基于两垂直相交线段的摄像机快速标定算法  

Fast camera calibration algorithm based on two perpendicular line segments

作　　者：贺科学[1,2] 李树涛[1] 

机构地区：[1]湖南大学电气与信息工程学院,长沙410082 [2]长沙理工大学电气与信息工程学院,长沙410076

出　　处：《仪器仪表学报》2013年第8期1696-1702,共7页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(61172161);电力与交通安全监控及节能技术教育部工程研究中心(2011kfj3);湖南省教育厅科学研究项目(09C076)资助

摘　　要：针对交通标线特有的几何特征,选取车道分界虚线的3个端点及其垂直线上的另一点作为标定参考点,推导出求解摄像机内外参数的公式,建立了图像像素坐标与其世界坐标的换算关系。标定算法的约束条件仅需交通标线中丁字形标定参考物的3个线段长度值、4个关键点及其在单幅图像中的坐标。算法具有标定参考物易选取、几何约束条件少等优点,非常适用于交通监控系统中单摄像机的快速标定。Aiming at the special geometric features of road lane markings, three collinear points on the broken white line of the road lane and another point on its perpendicular line are taken as the calibration reference objects. The formula for solving intrinsic and extrinsic camera parameters is derived and the conversion relationship between the image pixel coordinates and their world coordinates is established. The constraint conditions of the proposed method are only the lengths of three line segments, four key points and their corresponding coordinates in single image, which are in the calibration reference object of the road lane marking. The proposed method has the advantages, such as easy selection of the reference objects and fewer geometric constraints and etc. , which is suitable to be used for the mono-camera expeditious calibration in the traffic surveillance system.

关 键 词：摄像机标定 交通标线 视频监控 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]