| 面向ARV的视觉辅助水下对接方法研究 |
| |
| 引用本文： | 魏志祥,吴超.面向ARV的视觉辅助水下对接方法研究[J].海洋工程,2021,39(1):91-99,152. |
| |
| 作者姓名： | 魏志祥 吴超 |
| |
| 作者单位： | 上海交通大学 船舶建筑与海洋工程学院,上海 200240;上海交通大学 船舶建筑与海洋工程学院,上海 200240 |
| |
| 基金项目： | 国家重点研发计划项目（2016YFC0300700）；大洋两型新船装备研发项目（DYA2017HT036）；大洋重大装备研制项目（DY125-21-Js-06） |
| |
| 摘 要： | 针对水下机器人作业过程中的近距离引导水下对接问题,以自治缆控水下机器人(ARV)为研究对象,为实现快速、高精度的水下对接,设计了基于反射光源识别的单目视觉辅助水下对接方法。通过在对接口布置反光带并将其作为目标图像,设计了图像处理和特征点提取的算法,经过图像特征信息的分析处理,优化算法的时间复杂度,提高了特征点提取的准确率和识别效率。最后设计了传统的引导灯方案与反光带方案的对比试验,验证算法的可行性,证明在水下环境下识别反射光源的方法极大提升了位置估计的准确率,同时在计算速度上也有显著提高,弥补了传统水下对接方法中精度不足或计算量大的缺陷,更好地满足了ARV水下对接的需求,能为潜水器实现水下自主对接提供参考依据。
|
| 关 键 词： | 自治缆控水下机器人 水下对接 单目视觉 图像处理 特征点提取 |
| 收稿时间： | 2020/1/23 0:00:00 |
Study on docking method of ARV based on monocular vision |
| |
| WEI Zhixiang,WU Chao.Study on docking method of ARV based on monocular vision[J].Ocean Engineering,2021,39(1):91-99,152. |
| |
| Authors: | WEI Zhixiang WU Chao |
| |
| Institution: | School of Naval Architecture, Ocean and Civil Engineering, Shanghai Jiao Tong University, Shanghai 200240, China |
| |
| Abstract: | Aiming at the problem of close range underwater docking during underwater robot operation, we, studying ARV, designed an underwater docking method based on identifing reflection light source with monocular vision to achieve the quick and highly accurate underwater docking. We also designed a method to arrange reflective tape at the entrance of the ARV and set it as the target image. An algorithm of image processing and feature point extraction was designed. The algorithm was optimized by analyzing the features of the images to enhance the accuracy and recognition efficiency of the feature point extraction. In order to verify the feasibility of the method, an experiment on the guiding light solution in contrast with the reflective tape solution was designed to verify the accuracy of location estimation under the water. In the meanwhile, the method enhanced the velocity of calculation to meet the requirements of ARV underwater docking. The study results can provide references for autonomous underwater docking. |
| |
| Keywords: | autonomous and remotely operated vehicle underwater docking monocular vision image processing feature point extraction |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《海洋工程》浏览原始摘要信息 |
| 点击此处可从《海洋工程》下载免费的PDF全文 |
|
