多特征融合的无人船视觉目标跟踪

吴伟, 王宁, 王元元, 孙赫男

doi:10.16411/j.cnki.issn1006-7736.2023.04.005

大连海事大学学报 >

2023 , Vol. 49 >Issue 4: 37 - 45

DOI: https://doi.org/10.16411/j.cnki.issn1006-7736.2023.04.005

多特征融合的无人船视觉目标跟踪

展开

(1.大连海事大学 a. 船舶电气工程学院；b. 轮机工程学院，辽宁大连 116026；2.大连海事局甘井子海事处，辽宁大连 116000)

吴伟（1997 — ），男，硕士，研究方向：目标跟踪。E-mail：wuweimaster100@163.com。王宁*（1983 — ），男，博士，教授，博士生导师，研究方向：自主系统智能控制等。E-mail：n. wang. dmu. cn@ gmail. com。王元元（1990 — ），女，博士生，研究方向：目标检测等。E-mail：yyuanwang1010@163.com。孙赫男（1982 — ），男，学士，研究方向：海事管理。

收稿日期: 2023-05-05

修回日期: 2023-07-20

录用日期: 2023-07-20

网络出版日期: 2023-12-15

基金资助

国家自然科学基金资助项目(52271306)；船舶总体性能创新研究开放基金(31422120)

收起

Multi-feature fusion-based visual target tracking for unmanned surface vehicles

Expand

(1a. Marine Electrical Engineering College; 1b. Marine Engineering College, Dalian Maritime University, Dalian 116026, China；2. Ganjingzi Maritime Department, Dalian Maritime Safety Administration, Dalian 116000, China)

Received date: 2023-05-05

Revised date: 2023-07-20

Accepted date: 2023-07-20

Online published: 2023-12-15

Fold

摘要

针对海面光照变化、水花遮挡、水面倒影等耦合作用引起的无人船视觉目标跟踪漂移问题，提出一种基于多特征融合的尺度自适应相关滤波跟踪算法。通过多特征融合，增强了水面目标特征表达，避免了目标跟踪漂移。为减少环境干扰对跟踪目标外观描述的影响，设计尺度自适应跟踪滤波器，提升目标跟踪鲁棒性能。采用本文算法对多个代表性海上视频数据集进行验证，并与典型目标跟踪算法进行比较。结果表明，相较于基于单一CN特征的跟踪算法，本文算法的平均重叠率提升23.63%、平均中心误差减少53.79个像素点。；本文算法适用于处理由于海面环境剧烈变化、目标尺度变化导致的跟踪漂移问题，可为无人船作业自主性提供重要智能感知技术支持。

关键词： 无人船(USV); 视觉目标跟踪; 自适应相关滤波; 多特征融合

本文引用格式

吴伟, 王宁, 王元元, 孙赫男 . 多特征融合的无人船视觉目标跟踪[J]. 大连海事大学学报, 2023 , 49(4) : 37 -45 . DOI: 10.16411/j.cnki.issn1006-7736.2023.04.005

Abstract

A multi-feature fusion-based scale-adaptive correlation filter tracking algorithm was proposed to address the problem of visual target tracking drift in unmanned surface vehicles (USVs) caused by the coupling effects of sea surface illumination changes, water spray occlusion, water surface reflections, etc. By using multi-feature fusion, the algorithm enhanced the feature expression of the water surface target, thereby avoiding target tracking drift. To reduce the influence of environmental interference on the appearance description of the tracked target, a scale-adaptive tracking filter was designed to improve the robustness of target tracking. Several representative offshore video datasets were used to compare and analyze the proposed algorithm against typical target tracking algorithms. Results show that compared to a tracking algorithm based on a single CN feature, the proposed algorithm achieved 23.63% increase in average overlap score and 53.79 pixel reduction in average center location error. The proposed algorithm is suitable for handling the problem of tracking drift caused by drastic changes in sea surface environment and target scale, which can provide important intelligent perception technology support for the autonomy of unmanned surface vehicles operations.

Key words： unmanned surface vehicle（USV); visual target tracking; adaptive correlation filter; multi-feature fusion

参考文献

[1]WANG N, WANG Y, ER M J. Review on deep learning techniques for marine object recognition: Architectures and algorithms[J]. Control Engineering Practice, 2022, 118: 104458.

[2]WANG N, CHEN T, LIU S, et al. Deep learning-based visual detection of marine organisms: A survey[J]. Neurocomputing, 2023, 532: 1-32.

[3]CHEN T, WANG N, WANG R, et al. One-stage CNN detector-based benthonic organisms detection with limited training dataset[J]. Neural Networks, 2021, 144: 247-259.

[4]唐路源, 赵红, 王宁, 等. 基于M-DCGAN的缺陷检测数据集增广方法[J]. 大连海事大学学报,2023, 49(2): 148-160.

