基于可重构标定板的激光与视觉联合标定方法

目前,激光雷达与视觉传感器的联合标定方法包含动态在线标定与静态离线标定两大类.动态在线标定对标定的环境有较高的要求,且标定结果不稳定;静态离线标定通常采用标定板,因而对标定板的要求较高,手动选择与线拟合都易引入或放大误差,故提出了一种基于可重构标定板的激光雷达与视觉传感器的联合标定方法.首先,对传统的标定板进行拆分与重...

LiDAR-Camera Joint Calibration Method Based on Restructurable Calibration Board

The joint calibration methods of lidar and vision sensor includes two categories: dynamic online calibration and static offline calibration. Dynamic online calibration has high requirements for the calibration environment, and its results are unstable. As for static offline calibration, calibration board is a commonly used tool, in which feature points found by the ways of line fitting and manual selection are easy to cause errors and rely highly on the calibration boards. A joint calibration method of lidar and vision sensor based on a new kind of restructurable calibration board is proposed in this paper. Firstly, we split and reconstruct the traditional calibration board. Then we use barcode-like methods to recognize lidar feature points automatically. We also added the camera verification mechanism to alleviate the error caused by the instability of the camera recognition during the calibration process. Finally, in the experiment we calibrate the radar and binocular camera, and use reprojection method to verify the effectiveness of the proposed method through subjective visualization and objective quantitative indicators. The RMSE is 1.275cm in the method, and the error after reprojection is 3.2 pixels.