测绘通报 ›› 2019, Vol. 0 ›› Issue (12): 1-7.doi: 10.13474/j.cnki.11-2246.2019.0375

• 人工智能与高精度地图 •    下一篇

基于ORB-SLAM2算法的水下机器人实时定位研究

张阳1,2,3, 李海森1,2,3, 马礼4, 王健1,2,3, 郭企嘉1,2,3, 祁晶5, 闫铭6   

  1. 1. 哈尔滨工程大学水声工程学院, 黑龙江 哈尔滨 150001;
    2. 哈尔滨工程大学水声技术重点实验室, 黑龙江 哈尔滨 150001;
    3. 工业和信息化部海洋信息获取与安全工信部重点实验室(哈尔滨工程大学), 黑龙江 哈尔滨 150001;
    4. 中国地质大学(北京)信息工程学院, 北京 100083;
    5. 江苏海洋大学测绘与海洋信息学院, 江苏 连云港 222005;
    6. 上海大学力学与工程科学学院, 上海 200444
  • 收稿日期:2019-03-28 修回日期:2019-10-25 发布日期:2020-01-03
  • 通讯作者: 王健。E-mail:danlian674@126.com E-mail:danlian674@126.com
  • 作者简介:张阳(1991-),男,博士生,主要从事水下SLAM、水声技术等研究。E-mail:zhyang2018@gmail.com
  • 基金资助:
    国家重点研发计划(2017YFC0306000)

Research on real-time positioning of underwater robot based on ORB-SLAM2 algorithm

ZHANG Yang1,2,3, LI Haisen1,2,3, MA Li4, WANG Jian1,2,3, GUO Qijia1,2,3, QI Jing5, YAN Ming6   

  1. 1. College of Underwater Acoustic Engineering, Harbin Engineering University, Harbin, 150001, China;
    2. Acoustic Science and Technology Laboratory, Harbin Engineering University, Harbin 150001, China;
    3. Key Laboratory of Marine Information Acquisition and Security(Harbin Engineering University), Ministry of Industry and Information Technology, Harbin 150001, China;
    4. School of Information Engineering, China University of Geosciences, Beijing 100083, China;
    5. School of Geomatics and Marine Information, Jiangsu Ocean University, Lianyungang 222005, China;
    6. School of Mechanics and Engineering Science, Shanghai University, Shanghai 200444, China
  • Received:2019-03-28 Revised:2019-10-25 Published:2020-01-03

摘要: 随着近几年视觉SLAM的快速发展,为机器人、无人机、汽车等导航定位提供了更多选择。针对当前水下机器人定位技术存在的系统复杂、操作难度大等问题,本文提出了一种基于ORB-SLAM2算法的水下定位方案。利用单目相机作为传感器,构建了单目视觉SLAM水下机器人定位模型,完成像素坐标系到世界坐标系的转换,介绍了ORB-SLAM2算法定位涉及的关键技术。通过水下理想环境试验,对ORB-SLAM2算法在水下定位性能进行了综合评价,通过海洋环境试验证明ORB-SLAM2算法可以有效对水下机器人进行实时定位。

关键词: 水下机器人, 水下定位, SLAM, 视觉里程计, ORB-SLAM2

Abstract: With the rapid development of visual SLAM in recent years, it provides more choices for navigation and positioning of robots, drones, and automobiles. Aiming at the problems of complex systems and difficult operation of current underwater robot positioning technology, this paper proposes an underwater positioning scheme based on ORB-SLAM2 algorithm. Using the monocular camera as the sensor, the monocular vision SLAM underwater robot positioning model is constructed, and the conversion from pixel coordinate system to world coordinate system is completed. The key technologies involved in ORB-SLAM2 algorithm positioning are introduced. The underwater positioning performance of the ORB-SLAM2 algorithm is comprehensively evaluated by the underwater ideal environment experiment. The marine environment experiment proves that the ORB-SLAM2 algorithm can effectively locate the underwater robot in real time.

Key words: underwater robot, underwater positioning, SLAM, visual odometer, ORB-SLAM2

中图分类号: