network. Experimental results show that the proposed algorithm can adaptively track moving target in the shooting range image acquisition system, which can still maintain excellent robustness and adaptability in the context of complex environment.
Key words : range test;adaptive tracking;gated recurrent unit;convolutional neural network

0 引言

    某型号系统在进行靶场试验时,需准确定位并跟踪被试设备,确保其能处于相应试验系统范围中,这对单目标跟踪提出了更高的要求。单目标跟踪逐渐成为计算机视觉所需研究和应用的重点之一[1],为了满足某些复杂场景的使用需求,对视频中特定目标进行自适应处理逐渐成为重要的需求。随着近年来计算机技术的发展与算力的进步,单目标跟踪被广泛地应用于军事设施设备、安防监控、无人驾驶等领域[2-4]

    国内外相关学者根据不同的工作原理对跟踪算法做了大量研究工作。Henriques[5]等提出了核相关滤波算法,但该算法在遮挡等因素影响下会出现跟踪丢失的情况;Zhou[6]等提出了结合目标位置、形状、外观的多核相关滤波算法,对实际海洋雷达目标进行跟踪;卢杨[7]等通过改进纹理特征并应用于红外目标跟踪,验证了其鲁棒性与实时性;仇祝令[8]等考虑目标的空时域特性对正则化项进行约束求解,该方法在一定程度上提升了跟踪的实时性与精确度。




本文详细内容请下载:https://www.chinaaet.com/resource/share/2000005005




作者信息:

孙志成1,董一杰2,胡爱兰2,张瑞权2

(1.63861部队,吉林 白城137000;2.华北计算机系统工程研究所,北京100083)




wd.jpg

此内容为AET网站原创,未经授权禁止转载。