一种基于交通视频的摄像机标定方法

摘要

介绍了一种线性模型下的摄像机标定方法。在获取交通视频的基础上，利用实际路面上的参照物和图像处理的基本知识抽象出交通视频中的像素行和路面实际距离之间的几何模型，对几何模型进行解析，得到视频图像上像素行和实际路面的距离映射关系。试验表明，该方法能够较好地实现图像上像素行到路面实际距离之间的转换。%A method of camera calibration in linear model is introduced. On the basis of traffic video , geometric model between the pixel rows from the traffic video and actual road distance is abstracted by using the real pavement reference object and basic knowledge of image processing. Get mapping relation between video image pixel rows and actual pavement distance after analyzing geometric model. Experiments show that the method can better realize converting the image pixel rows into actual pavement distance.