【摘要】 照相机内外参数的标定是利用计算机视觉理论进行交通事故现场测量的基础。线性成像模型是照相机最简单、最基本的成像模型。根据交通现场的静态特点,可以对事故现场摄影时照相机的成像参数进行现场标定。为了提高测量速度,对标定参照物上的标定点进行交互式自动识别。本文提出的标定点设置方式和提取方法,可以克服阴影干扰。用基于误差统计的方法对标定点进行优选,可以确保对照相机线性成像模型的可靠标定。更多还原

【Abstract】 Camera calibration is the base of the measuring traffic accident scene using computer vision theory. The most important and simplest camera model is linear one. Because the objects measured in accident scene are generally static, camera can be calibrated at traffic scene using certain poults on reference object. In order to calibrate camera quickly, reference points must be recognized automatically. Using the method given in this paper, reference points could be recognized rightly and automatically, and the effect of shadow could be overcome. Reference points were selected on the base of the error of reference points. Camera could be robust calibrated using those points.更多还原