基于HALCON的车牌识别研究

Study on license plate recognition based on HALCON

为了实现车牌字符快速准确的识别,基于Halcon软件提出了一种二次阈值的车牌定位方法,根据车牌在HSV颜色模式中S通道的纹理特征进行一次阈值,获取车牌候选区域,其次校正车牌候选区域,最后进行第二次阈值,通过字符区域的最大(小)行(列)值确定出精确的车牌区域;在字符分割模块提出了一种汉字分割与其他字符连通域法相结合的方法。实验结果表明,这两种方法简单快捷、识别正确率高。

In order to achieve accurate and quick recognition of license plates,a license plate positioning method with secondary threshold is proposed based on Halcon. A primary threshold is made based on the license plate texture characteristics of S channel in HSV color mode to get the candidate region of a license plate,and then adjusting the candidate region to make the second threshold and determine the exact license plate region according to the maximum（minimum）row（column）value of the character region. As for the character segmentation module,a method to combine Chinese character segmentation with other character connected domain method is proposed. The experimental results indicate that both of the two methods are simple and quick,and have high recognition rate.