一种多目相机的外参数两步联合在线标定方法

A Two-Step Online Calibration Method of Multi-Camera Extrinsic Parameters

相机外参数的标定是计算机视觉和摄影测量领域的研究热点。但是由于实际标定成本和相机安装环境的限制,难以频繁对其进行外参数标定。针对静态多目相机对外参数连续、在线、稳定的标定需求,提出一种多目相机外参数两步联合在线标定方法。该算法通过对共线交会的外参数初值进行稀疏光束法平差,使外参数存在系统噪声的情况下仍获取稳定估计值。外参数标定的实测试验和仿真分析结果表明,该方法对外参数初值系统误差具有鲁棒性强和收敛速度快的优点,解决了多目相机外参数的在线快速标定问题;满足对静态、连续、在线外参数标定有较高需求的实际工程应用,对视觉导航有参考和借鉴意义。

The calibration of extrinsic parameters is a hot research focus in the field of computer vision and photogrammetry.However,limited to the actual calibration costs and the installation of the camera environment,it is difficult to calibrate external parameters frequently.In consideration of the requirements of continuous,online and stable external parameters calibration for static multi-camera,a two-step combined online calibration of extrinsic parameters of multi-cameras is proposed,which can obtain the stable estimated external parameter in sytem noise environment through sparse bundle adjustment on initial external parameter.The experiment and simulation analysis of the external parameters calibration are carried out,whose results indicate that this method has the advantages of strong robustness and fast convergence speed in the system error of extrinsic parameters initial calibration.This method solves the problem of on-line fast calibration of multi camera parameters,and meets the needs of the actual engineering application of the static,continuous,online external parameters calibration.This method has certain reference value of the field of vision navigation.