水下连接器的对接作业是水下工程作业的重要环节，针对目前依靠摄像机传回视频进行作业过程中存在的操作难度大、依赖操作员经验的问题，设计基于双目视觉的水下连接器位姿测量方法。该方法首先根据水下连接器的颜色特征确定检测范围，之后在检测范围中以水下连接器为模板进行初步定位，然后根据水下连接器端面的成像特点检测椭圆特征，并进行双目匹配获得相关三维点坐标，最后计算得出水下连接器的位姿。实验表明，该方法位置测量平均误差1.3%，姿态测量平均误差3.5°，可以较好地为水下连接器对接作业提供参考。
The docking operation of underwater connectors is an important part of underwater engineering operations. In view of the problems existing in the operation process of relying on the video returned by the camera, such as the difficulty of operation and the dependence on the operator's experience, a method of underwater connector pose measurement based on binocular vision is designed. This method firstly determines the detection range according to the color characteristics of the underwater connector, and then uses the underwater connector as the template for preliminary positioning in the detection range, then detects the ellipse characteristics of the end face of the underwater connector, and carries out binocular matching to obtain the relevant three-dimensional point coordinates, and finally calculates the position and orientation of the underwater connector.Experiments show that this method has an average position measurement error of 1.3% and an attitude measurement average error of 3.5°, which can provide a good reference for underwater connector docking operations.
2021,43(4): 64-67 收稿日期：2020-04-17
DOI：10.3404/j.issn.1672-7649.2021.04.013
分类号：TP242.3
基金项目：国家重点研发计划资助项目(2017YFC0307002)
作者简介：陈瑞(1994-),男,硕士研究生,研究方向为水下机器人与机器视觉
参考文献：
[1] 季旭全, 王君臣, 赵江地, 等. 基于机器人与视觉引导的星载设备智能装配方法[J]. 机械工程学报, 2018, 54(23): 63-72
[2] 殷莉甜. 水下双目视觉定位系统开发与应用研究[D]. 广州: 华南理工大学, 2016.
[3] 王从政, 胡松, 冯常, 等. 基于水下双目视觉的燃料组件变形检测系统[J]. 应用光学, 2019, 40(2): 246-252
[4] 盛明伟, 周浩, 黄海, 等. 一种水下双目视觉测距方法研究[J]. 自然科学版, 2018, 46(8): 93-98
[5] ZHANG Z. A flexible new technique for camera calibration[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2000, 22(11): 1330-1334
[6] FUSIELLO A, TRUCCO E, VERRI A. A compact algorithm for rectification of stereo pairs[J]. Machine Vision & Applications, 2000, 12(1): 16-22
[7] FORNACIARI M, PRATI A, CUCCHIARA R. A fast and effective ellipse detector for embedded vision applications[J]. Pattern Recognition, 2014, 47(11): 3693-3708
