本文提出一种结合声学和光学定位手段的UUV水下回收方法，并重点对视觉定位进行研究。针对光照剧烈变化情况下视觉定位稳定性差的问题，设计一种基于先验信息的目标轮廓提取算法，充分利用环境亮度信息和合作目标的物理属性，通过筛选兴趣区（ROI）、动态阈值等方法，有效解决图像背景亮度过高或亮度不均衡导致的目标漏检、误检等问题，提升对目标轮廓的分割精度，从而优化目标图像坐标及空间坐标的计算准度。在湖验中应用该算法，实现全天时、全天候对UUV目标的可靠检测与准确定位，验证了该算法对环境亮度变化的鲁棒性。
This paper proposes a method for UUV underwater recovery that combines acoustic and optical positioning methods, with a focus on visual positioning. Aiming at the problem of poor stability of visual positioning in the case of dramatic change of brightness, a target contour extraction algorithm based on prior information is designed. It makes full use of the environmental brightness information and the physical attributes of the cooperative target. Solving the problems of target missing and false detection caused by the high or uneven brightness of the background by filtering the region of interest (ROI), dynamic threshold and other methods, so as to optimize the accuracy of target image coordinates and spatial coordinates. The algorithm is applied in the lake experiment to realize the reliable detection and accurate positioning of UUV targets, all-day and all-weather, which verifies that this algorithm is robust to change of environment brightness.
2024,46(10): 98-101 收稿日期：2023-06-12
DOI：10.3404/j.issn.1672-7649.2024.10.017
分类号：TP242
作者简介：王嘉(1996-),男,硕士,助理工程师,研究方向为水下无人系统
参考文献：
[1] 黎南, 湛鑫, 陈涛, 等. UUV水下回收中的视觉和短基线定位融合[J]. 智能系统学报, 2013, 8(2): 156-161.
LI N, ZHAN X, CHEN T, et al. Data fusion method of vision and SBL position for UUV underwater docking[J]. CAAI Transactions on Intelligent Systems, 2013, 8(2): 156-161.
[2] 严浙平, 王璐. UUV水下定位方法的研究现状与进展[J]. 哈尔滨工程大学学报, 2017, 38(7): 989-1000.
YAN Z P, WANG L. Research status and progress of UUV underwater localization[J]. Journal of Harbin Engineering University, 2017, 38(7): 989-1000.
[3] 蔡迎波, 李德彪. 基于单目视觉的AUV水下定位方法[J]. 中国惯性技术学报, 2015, 23(4): 489-492.
CAI Y B, LI D B. AUV underwater positioning method based on monocular-vision[J]. Journal of Chinese Inertial Technology, 2015, 23(4): 489-492.
[4] 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 W, LI Z X, GONG P, et al. A positioning method of monocular visual guided light array in UUV docking system[J]. Journal of Unmanned Undersea Systems, 2019, 27(3): 313-318.
[5] 张伟, 潘珺, 宫鹏, 等. 面向UUV回收过程的单目视觉导引灯阵跟踪方法[J]. 水下无人系统学报, 2021, 29(4): 435-441.
ZHANG W, PAN J, GONG P, et al. Monocular vision guided lamp array tracking method for the UUV recovery process[J]. Journal of Unmanned Undersea Systems, 2021, 29(4): 435-441.
[6] 韩泽凯, 朱兴华, 韩晓军, 等. 基于卷积神经网络目标跟踪的AUV回收视觉导引算法[J]. 水下无人系统学报, 2022, 30(6): 801-808.
HAN Z K, ZHU X H, HAN X J, et al. Visual guidance algorithm for AUV recovery based on CNN object tracking[J]. Journal of Unmanned Undersea Systems, 2022, 30(6): 801-808.
[7] 刘建业, 庄苏锋, 屠大维, 等. 双目立体视觉成像系统水下建模与标定方法[J]. 光子学报, 2022, 51(12): 11-23.
LIU J Y, ZHUANG S F, TU D W, et al. Underwater modeling and calibration method of binocular stereo vision imaging system[J]. Acta Photonica Sinica, 2022, 51(12): 11-23.
[8] 陈建平, 秦斌, 王欣. 非均匀光照图像的自适应阈值分割[J]. 湖南工业大学学报, 2016, 30(4): 36-40.
CHEN J P, QIN B, WANG X. An adaptive threshold segmentation method based on inhomogeneous illumination images[J]. Journal of Hunan University of Technology, 2016, 30(4): 36-40.
[9] 白雪. 自适应阈值法的舰船图像自动分割简[J]. 舰船科学技术, 2018, 40(6): 136–138.
BAI X. Automatic segmentation of ship image based on adaptive threshold method[J]. Ship Science and Technology, 2018, 40(6): 136–138.
[10] 杨琨, 曾立波, 王殿成. 数学形态学腐蚀膨胀运算的快速算法[J]. 计算机工程与应用, 2005, 41(34): 54-56.
YANG K, ZENG L B, WANG D C. A fast arithmetic for the erosion and dilation operations of mathematical morphology[J]. Computer Engineering and Applications, 2005, 41(34): 54-56.
