基于孪生网络的轻量型无人机单目标跟踪算法

doi:10.7527/S1000-6893.2024.30925

Abstract

Abstract:

Currently， some single object tracking algorithms have achieved leading performance， but their large models limit their application on resource-constrained platforms such as Unmanned Aerial Vehicles （UAVs）. This paper designs a lightweight single object tracking algorithm for UAVs based on the Siamese network to achieve efficient tracking with less resource consumption. Firstly， a lightweight backbone network for Siamese feature extraction is designed based on MobileNetV3， significantly reducing the computation and parameter volume of the network without compromising feature extraction capabilities. Secondly， a dual cross-correlation module is designed. The module uses Pointwise Cross-Correlation to quickly calculate the similarity between template image features and search image features and also uses Depthwise Cross-Correlation to supplement missing features， effectively enhancing feature matching accuracy and robustness. Then， a lightweight prediction head is designed by stacking multiple depthwise separable convolution layers， obtaining accurate target representations with minimal performance consumption. Finally， classification ranking loss is introduced to traditional classification and regression losses， enhancing the network’s ability to learn target foreground and suppress background interference， and further improving tracking performance. Comprehensive experiments show that the proposed algorithm achieves 82.1%， 81.2%， and 64.6% precision and 63.4%， 61.8%， and 49.6% success rate on DTB70， UAV123， and UAV20L datasets， respectively， with only 5.3×10⁵ parameters and 1.1×10⁸ floating point operations. It achieves the performance comparable to state-of-the-art tracking algorithms， while having significantly fewer parameters and computational load， and it can run at the speeds above 100 fps， meeting the real-time requirements of UAV object tracking.

Key words: UAV, Siamese network, single object tracking, lightweight, cross-correlation

CLC Number:

V279

Qishuai DING, Bangjun LEI, Zhengping WU. A lightweight single object tracking algorithm for UAVs based on Siamese network[J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(4): 330925.

