基于OpenCV的水下机器人单目定位技术研究与仿真

Research and Simulation of Underwater Vehicle Monocular Positioning Technology Based on OpenCV

摘要:

针对水下机器人自主回收对接时的定位问题,提出了一种基于OpenCV的单目定位技术研究与仿真方法;首先确定回收装置在回收侧的光源标记点坐标信息；然后借助OpenCV算法库,通过对水下机器人自带摄像机的标定,得出反应摄像机固有信息的内参数,通过摄像机对回收装置光源标记点的识别,得出光源标记点在图像上的像素坐标,结合其世界坐标,得出反映回收装置在摄像机坐标系下位置和姿态信息关系的平移向量和旋转向量,进而确定水下机器人在回收装置坐标系下的位置和姿态信息；最后运用CATIA软件对摄像机拍摄模型进行建模和仿真,结果表明,所提方法不仅能快速地获得水下机器人的位置和姿态信息,而且定位精度高,满足水下机器人在自主回收对接时的设计要求。

Abstract:

Focused on obtaining the position and attitude information for automation recovery of underwater vehicle, a research and simulation method of monocular positioning technology based on OpenCV was proposed. Firstly, the light source tag coordinate information for the recovery side of the recovery unit was determined;Then, with the help of OpenCV library, the intrinsic parameters of the camera's inherent information were obtained by calibrating the camera of the underwater vehicle. Through the recognition of the recovery markers, the translation and rotation vector which reacted position and attitude relationship of the reclaimer in the camera coordinate system were obtained. Then the position and attitude of the underwater vehicle in the world coordinate system of the reclaimer were determined. Finally, the camera shooting model was modeled and simulated by CATIA, the experimental results demonstrate that the method can obtain the position and attitude information of underwater vehicle quickly and accurately, and the positioning accuracy is high,which meets the design requirements of autonomous unterwater vehicle for recovery and docking.