基于惯性预测的遥感场景舰船多目标跟踪模型

doi:10.12305/j.issn.1001-506X.2025.01.05

Abstract

Abstract:

Ship object tracking in remote sensing scenario is of great strategic significance and economic value. How to overcome the influence of ship orientation arbitrariness, close arrangement of offshore ship targets on tracking performance from remote sensing perspective is a challenging task. Multiple-object tracking (MOT) tasks with large aspect ratio such as ships in remote sensing scenarios are studied. An inertial predicting multiple object tracker (IPMOT) is proposed. Firstly, the tracing-by-detection (TBD) paradigm is used to cascade the detector and tracker to effectively avoid the dependence of the training process on the time sequence relationship, and the detector is trained through the open target detection data set to solve the problem of lack of tracking data set. Secondly, addressing the problem of missed detections in the detection phase of the TBD paradigm, which significantly affects tracking performance, an inertial tracking model (ITM) is built. It uses multi-step prediction to track the errors of detector, and uses angle correction to eliminate the impact of angle mutation at the edge. Finally, in order to realize the model training and performance testing of the proposed algorithm, the ship MOT (SMOT) dataset is constructed. The experimental results show that the proposed model improves the performance of MOT accuracy (MOTA) and inverse document frequency (IDF1) indexes by 3.9% and 7.2%, respectively, and the performance of IDs and Frag indexes is significantly improved, with good tracking accuracy and stability.

Key words: remote sensing scenario, ship target, multiple-object tracking (MOT), inertial predicting, tracing-by-detection (TBD) paradigm

CLC Number:

P237

Chaofan PAN, Runsheng LI, Qing HU, Quanfu BAO, Yongqiang BAO. Ship multiple-object tracking model in remote sensing scene based on inertial prediction[J]. Systems Engineering and Electronics, 2025, 47(1): 41-51.