Figures/Tables 15

Fig.1

Table 1

Fig.2

Fig.3

Fig.4

Table 2

Table 3

Fig.5

Fig.6

Fig.7

Table 4

Table 5

Table 6

Fig.8

Fig.9

References 40

1	BERTINETTO L， VALMADRE J， HENRIQUES J F， et al. Fully-convolutional Siamese networks for object tracking［C］∥European Conference on Computer Vision. Cham： Springer， 2016： 850-865.
2	LI B， YAN J J， WU W， et al. High performance visual tracking with Siamese region proposal network‍［C］‍∥2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE Press， 2018： 8971-8980.
3	ZHU Z， WANG Q， LI B， et al. Distractor-aware siamese networks for visual object tracking‍［C］‍∥Proceedings of the European Conference on Computer Vision （ECCV）. Cham： Springer， 2018： 101-117.
4	LI B， WU W， WANG Q， et al. Siamrpn++： Evolution of siamese visual tracking with very deep networks［C］‍∥Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE Press， 2019： 4282-4291.
5	HE K， ZHANG X， REN S， et al. Deep residual learning for image recognition［C］∥Proceedings of the IEEE Conference on Computer vision and Pattern Recognition. Piscataway： IEEE Press， 2016： 770-778.
6	HU W， WANG Q， ZHANG L， et al. Siammask： A framework for fast online object tracking and segmentation［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2023， 45（3）： 3072-3089.
7	雷帮军，丁奇帅，牟乾西，等. 基于模板更新和双特征增强的视觉跟踪算法［J/OL］. 北京航空航天大学学报，（2024-02-27）［2024-07-01］. .
	LEI B J， DING Q S， MOU Q X， et al. Visual tracking algorithm based on template updating and dual feature enhancement［J］. Journal of Beijing University of Aeronautics and Astronautics，（2024-02-27 ［2024-07-01］. （in Chinese）.
8	CHEN Z D， ZHONG B N， LI G R， et al. Siamese box adaptive network for visual tracking‍［C］‍∥2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2020： 6667-6676.
9	GUO D Y， WANG J， CUI Y， et al. SiamCAR： Siamese fully convolutional classification and regression for visual tracking‍［C］‍∥2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2020： 6268-6276.
10	LIN L， FAN H， ZHANG Z， et al. SwinTrack： A simple and strong baseline for transformer tracking［DB/OL］. arXiv preprint： 2112. 00995， 2021.
11	CHEN X， YAN B， ZHU J W， et al. Transformer tracking［C］∥2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2021： 8122-8131.
12	CHEN X， PENG H W， WANG D， et al. SeqTrack： Sequence to sequence learning for visual object tracking［C］∥2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2023： 14572-14581.
13	YAN B， PENG H W， WU K， et al. LightTrack： Finding lightweight neural networks for object tracking via one-shot architecture search［C］∥2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2021： 15175-15184.
14	CAO Z A， HUANG Z Y， PAN L， et al. TCTrack： Temporal contexts for aerial tracking［C］∥2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2022： 14778-14788.
15	CAO Z A， HUANG Z Y， PAN L， et al. Towards real-world visual tracking with temporal contexts［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2023， 45（12）： 15834-15849.
16	FU C H， CAO Z A， LI Y M， et al. Onboard real-time aerial tracking with efficient Siamese anchor proposal network［J］. IEEE Transactions on Geoscience and Remote Sensing， 2022， 60： 5606913.
17	CAO Z A， FU C H， YE J J， et al. SiamAPN： Siamese attentional aggregation network for real-time UAV tracking［C］∥2021 IEEE/RSJ International Conference on Intelligent Robots and Systems （IROS）. Piscataway： IEEE Press， 2021： 3086-3092.
18	XING D T， EVANGELIOU N， TSOUKALAS A， et al. Siamese transformer pyramid networks for real-time UAV tracking［C］∥2022 IEEE/CVF Winter Conference on Applications of Computer Vision （WACV）. Piscataway： IEEE Press， 2022： 1898-1907.
19	刘芳，孙亚楠. 基于自适应融合网络的无人机目标跟踪算法［J］. 航空学报， 2022， 43（7）： 359-369.
	LIU F， SUN Y N. UAV target tracking algorithm based on adaptive fusion network［J］. Acta Aeronautica et Astronautica Sinica， 2022， 43（7）： 359-369 （in Chinese）.
20	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）： 329834
	XU X Y， CHEN J. UAV object tracking for air-ground targets based on status detection and Kalman filter［J］.Acta Aeronautica et Astronautica Sinica， 2024， 45（16）： 329834 （in Chinese）.
21	薛远亮，金国栋，谭力宁，等. 基于多尺度融合的自适应无人机目标跟踪算法［J］. 航空学报， 2023， 44（1）： 326107 .
	XUE Y L， JIN G D， TAN L N， et al. Adaptive UAV target tracking algorithm based on multi-scale fusion［J］. Acta Aeronautica et Astronautica Sinica， 2023， 44（1）： 326107 （in Chinese）.
22	HOWARD A， SANDLER M， CHEN B， et al. Searching for MobileNetV3‍［C］‍∥2019 IEEE/CVF International Conference on Computer Vision （ICCV）. Piscataway： IEEE Press， 2019： 1314-1324.
23	TANG F， LING Q. Ranking-based Siamese visual tracking［C］∥2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2022： 8731-8740.
24	LI S Y， YEUNG D Y. Visual object tracking for unmanned aerial vehicles： A benchmark and new motion models［J］. Proceedings of the AAAI Conference on Artificial Intelligence， 2017， 31（1）： 4140-4146.
25	MUELLER M， SMITH N， GHANEM B. A benchmark and simulator for UAV tracking［M］∥Computer Vision‐ECCV 2016. Cham： Springer， 2016： 445-461.
26	HOWARD A G， ZHU M L， CHEN B， et al. MobileNets： Efficient convolutional neural networks for mobile vision applications［DB/OL］. arXiv preprint： 1704. 04861， 2017.
27	SANDLER M， HOWARD A， ZHU M L， et al. MobileNetV2： inverted residuals and linear bottlenecks［C］‍∥2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE Press， 2018： 4510-4520.
28	HU J， SHEN L， SUN G. Squeeze-and-excitation networks［C］‍∥2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE Press， 2018： 7132-7141.
29	GUO D Y， SHAO Y Y， CUI Y， et al. Graph attention tracking［C］∥2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2021： 9538-9547.
30	LIAO B Y， WANG C Y， WANG Y Y， et al. PG-net： Pixel to global matching network for visual tracking［C］∥European Conference on Computer Vision. Cham： Springer， 2020： 429-444.
31	LIN T Y， MAIRE M， BELONGIE S， et al. Microsoft COCO： Common objects in context［M］∥Lecture Notes in Computer Science. Cham： Springer International Publishing， 2014： 740-755.
32	REAL E， SHLENS J， MAZZOCCHI S， et al. Youtube-boundingboxes： A large high-precision human-annotated data set for object detection in video［C］∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE Press， 2017： 5296-5305.
33	RUSSAKOVSKY O， DENG J， SU H， et al. Imagenet large scale visual recognition challenge［J］. International Journal of Computer Vision， 2015， 115： 211-252.
34	HUANG L H， ZHAO X， HUANG K Q. GOT-10k： A large high-diversity benchmark for generic object tracking in the wild［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2021， 43（5）： 1562-1577.
35	ZHANG Z， PENG H， FU J， et al. Ocean： Object-aware anchor-free tracking‍［C］‍∥Computer Vision‍-‍ECCV 2020. Cham： Springer， 2020： 771-787.
36	ZHANG Z P， PENG H W. Deeper and wider Siamese networks for real-time visual tracking［C］‍∥2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2019： 4586-4595.
37	GUO Q， FENG W， ZHOU C， et al. Learning dynamic Siamese network for visual object tracking‍［C］‍∥2017 IEEE International Conference on Computer Vision （ICCV）. Piscataway： IEEE Press， 2017： 1781-1789.
38	VALMADRE J， BERTINETTO L， HENRIQUES J， et al. End-to-end representation learning for correlation filter based tracking‍［C］‍∥2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2017： 5000-5008.
39	DANELLJAN M， BHAT G， KHAN F S， et al. ECO： Efficient convolution operators for tracking‍［C］‍∥2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2017： 6931-6939.
40	LI X， MA C， WU B Y， et al. Target-aware deep tracking‍［C］‍∥2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2019： 1369-1378.