TANG L, ZHAO H, WANG N, et al. An augmentation method of defect detection dataset based on M-DCGAN[J]. Journal of Dalian Maritime University, 2023, 49(2): 148-160. (in Chinese)

[5]王宁, 王仁慧, 鲁挺. 推进器饱和约束的水面无人艇固定时间精准跟踪控制[J]. 控制理论与应用, 2023, 40(1):10.

WANG N, WANG R, LU T, et al. Fixed-time precision tracking control of an unmanned surface vehicle constrained by thruster saturations[J]. Control Theory & Applications,2023,40(1):10. (in Chinese)

[6]WANG N, GAO Y, LIU Y, et al. Self-learning-based optimal tracking control of an unmanned surface vehicle with pose and velocity constraints[J]. International Journal of Robust and Nonlinear Control,2022,32(5):2950-2968.

[7]HE H, WANG N. Monocular Visual Servo of Unmanned Surface Vehicles with View-field Constraints[C]//2021 33rd Chinese Control and Decision Conference (CCDC).Kunming,China,2021:973-978.

[8]何红坤, 王宁. 欠驱动无人船单目视觉伺服镇定控制[J]. 中国舰船研究,2022,17(5):166-174.

HE H, WANG N. Monocular visual servo-based stabilization control of underactuated unmanned surface vehicle[J]. Chinese Journal of Ship Research, 2022,17(5):166–174. (in Chinese)

[9]WANG N, GAO Y, YANG C, et al. Reinforcement learning-based finite-time tracking control of an unknown unmanned surface vehicle with input constraints[J]. Neurocomputing, 2022, 484: 26-37.

[10]JAVED S, DANELLJAN M, KHAN F S, et al. Visual object tracking with discriminative filters and siamese networks: a survey and outlook[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022.

[11]HUANG J. Auto-attentional mechanism in multi-domain convolutional neural networks for improving object tracking[J]. International Journal of Intelligent Computing and Cybernetics,2021,15(1):41-60.

[12]DANELLJAN M, GOOL L V, TIMOFTE R. Probabilistic regression for visual tracking[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition.2020:7183-7192.

[13]NAM H, HAN B. Learning multi-domain convolutional neural networks for visual tracking[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.2016:4293-4302.

[14]ZHU Z, WU W, ZOU W, et al. End-to-end flow correlation tracking with spatial-temporal attention[C]//Proceedings of the IEEE conference on computer vision and pattern recognition.2018:548-557.

[15]RAHMAD C, ASMARA R A, PUTRA D R H, et al. Comparison of Viola-Jones Haar Cascade classifier and histogram of oriented gradients (HOG) for face detection[C]//IOP conference series: materials science and engineering. IOP Publishing, 2020, 732(1): 012038.

[16]DANELLJAN M, SHAHBAZ K F, FELSBERG M, et al. Adaptive color attributes for real-time visual tracking[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2014: 1090-1097.

[17]DANELLJAN M, HÄGER G, KHAN F, et al. Accurate scale estimation for robust visual tracking[C]//British Machine Vision Conference, Nottingham, September 1-5, 2014. Bmva Press, 2014.

[18]BOLME D S, BEVERIDGE J R, DRAPER B A, et al. Visual object tracking using adaptive correlation filters[C]//2010 IEEE computer society conference on computer vision and pattern recognition. IEEE, 2010: 2544-2550.

[19]HENRIQUES J F, CASEIRO R, MARTINS P, et al. Exploiting the circulant structure of tracking-by-detection with kernels[C]//European conference on computer vision. Springer, Berlin, Heidelberg, 2012: 702-715.

[20]BERTINETTO L, VALMADRE J, GOLODETZ S, et al. Staple: Complementary learners for real-time tracking[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 1401-1409.

[21]DANELLJAN M, ROBINSON A, SHAHBAZ K F, et al. Beyond correlation filters: Learning continuous convolution operators for visual tracking[C]//European conference on computer vision. Springer, Cham, 2016: 472-488.

[22]DANELLJAN M, BHAT G, SHAHBAZ K F, et al. Eco: Efficient convolution operators for tracking[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2017: 6638-6646.

[23]LIU T, WANG G, YANG Q. Real-time part-based visual tracking via adaptive correlation filters[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015: 4902-4912.

[24]MONTERO A S, LANG J, LAGANIERE R. Scalable kernel correlation filter with sparse feature integration[C]//2015 IEEE International Conference on Computer Vision Workshop (ICCVW). IEEE, 2015: 587-594.

[25]LI Y, FU C, DING F, et al. AutoTrack: Towards high-performance visual tracking for UAV with automatic spatio-temporal regularization[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2020: 11923-1193.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献