基于HSV空间和连通域的车牌定位研究

Research on License Plate Location Method Based on HSV Space and Connected Components

随着智能交通技术的快速发展,车牌识别系统(License Plate Recognition,LPR)已经成为其中的一个重要环节。在分析现有车牌定位算法基础上,提出一种HSV颜色空间和连通域相结合的定位方法。利用车牌颜色信息对图片提取出可能存在的车牌区域,通过数学形态学的分析得到候选连通区域,再结合车牌先验知识长宽比特征进行区域搜索判断,根据车牌字符特征垂直方向对车牌二值图像精确定位得出正确的车牌区域。实验结果表明,所提出的方法可以有效地对车牌进行定位。

With the fast development of Intelligent Transportation System（ITS） ,License Plate Recognition（LPR） is a key point of it.On the basis of current vehicle license plate location algorithm,a method for license plate location is proposed based on HSV space and connected components.The color information of license plate is used to get the possible plate region, and the connected region is obtained by mathematical morphology techniques.Based on the prior knowledge of the license plate aspect ratio, the candidate regions of license plate is searched and judged.Finally, the precise positioning for binary image of license plate is performed based on its character feature vertical direction to obtain the right license plate area.The experimental results show that the proposed algorithm can effectively locate the license plate.