层数	输入	输出	卷积核	步长	SE
0	255²×3	128²×16	3×3	2
1	128²×16	64²×16	3×3	2	√
2	64²×16	32²×24	3×3	2
3	32²×24	32²×24	3×3	1
4	32²×24	16²×40	5×5	2	√
5	16²×40	16²×40	5×5	1	√
6	16²×40	16²×40	5×5	1	√
7	16²×40	16²×48	5×5	1	√
8	16²×48	16²×48	5×5	1	√
9	16²×48	16²×96	5×5	1	√

算法类别	跟踪算法	准确率	成功率	浮点数运算量/10⁹	参数量/10⁶	速度/fps
非轻量型算法	SiamDW_FC^［1］	0.722	0.489	8.00	7.40	76.70
	Ocean^［35］	0.636	0.455	4.40	25.9	61.60
	SiamMask^［6］	0.777	0.575	16.70	18.80	53.80
	SiamGAT^［29］	0.754	0.583	17.28	14.23	46.40
	SiamRPN++^［4］	0.795	0.589	48.90	53.95	51.60
轻量型算法	DaSiamRPN^［3］	0.694	0.472	21.10	19.60	163.20
	LightTrack^［13］	0.763	0.587	0.79	3.13
	SiamAPN^［16］	0.784	0.586	10.30	15.10	127.50
	SiamAPN++^［17］	0.792	0.594	9.20	12.20	93.40
	SiamTPN^［18］	0.728	0.572	3.73	10.77
	TCTrack^［14］	0.814	0.620	4.51	6.26	76.40
	Ours	0.821	0.634	0.11	0.53	174.69

跟踪算法	准确率	成功率
SiamFC^［1］	0.719	0.498
TCTrack++^［15］	0.731	0.519
SiamDW^［36］	0.776	0.536
SiamRPN^［2］	0.772	0.581
SiamMask^［6］	0.794	0.602
SiamBAN^［8］	0.795	0.604
TCTrack^［14］	0.800	0.604
SiamRPN++^［4］	0.804	0.611
SiamCAR^［9］	0.804	0.615
Ours	0.812	0.618

跟踪算法	准确率	成功率
DSiam^［37］	0.603	0.391
CFNet^［38］	0.588	0.399
SiamFC^［1］	0.599	0.402
ECO^［39］	0.589	0.427
TADT^［40］	0.609	0.459
DaSiamRPN^［3］	0.665	0.465
Ours	0.646	0.496

组合	本文 MobileNetV3	原始 MobileNetV3	轻量化预测头部	深度互相关	像素互相关	分类排名损失	浮点数运算量/10⁶	参数量/10³	准确率	成功率
1		√	√	√			101.89	394.68	0.709	0.531
2	√		√	√			104.42	465.16	0.728	0.556
3		√	√		√		103.66	414.52	0.711	0.538
4	√		√		√		106.19	484.99	0.736	0.563
5		√	√	√	√		113.03	470.97	0.764	0.574
6	√		√	√	√		115.56	541.45	0.786	0.600
7		√	√	√	√	√	113.03	470.97	0.770	0.578
8	√		√	√	√	√	115.56	541.45	0.812	0.618

