基于视频的车型分类算法研究

对视频图像处理中车型分类算法涉及的摄像机标定和目标分割算法进行研究。通过摄像机标定实现像素坐标到世界坐标的映射,进而准确获取车辆长度及宽度信息。建立车辆高度信息计算模型,获得车辆高度信息。对实时视频图像进行目标分割,并通过车辆长度、高度及长高比等信息实现高速公路中车型分类。经过验证,该方法能够比较准确地进行车型分类,满足实际应用要求。

Study on vehicle type classification algorithm based on video

The camera calibration and target segmentation algorithm involved with the vehicle type classification algorithm in video image processing are studied in this paper. The length and width information of vehicle accurately is achieved by map-ping the pixel coordinates to world coordinates through camera calibration. The vehicle height information model is established to obtain the vehicle height information. Classification of the vehicles on expressway can be realized by vehicle information of length,height and length to height ratio,as well as the target segmentation of real-time video images. The field test results show that the algorithm can classify the vehicles accurately and meet the requirements in practical applications.