用于车辆监控系统的单目视觉二维标定新算法  

New Two-dimensional Monocular Vision Calibration Algorithm for Vehicle Monitoring Systems

作　　者：王庆峰[1] 张全法[1] 李焕[1] 任朝栋[1] 

机构地区：[1]郑州大学物理工程学院,郑州450001

出　　处：《中原工学院学报》2011年第3期50-53,共4页Journal of Zhongyuan University of Technology

摘　　要：为了在基于单目视觉的车辆监控系统中实现快速现场二维标定,提出了一种新的标定算法.该算法通过变量分离简化了路面坐标与图像坐标之间的关系,使得能够利用一元多项式逼近这种关系,并通过求解线性方程组快速地确定它;利用交通标线作为标定参照物,以便获得足够大的标定范围并方便标定操作.实验结果表明,该算法具有可行性.To meet the fast on-site two-dimensional calibration requirement of vehicle monitoring systems based on monocular vision,a new calibration algorithm is proposed.This algorithm simplifies the relations between image and road coordinates through variable separation,thus makes possible of using unary multinomials as their appoximating functions and solving linear equations to determine them quickly.It also makes use of traffic markings as reference templates to make operations easy and get calibrated ranges as large as requirements.Experiment shows that it is feasible for practice.

关 键 词：车辆监控系统 视觉标定 交通标线 坐标变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]