针对低对比度和环境复杂情况下,车牌定位时间长和准确率低的问题,提出了一种快速车牌定位的新方法.该方法首先根据边缘密度以及亮度方差进行局部图像增强,然后采用一种高效的算法去除冗余边缘,最后使用水平投影及自适应搜索窗来精确定位并分割车牌.实验验证了该方法完全满足复杂环境下实时车牌定位的要求.
To solve the problems of current methods in complex scenes,this paper presented a novel method for fast license plate location.Based on the edge density and luminance variance,the vehicle image was enhanced by local image enhancement firstly,then an efficient method was introduced for removing redundant edges to improve the edge image,and finally the plate was located by horizontal projection and sliding a adaptive window in the residual edge image.Experimental results demonstrated the new method could meet the requirements of real-time license plate location in complex scenes.