基于纹理分析和垂直投影的车牌定位算法

根据车牌区域字符的纹理特征和统计规律，提出了一种综合纹理分析和垂直投影的车牌定位方法。由于光照、拍摄角度会对图像产生对比度偏低、图像倾斜和变形等不良影响，利用车牌字符纹理丰富的特征寻找车牌区域可以避开了这些不良影响，并利用灰度垂直投影的统计特征进一步删除由于复杂背景而产生的伪车牌区域。实验结果表明，该方法对于背景复杂的图像，可以得到很好的定位效果，解决了复杂背景下车牌图像提取的难点问题。

Vehicle License Plate Segmentation Based on Texture Analysis and Vertical Projection

Based upon the combination of texture analysis and gray scale vertical projection, a license plate segmentation algorithms is proposed. It uses the texture characteristic and the statistics law of the characters in the license plate area. The algorithm has low sensitivity to plate rotation, distortion and strong daylight. The technique can accurately wipe off fake plate areas brought by the complicated background. The experiment results show the efficient of the method.