Figures/Tables 12

Fig.1

Fig.2

Fig.3

Fig.4

Table 1

Fig.5

Table 2

Table 3

Table 4

Fig.6

Table 5

Fig.7

References 30

1	BOLME D S, BEVERIDGE J R, DRAPER B A, et al. Visual object tracking using adaptive correlation filters[C]//Proc. of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2010: 2544-2550.
2	HENRIQUES J F, CASEIRO R, MARTINS P, et al. Exploiting the circulant structure of tracking-by-detection with kernels[C]// Proc. of the European Conference on Computer Vision, 2012: 702-715.
3	HENRIQUES J F , CASEIRO R , MARTINS P , et al. High-speed tracking with kernelized correlation filters[J]. IEEE Trans. on Pattern Analysis & Machine Intelligence, 2015, 37 (3):583-596.
4	KIANI G H, FAGG A, LUCEY S. Learning background-aware correlation filters for visual tracking[C]//Proc. of the IEEE International Conference on Computer Vision, 2017: 1135-1143.
5	LI Y, ZHU J K. A scale adaptive kernel correlation filter tracker with feature intergration[C]//Proc. of the European Conference on Computer Vision, 2014: 254-265.
6	YUN S D, CHOI J W, YOO Y J, et al. Action-decision networks for visual tracking with deep reinforcement learning[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 2711-2720.
7	刘志强, 任世恒. 复杂场景下基于OSA改进的多目标跟踪算法研究[J]. 信息技术, 2022, (4): 123- 129.
	LIU Z Q , REN S H . Research on multi-target tracking algorithm based on improved OSA in complex scenes[J]. Information Technology, 2022, (4): 123- 129.
8	王彬彬. 基于SVM与Meanshift跟踪算法的运动视频目标跟踪[J]. 现代电子技术, 2022, 45 (1): 56- 60.
	WANG B B . Moving video object tracking based on SVM and meanshift tracking algorithm[J]. Modern Electronics Technique, 2022, 45 (1): 56- 60.
9	张文利, 辛宜桃, 杨堃, 等. 基于改进的Transformer加Anchor-free网络的多目标跟踪算法[J]. 测控技术, 2022, 41 (2): 20- 28.
	ZHANG W L , XIN Y T , YANG K , et al. Improved Transformer plus Anchor-free network based on multi-object tracking algorithm[J]. Measurement & Control Technology, 2022, 41 (2): 20- 28.
10	BEWLEY A, GE Z, OTT L, et al. Simple online and realtime tracking[C]//Proc. of the IEEE International Conference on Image Processing, 2016: 3464-3468.
11	WOJKE N, BEWLEY A, PAULUS D. Simple online and realtime tracking with a deep association metric[C]//Proc. of the IEEE International Conference on Image Processing, 2017: 3645-3649.
12	ZHANG Y F, SUN P Z, JIANG Y, et al. ByteTrack: multi-object tracking by associating every detection box[EB/OL]. [2023-01-07]. https://arXivpreprintarXiv:2110.06864, 2021.
13	WANG Z D, ZHENG L, LIU Y X, et al. Towards real-time multi-object tracking[C]//Proc. of the European Conference on Computer Vision, 2020: 107-122.
14	ZHANG Y F, WANG C Y, WANG X G, et al. A simple baseline for multi-object tracking[EB/OL]. [2023-01-07]. https://arXivpreprintarXiv:2004.01888.
15	PENG J L, WANG C G, WAN F B, et al. Chained-tracker: chaining paired attentive regression results for end-to-end joint multiple-object detection and tracking[C]//Proc. of the European Conference on Computer Vision, 2020: 145-161.
16	ZHOU X Y, KOLTUN V, KRÄHENBVHL P. Tracking objects as points[C]//Proc. of the European Conference on Computer Vision, 2020: 474-490.
17	SUN P Z, JIANG Y, ZHANG R F, et al. Transtrack: multiple object tracking with transformer[EB/OL]. [2023-01-07]. https://arXivpreprintarXiv:2012.15460, 2020.
18	MEINHARDT T, KIRILLOV A, LEAL-TAIXE L, et al. Trackformer: multi-object tracking with transformers[C]//Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 8844-8854.
19	ZENG F G, DONG B, ZHANG Y, et al. Motr: end-to-end multiple-object tracking with transformer[EB/OL]. [2023-01-07]. https://arXivpreprintarXiv:2105.03247, 2021.
20	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proc. of the Advances in Neural Information Processing Systems, 2017.
21	SHI X J, CHEN Z R, WANG H, et al. Convolutional LSTM network: a machine learning approach for precipitation nowcasting[C]//Proc. of the 28th International Conference on Neural Information Processing Systems, 2015: 802-810.
22	KALMAN R E . A new approach to linear filtering and prediction problems[J]. Journal of Basic Engineering, 1960, 82 (D): 35- 45.
23	ZAREMBA W, SUTSKEVER I, VINYALS O. Recurrent neural network regularization[EB/OL]. [2023-01-07]. https://arXivpreprintarXiv:1409.2329, 2014.
24	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770-778.
25	BALDUZZI D, FREAN M, LEARY L, et al. The shattered gradients problem: if resnets are the answer, then what is the question?[C]//Proc. of the International Conference on Machine Learning, 2017: 342-350.
26	LECUN Y , BOTTOU L , BENGIO Y , et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86 (11): 2278- 2324.
27	ZHANG F , WANG X Y , ZHOU S L , et al. Arbitrary-oriented ship detection through center-head point extraction[J]. IEEE Trans. on Geoscience and Remote Sensing, 2021, 60 (1): 5612414.
28	BERNARDIN K , STIEFELHAGEN R . Evaluating multiple object tracking performance: the clear mot metrics[J]. EURASIP Journal on Image and Video Processing, 2008, 246309.
29	RISTANI E, SOLERA F, ZOU R S, et al. Performance mea-sures and a data set for multi-target, multi-camera tracking[C]// Proc. of the European Conference on Computer Vision, 2016: 17-35.
30	LI Y, HUANG C, NEVATIA R. Learning to associate: hybridboosted multi-target tracker for crowded scene[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2009: 2953-2960.

变量名	含义
<frame>	帧序
<target id>	轨迹编号
<rbox_x>	定向框中心点横坐标
<rbox_y>	定向框中心点纵坐标
<rbox_w>	定向框宽度
<rbox_h>	定向框长度
<rbox_angle>	定向框角度
<tr>	是否忽略目标
<classes>	目标类别
<visibility ratio>	目标保留率

数据集	视频序列数目	轨迹数目	真值框数目	帧数	分辨率	目标类别
训练集	12	299	107 940	1 000~1 300	800×800	1: Air. 2: Was. 3: Tar 4: Oth. 5: Aus. 6: Whi. 7: San. 8: New. 9: Tic. 10: Bur. 11: Per. 12: Lew. 13: Sup. 14: Kai. 15: Hop. 16: Mer. 17: Fre. 18: Ind. 19: Ave. 20: Sub.
验证集	3	64	30 051	1 000~1 300	800×800
测试集	5	107	44 661	1 000~1 300	800×800

ResLSTM	隐藏层宽度	MOTA↑/%	MOTP↑/%	IDF1↑/%	IDs↓	MT↑/%	Frag↓	Size↓/kB
结构1	32	60.4	79.9	64.3	162	45.2	981	159
	64	59.9	79.8	62.7	179	46.9	1 002	483
	128	59.9	79.9	63.1	185	48.1	980	1 709
	256	59.7	79.9	63.4	173	43.3	996	6 463
	512	59.8	79.8	62.8	172	44.5	995	25 187
结构2	32	53.0	80.1	56.6	281	37.2	1 227	440
	64	53.4	79.3	56.9	299	38.9	1 254	1 362
	128	54.9	79.9	59.8	238	39.9	1 236	4 934
	256	55.0	79.9	59.2	253	41.8	1 202	18 990
	512	56.8	79.8	61.9	213	43.3	1 141	74 750
结构3	32	60.9	79.8	65.3	161	46.9	985	94
	64	61.2	79.8	65.7	163	48.8	977	253
	128	59.7	79.9	64.1	173	44.0	990	858
	256	59.4	79.9	63.4	183	44.0	1 028	3 220
	512	59.8	79.8	63.6	186	45.2	1 007	12 552

