水下机器人的单目视觉定位系统

Mono-vision localization system of underwater robot

介绍了OutLand 1000型水下小型远程遥控机器人(ROV)所携带的视觉系统,并通过实验对水下摄像机进行了标定,获得了像机的内参数。不同于陆地机器人的简化摄像机模型,针对水下环境对像机模型的影响,建立了水下像机的非线性畸变模型,在此基础上,对陆地单目视觉定位算法做了改进,并完成了系列水下实验。实验结果表明:经改进后的视觉定位方法提高了水下定位精度。

Vision system of OutLand 1000 ROV is introduced. Underwater camera is calibrated by experiment to obtain intrinsic parameters. Different from the simplified camera model adopted by robot on land, the nonlinear distortion model of camera are considered for the underwater environment. An improved algorithm is proposed based on the mono-vision localization algorithm on land, and a series of underwater experiments are performed. The experiment results show that the precision of underwater localization is improved.