面向仿真场景的无人机目标跟踪算法研究

doi:10.16182/j.issn1004731x.joss.24-1159

摘要/Abstract

摘要：

针对UAV在仿真实验中自动跟踪移动目标的需求，提出基于改进CSRT(channel and spatial reliability-aware tracker)算法的无人机长时自动跟踪方法。通过导向滤波加拉普拉斯算子LOGF(laplacian of guided filter)检测获取目标边缘特征，再与HOG(histogram of oriented gradient)和CN(color names)特征融合，增强算法对目标的判别能力；使用平均峰值相关能量和感知哈希汉明距离来综合判定目标状态，当判定目标被遮挡时，采用YOLOv8定位目标，再将定位结果传输至跟踪算法继续跟踪目标。仿真结果表明：在搭建的仿真环境中算法能够在目标被遮挡时仍能长时稳定的跟踪目标，为无人机目标跟踪算法研究提供了良好的仿真实验环境。

关键词: CSRT(channel and spatial reliability-aware tracker), 边缘特征, 定位目标, YOLOv8, 虚拟目标数据集

Abstract:

To address the need for automatic UAV tracking of moving targets in simulated experiments, this paper proposed a long-term automatic tracking method based on an improved channel and spatial reliability-aware tracker (CSRT) algorithm. The target edge features were detected using the Laplacian of guided filter (LOGF) through guided filtering and then fused with the histogram of oriented gradient (HOG) and color names (CN) features to enhance the algorithm's discriminative ability for the target. To evaluate the target state, the paper used average peak correlation energy and perceptual hash Hamming distance. When the target was occluded, the paper employed YOLOv8 for target localization and fed the localization results back into the tracking algorithm to continue tracking. Simulation results demonstrate that the algorithm can maintain long-term stable tracking of the target even when the target is occluded in the constructed simulation environment, providing a favorable simulation experimental environment for the research on UAV target tracking algorithms.

Key words: channel and spatial reliability-aware tracker(CSRT), edge feature, target positioning, YOLOv8, virtual target dataset

中图分类号:

TP391.9

李心仪,王振飞,吴晗 . 面向仿真场景的无人机目标跟踪算法研究[J]. 系统仿真学报, 2026, 38(2): 346-359.

Li Xinyi,Wang Zhenfei,Wu Han . Research on UAV Target Tracking Algorithm for Simulation Scenarios[J]. Journal of System Simulation, 2026, 38(2): 346-359.