方法	ResLSTM	ACM	MOTA↑/%	MOTP↑/%	IDF1↑/%	IDs↓	MT↑/%	ML↓/%	Frag↓	FP↓	FN↓	FPS↑
基线模型	√	-	57.3	80.2	58.5	226	35.6	8.5	1 490	3 235	15 868	14.0
	√	-	59.2	79.6	62.6	190	45.6	6.1	1 020	4 625	13 614	14.2
	√	√	58.2	80.3	60.5	186	36.5	9.3	1 463	3 064	15 665	14.4
本文方法	√	√	61.2	79.8	65.7	163	48.8	6.1	977	4 054	13 371	14.1

方法	MOTA↑/%	IDF1↑/%	IDs↓	MT↑/%	ML↓/%	Frag↓	FP↓	FN↓	FPS↑
SORT^[10]	54.8	42.8	811	34.4	11.6	1 144	2 766	16 925	43.1
ByteTrack^[11]	55.4	44.0	730	38.5	10.7	1 076	4 007	15 461	40.8
DeepSORT^[12]	57.3	58.5	226	35.6	8.5	1 490	3 235	15 868	14.0
IPMOT	61.2	65.7	163	48.8	6.1	977	4 054	13 371	14.1

Ship multiple-object tracking model in remote sensing scene based on inertial prediction

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 12

References 30

Related Articles 12

Recommended Articles

Metrics

Comments

[1]	Hongmeng CHEN, Jun LI, Jing LIU, Wei HUANG, Yingjie ZHANG, Yan CHEN, Yaobing LU. SAR-ISAR hybrid imaging method for sea surface ship target based on Radon time-frequency analysis [J]. Systems Engineering and Electronics, 2025, 47(1): 109-116.
[2]	Qian CHENG, Jia LI, Juan DU. Ship target detection algorithm of optical remote sensing image based on YOLOv5 [J]. Systems Engineering and Electronics, 2023, 45(5): 1270-1276.
[3]	Runlin LI, Huanxin ZOU, Xu CAO, Fei CHENG, Shitian HE, Meilin LI. Multi-direction remote sensing ship detection based on center point and semantic information [J]. Systems Engineering and Electronics, 2022, 44(6): 1772-1781.
[4]	Chaofan PAN, Runsheng LI, Yan XU, Qing HU, Chaoyang NIU, Wei LIU. Ship detection of optical remote sensing images based on aware vectors [J]. Systems Engineering and Electronics, 2022, 44(12): 3631-3640.
[5]	Yu LEI, Xiangguang LENG, Xiaoyan ZHOU, Zhongzhen SUN, Kefeng JI. Recognition method of ship target in complex SAR image based on improved ResNet network [J]. Systems Engineering and Electronics, 2022, 44(12): 3652-3660.
[6]	Xiaoya JIA, Hongqiao WANG, Yadan YANG, Zhongma CUI, Bin XIONG. Anchor free SAR image ship target detection method based on the YOLO framework [J]. Systems Engineering and Electronics, 2022, 44(12): 3703-3709.
[7]	Yonggang LI, Weigang ZHU, Qiongnan HUANG, Yuntao LI, Yonghua HE. Near-shore ship target detection with SAR images in complex background [J]. Systems Engineering and Electronics, 2022, 44(10): 3096-3103.
[8]	ZHAO Zhiguo, XIAO Hui. Approach to correct ionospheric phase perturbation based on Toeplitz matrix [J]. Systems Engineering and Electronics, 2017, 39(5): 1024-1029.
[9]	GUO Shao-jun, SHEN Tong-sheng, XU Jian, MA Xin-xing. Detection of multi-ship targets at sea based on ObjectNess BING [J]. Systems Engineering and Electronics, 2016, 38(1): 14-20.
[10]	ZHAO Zhiguo, XIAO Hui. Spread Doppler clutter suppression approach in FRFT field for hybrid skysurfacewave radars [J]. Systems Engineering and Electronics, 2014, 36(12): 2411-2416.
[11]	ZHAO Zhi-guo， CHEN Jian-wen， BAO Zheng. Modified adaptive ocean clutter suppression approach in OTHR [J]. Journal of Systems Engineering and Electronics, 2012, 34(5): 909-914.
[12]	YU Ding-feng，CHEN Qian，YAO Jing-jing，HE Si-yuan，ZHU Guo-qiang，HU Wei-dong. Simulation of HRRP characteristics for maritime ship targets [J]. Journal of Systems Engineering and Electronics, 2010, 32(12): 2552-2556.