A lightweight single object tracking algorithm for UAVs based on Siamese network

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 15

References 40

Related Articles 15

Recommended Articles

Metrics

Comments

堆叠次数N	浮点数运算量/10⁶	参数量/10³	UAV123			DTB70
堆叠次数N	浮点数运算量/10⁶	参数量/10³	准确率	成功率	速度/fps	准确率	成功率	速度/fps
1	91.16	436.81	0.512	0.455	198.35	0.522	0.479	208.67
2	96.04	475.73	0.558	0.479	183.96	0.586	0.504	200.01
3	100.92	478.66	0.612	0.522	172.46	0.636	0.533	192.74
4	105.80	499.59	0.759	0.575	161.70	0.779	0.588	186.58
5	110.68	520.52	0.798	0.602	149.68	0.801	0.613	180.42
6	115.56	541.45	0.812	0.618	142.77	0.821	0.634	174.69
7	120.45	562.37	0.801	0.601	133.24	0.818	0.632	165.36
8	125.34	583.29	0.806	0.599	128.69	0.814	0.630	158.64

[1]	Nuo MA, Shechun WEI, Junhui MENG, Qingyang LIU, Yusheng LEI. Flow field characteristics and dynamics of internal supply chamber separating from UAV considering effect of deceleration parachutes [J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(3): 130755-130755.
[2]	Yiquan WU, Kang TONG. Research advances on deep learning-based small object detection in UAV aerial images [J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(3): 30848-030848.
[3]	Ji MA, Xiangyong CHEN, Guanghui WEN, Long CHENG, Jianlong QIU. Predefined time tracking control of unmanned aerial vehicles under stochastic switching topology [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(S1): 730793-730793.
[4]	Ting LIU, Guoxin ZHOU, Yang XU, Delin LUO, Zhengyu GUO, Mengjie YANG. Optimised Harris hawks multi-UAV dynamic target search with fused infographics [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(S1): 730773-730773.
[5]	Haifeng WANG, Kunpeng LIU, Hongxin JIANG, Chenxi DU. Aerodynamic optimization method of propeller multi⁃design points and variable pitch angle strategy [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(9): 528831-528831.
[6]	Jing ZHAO, Dan SONG. Integrity monitoring method for GNSS/IMU integrated navigation system of UAV [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(7): 328943-328943.
[7]	Chuanyun WANG, Yang SU, Linlin WANG, Tian WANG, Jingjing WANG, Qian GAO. Multi-object continuous robust tracking algorithm for anti-UAV swarm [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(7): 329017-329017.
[8]	Xudong LUO, Yiquan WU, Jinlin CHEN. Research progress on deep learning methods for object detection and semantic segmentation in UAV aerial images [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(6): 28822-028822.
[9]	Hongyu YIN, Yu WU, Tianjiao LIANG. Cooperative path planning for patrol coverage of fixed wing UAV [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(6): 328944-328944.
[10]	Chuihuan KONG, Dawei WU, Zhaoguang TAN, Lijun PAN, Rubing MA, Jiangtao SI. Design of fully electric scheme for three⁃surface verification aircraft [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(6): 629618-629618.
[11]	Guangjia LI, Hongbo WANG, Kai ZHANG, Zhisheng YI. Lift enhancement and drag reduction technologies of solar powered unmanned aerial vehicles in near space: Review [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(5): 529644-529644.
[12]	Yunpeng CAI, Dapeng ZHOU, Jiangchuan DING. Intelligent collaborative control of UAV swarms with collision avoidance safety constraints [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(5): 529683-529683.
[13]	Rui SI, Yong CHEN. Application trends of additive manufacturing technology for civil aircraft [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(5): 529677-529677.
[14]	Weihong ZHANG, Changhong TANG. Lightweighting of aerospace and aeronautical equipment: Challenges and perspectives [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(5): 529965-529965.
[15]	Gaojie ZHENG, Xiaoming HE, Dongpo LI, Huijun TAN, Kun WANG, Zhenlong WU, Depeng WANG. Double 90° deflection inlet/volute coupling flow characteristics of tail-powered unmanned aerial vehicle [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(4): 128782-128782.