一种基于特征点的车牌识别改进算法

A License Plate Recognition Algorithm Based on Characteristic Point

车牌识别技术已经成为公路交通自动控制与管理(RTACM)以及智能运输系统(ITS)中的一个重要组成部分。提出一种基于特征点的车牌识别改进算法,利用车牌的纹理特征和形状特征定位车牌区域,采用垂直投影分割车牌字符,通过统计特征点进行字符识别。实验结果表明,该算法能显著提高由于拍摄角度引起的车牌图像中字符拉伸、变形等情况下的识别率,同时缩短了识别时间。

License plate recognition technology had become an important part of Road Traffic Automatic Control ＆ Management and Intelligent Transportation System. An improved license plate recognition algorithm based on characteristic point is proposed. In the algorithm, the texture feature and shape feature is used to locate the license plate, and the vertical projection is used to segment character of license plate, and characteristic points are counted to recognize characters. Experimental results show that the algorithm can signifi- cantly improve the recognition rate and time in the case of the stretching deformation of characters in license plate image by the camera angle cause.