图/表 13

图1

图2

图3

图4

图5

图6

表1

图 7

图8

表2

图9

图10

图11

参考文献 18

[1]	蒲良, 张学军. 基于深度学习的无人机视觉目标检测与跟踪[J]. 北京航空航天大学学报, 2022, 48(5): 872-880.
	Pu Liang, Zhang Xuejun. Deep Learning Based UAV Vision Object Detection and Tracking[J]. Journal of Beijing University of Aeronautics and Astronautics, 2022, 48(5): 872-880.
[2]	王法胜, 李富, 尹双双, 等. 全天实时跟踪无人机目标的多正则化相关滤波算法[J]. 自动化学报, 2023, 49(11): 2409-2425.
	Wang Fasheng, Li Fu, Yin Shuangshuang, et al. All-day and Real-time Multi-regularized Correlation Filter for UAV Object Tracking[J]. Acta Automatica Sinica, 2023, 49(11): 2409-2425.
[3]	赵浩光, 孟琭, 耿欢, 等. 尺度自适应的多特征融合相关滤波目标跟踪算法[J]. 控制与决策, 2021, 36(2): 429-435.
	Zhao Haoguang, Meng Lu, Geng Huan, et al. Scale Adaptation and Multi-feature Fusion Correlation Filtering Object Tracking Algorithm[J]. Control and Decision, 2021, 36(2): 429-435.
[4]	程月英. 复杂场景下的单目标跟踪算法研究[D]. 南京: 南京邮电大学, 2021.
	Cheng Yueying. Research on Single Target Tracking Algorithm in Complex Scenes[D]. Nanjing: Nanjing University of Posts and Telecommunications, 2021.
[5]	张博强. 融合视觉注意力机制的相关滤波目标跟踪研究[D]. 阜阳: 辽宁工程技术大学, 2022.
	Zhang Boqiang. Research on Correlation Filter Object Tracking with Visual Attention Mechanism[D]. Fuyang: Liaoning Technical University, 2022.
[6]	Henriques João F, Rui Caseiro, Martins Pedro, et al. High-speed Tracking with Kernelized Correlation Filters[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(3): 583-596.
[7]	Lukežic Alan, Vojír Tomáš, Luka Cehovin Zajc, et al. Discriminative Correlation Filter with Channel and Spatial Reliability[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2017: 4847-4856.
[8]	Dalal N, Triggs B. Histograms of Oriented Gradients for Human Detection[C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05). Piscataway: IEEE, 2005: 886-893.
[9]	Li Bo, Yan Junjie, Wu Wei, et al. High Performance Visual Tracking with Siamese Region Proposal Network[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 8971-8980.
[10]	徐心宇, 陈建. 无人机状态检测Kalman滤波空地目标跟踪算法[J]. 航空学报, 2024, 45(16): 227-246.
	Xu Xinyu, Chen Jian. UAV Object Tracking for Air-ground Targets Based on Status Detection and Kalman Filter[J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(16): 227-246.
[11]	Fei Mengjuan, Ju Zhaojie, Zhen Xiantong, et al. Real-time Visual Tracking Based on Improved Perceptual Hashing[J]. Multimedia Tools and Applications, 2017, 76(3): 4617-4634.
[12]	牛轶峰, 凌黎华. 无人机规避或跟踪空中目标的自适应运动导引方法[J]. 国防科技大学学报, 2017, 39(4): 116-124.
	Niu Yifeng, Ling Lihua. Adaptive Motion Guidance Method of Avoiding or Tracking Aerial Targets for Unmanned Aerial Vehicles[J]. Journal of National University of Defense Technology, 2017, 39(4): 116-124.
[13]	林传健, 章卫国, 史静平, 等. 无人机跟踪系统仿真平台的设计与实现[J]. 哈尔滨工业大学学报, 2020, 52(10): 119-127.
	Lin Chuanjian, Zhang Weiguo, Shi Jingping, et al. Design and Implementation of Simulation Platform for UAV Tracking System[J]. Journal of Harbin Institute of Technology, 2020, 52(10): 119-127.
[14]	王蓓, 李东文, 陈佳. 结合扩展卡尔曼滤波的快速判别尺寸空间滤波跟踪算法[J]. 西安工业大学学报, 2023, 43(3): 287-294.
	Wang Bei, Li Dongwen, Chen Jia. Fast Discriminative Scale Space Filter Tracking Algorithm Based on Extended Kalman Filter[J]. Journal of Xi'an Technological University, 2023, 43(3): 287-294.
[15]	史云飞, 刘克辉, 李向伟, 等. 一种含室内空间结构的建筑物3D LOD概念模型与生成方法[J]. 武汉大学学报(信息科学版), 2022, 47(4): 561-569, 596.
	Shi Yunfei, Liu Kehui, Li Xiangwei, et al. A 3D LOD Conceptual Model of Building with Indoor Spatial Structure and Its Generating Method[J]. Geomatics and Information Science of Wuhan University, 2022, 47(4): 561-569, 596.
[16]	靳晓芳, 岳鼎, 刘金羽. 基于YOLOv3-tiny的智能侦察虚拟训练系统研究[J]. 兵器装备工程学报, 2023, 44(8): 186-190.
	Jin Xiaofang, Yue Ding, Liu Jinyu. Research on Virtual Intelligent Reconnaissance Training System Based on YOLOv3-tiny[J]. Journal of Ordnance Equipment Engineering, 2023, 44(8): 186-190.
[17]	褚治广, 李万杰, 张兴, 等. 基于Unity的"平行空间"冒险解谜游戏的开发设计[J]. 辽宁工业大学学报(自然科学版), 2018, 38(6): 374-377.
	Chu Zhiguang, Li Wanjie, Zhang Xing, et al. Development and Design of "Parallel Space" Adventure Puzzle Game Based on Unity[J]. Journal of Liaoning University of Technology(Natural Science Edition), 2018, 38(6): 374-377.
[18]	Babenko B, Yang M H, Belongie S. Robust Object Tracking with Online Multiple Instance Learning[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(8): 1619-1632.

测试数据	总帧数	POC	FOC	IV	FM	ARV
Tank1	235	否	否	否	是	是
Tank2	275	否	否	否	是	否
Tank3	916	是	否	是	否	否
Militaryvehicle1	293	否	否	是	否	否
Militaryvehicle2	264	是	否	否	否	否
wakeboard1	421	是	否	是	是	是
boat1	901	否	否	否	否	是
car18	1 207	否	否	否	是	是
person12	1 621	是	是	是	否	是
CarScale	253	是	否	否	是	是

属性	评价指标	本文算法	Staple	MIL	KCF	SiamRPN	KPsiamfc
POC	距离精度	88.5	88.3	57.3	55.3	86.3	86.4
POC	跟踪成功率	79.4	77.2	51.7	51.1	72.8	75.6
FOC	距离精度	77.6	65.4	18.8	12.8	75.8	76.1
FOC	跟踪成功率	71.3	58.1	25.8	18.7	70.2	71.1
IV	距离精度	87.3	84.4	62.9	8.60	84.1	85.6
IV	跟踪成功率	81.8	81.3	61.1	51.0	80.0	77.6
FM	距离精度	73.0	52.8	5.40	6.50	82.9	74.6
FM	跟踪成功率	73.5	64.7	7.30	9.10	79.6	73.7
ARV	距离精度	79.7	72.9	38.3	6.70	87.8	78.8
ARV	跟踪成功率	71.8	70.3	35.8	5.70	77.9	72.2