ZHAO Minghui, WANG Jianhua, ZHENG Xiang, et al. Monocular vision based water-surface target distance measurement method for unmanned surface vehicles[J]. 2021, (2): 47-50+54. DOI: 10.13873/J.1000-9787(2021)02-0047-04.
基于单目视觉的无人水面艇水面目标测距方法
摘要
针对单目测距时相机高度和俯仰角难以测量的问题
提出一种相机高度和俯仰角的标定方法
在此基础上提出一种水面目标的单目测距方法
并分析影响测距精度的因素。通过相机标定
获得与水面共面的标定板上的网格坐标系到相机坐标系的外参
计算相机高度和俯仰角;识别出图像中水面目标所在区域
计算该区域像素纵坐标最大值
及其对应的横坐标的平均值
将组合得到的坐标作为观测点的像素坐标;根据相机内参、透视投影关系和刚体变换
获得观测点在无人艇附体坐标系下的三维信息
进而计算观测点的距离。通过实验验证了所提方法的正确性和有效性。
Abstract
Aiming at the problem that the height and pitch angle are difficult to measure during monocular ranging
a method for calibrating camera height and pitch angle is proposed
and a water-surface target ranging method based on monocular vision is proposed
and the factors affecting the ranging precision are analyzed.Through camera calibration
obtain the external coordinate of the grid coordinate system on the calibration plate that is coplanar with the water surface to the camera coordinate system
calculate the camera height and the pitch angle. The area of the target on water surface in the image is identified
the maximum value of the ordinate of the pixel in the area
and the average value of the corresponding horizontal coordinate are calculated
and the combined coordinate is taken as the pixel coordinate of the observation point. According to the camera internal reference
perspective projection relationship and rigid body transformation
the three-dimensional information of the observation point in the coordinate system of the unmanned boat is obtained
and then the distance of the observation point is calculated. The correctness and effectiveness of the proposed method are verified by experiments.