面向UUV回收过程的单目视觉导引灯阵跟踪方法

张 伟; 潘 珺; 宫 鹏; 李子轩; 郭帅克; 高赛博

doi:10.11993/j.issn.2096-3920.2021.04.010

面向UUV回收过程的单目视觉导引灯阵跟踪方法

doi: 10.11993/j.issn.2096-3920.2021.04.010

哈尔滨工程大学智能科学与工程学院, 黑龙江哈尔滨, 150001

基金项目: 国家自然科学基金(5207110590).

详细信息

通讯作者:
潘珺(1993-), 男, 在读硕士, 主要研究方向为控制科学与工程.

中图分类号: TJ630.33 TP242.62
计量
- 文章访问数: 725
- HTML全文浏览量: 17
- PDF下载量: 67
- 被引次数: 0
出版历程
- 刊出日期: 2021-08-31

Monocular Vision Guided Lamp Array Tracking Method for the UUV Recovery Process

College of Automation, Harbin Engineering University, Harbin 150001, China

摘要

摘要: 无人水下航行器(UUV)在进行水下作业过程中, 受到自身可携带载荷和环境的制约, 导致其作业时间不能过长和作业距离不能过远, 需要在一定时间内回收以供给能量、获得数据并布置新任务。深入研究UUV视觉跟踪问题, 对实现UUV快速、准确、安全的回收有着重要意义。基于此, 文中对UUV回收的单目视觉多目标跟踪问题进行了研究。以非线性卡尔曼滤波器为基础, 提出基于非对称光阵列的UUV回收跟踪方法。首先, 对单目摄像机采集的水下图像进行预处理, 提取光源阵列的特征信息。其次, 初始化目标轨迹, 提出了基于非对称L型光源阵列的轨迹确定方法。再次, 权衡观测器与滤波器设计了寻优代价矩阵, 应用改进的匈牙利算法将检测与目标轨迹相匹配, 并根据匹配结果对卡尔曼滤波器的先验结果进行修正。最后, 通过水下目标跟踪实验, 验证了所提方法的工程有效性。
- 无人水下航行器 /
- 回收 /
- 多目标跟踪 /
- 单目视觉 /
- 非线性卡尔曼滤波器
Abstract: An unmanned undersea vehicle(UUV) is restricted by its carrying load and environment during underwater operation, reducing the maximum operation time and distance. As a result, periodical UUV recovery is necessary to supply it with energy, obtain data, and arrange new tasks. Further study on UUV vision tracking is important to realize a rapid, accurate, and safe UUV recovery. In this study, problems caused by the monocular visual multi-target tracking for UUV recovery are investigated. A recovery tracking method of UUVs based on the nonlinear Kalman filter and asymmetric light array is proposed. First, underwater images collected by the monocular camera are preprocessed to extract characteristic information of the light source array. Afterward, the target trajectory is determined through a method based on the asymmetric L-shaped light source array. The optimal cost matrix is thereafter established by weighing the observer and the filter, and the detection is matched to the target trajectory using the improved Hungarian algorithm. Prior results of the Kalman filter are modified according to these results. Finally, the engineering effectiveness of the proposed method is verified through underwater target tracking experiments.
- unmanned undersea vehicle /
- recovery /
- multi-target tracking /
- monocular vision /
- nonlinear Kalman filter

HTML全文

参考文献(14)

[1]	钟宏伟. 国外无人水下航行器装备与技术现状及展望[J]. 水下无人系统学报, 2017, 25(3): 215-225. Zhong Hong-wei. Review and Prospect of Equipment and Techniques for Unmanned Undersea Vehicle in Foreign Countries[J]. Journal of Unmanned Undersea Systems, 2017, 25(3): 215-225.
[2]	孙大军, 郑翠娥, 张居成, 等. 水声定位导航技术的发展与展望[J]. 中国科学院院刊. 2019, 34(3): 331-338. Sun Da-jun, Zheng Cui-e, Zhang Ju-cheng, et al. Development and Prospect for Underwater Acoustic Positioning and Navigation Technology[J]. Bulletin of Chinese Academy of Sciences, 2019, 34(3): 331-338.
[3]	Henriques J F, Caseiro R, Martins P, et al. High-Speed Tracking with Kernelized Correlation Filters[C]//IEEE Conference on Industrial Electronics and Applications. Wollongong, NSW, Australia: IEEE, 2016.
[4]	Wirtz M, Hildebrandt M, Gaudig C. Design and Test of a Robust Docking System for Hovering AUVs[C]//Oceans 2012. Yeosu, South Korea: IEEE, 2012.
[5]	Zhang K, Zhang L, Yang M H, et al. Fast Tracking via Spatio-Temporal Context Learning[EB/OL]. arXiv: 1311. 1939, 2013. (2013-11-08)[2020-03-05]. https://www. researchgate.net/publication/258345794.
[6]	边信黔, 王晓娟．基于散射模型的水下图像复原[J]．机器人, 2010, 32(6): 721-725． Bian Xin-qian, Wang Xiao-juan. Underwater Image Restoration Based on Scattering Model[J]. Robot, 2010, 32(6): 721-725.
[7]	马永华, 耿瑞芳, 张罡. 基于Blob算法的机器视觉图像识别方法的研究[J]. 仪器仪表用户, 2008, 15(4): 1-2. Ma Yong-hua, Geng Rui-fang, Zhang Gang. Research on Machine Vision Image Recognition Method Based on Blob Algorithm[J]. Instrument Users, 2008, 15(4): 1-2.
[8]	Wu Y, Lim J, Yang M H. Object Tracking Benchmark[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 37(9): 1834-1848.
[9]	Otsu N. A Threshold Selection Method from Gray-Level Histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 2007, 9(1): 62-66.
[10]	Zhang W, Wang X, Chen T, et al. Fast Target Extraction Based on Bayesian BLOB Analysis and Simulated Annealing for Underwater Images[J]. IEEE Journal on Robotics and Automation, 2017, 32(2): 101-108.
[11]	Einicke G A, White L B. Robust Extended Kalman Filtering[J]. IEEE Transactions on Signal Processing, 1999, 47(9): 2596-2599.
[12]	Garcia-Fernandez F, Morelande R, Grajal J, et al. Truncated Unscented Kalman Filtering[J]. IEEE Transactions on Signal Processing, 2012, 60(7): 3372-3386.
[13]	Bernardin K, Stiefelhagen R. Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics[J/OL]. EURASIP Journal on Image and Video Processing, 2008(1). [2020-03-05]. http://citeseerx.ist.psu.edu/my-citeseer/login.
[14]	张伟, 李子轩, 宫鹏, 等. UUV接驳单目视觉导引灯阵定位方法[J]. 水下无人系统学报, 2019, 27(3): 313-318. Zhang Wei, Li Zi-xuan, Gong Peng, et al. Positioning Method of Monocular Vision Guided Lamp Array with UUV Connection[J]. Journal of Underwater Unmanned Systems, 2019, 27(3): 313-318.