基于Gabor滤波器组的车牌汉字特征提取

CHINESE CHARACTER FEATURE EXTRACTION IN THE VEHICLE-LICENSE-PLATE BASED ON GABOR FILTERS

提出基于Gabor滤波器组的特征提取新方法。利用汉字图像的统计信息及归一化信息,提出针对车牌汉字的有效的Ga-bor滤波器组参数优化方法,并设计一组Gabor滤波器用于提取车牌汉字图像中不同中心频率和方位的笔画纹理特征,实现直接对灰度图像的特征提取。实验结果表明,相比传统二值化特征提取方法,采用基于Gabor滤波器组的小波变换提取特征能够获得更良好的识别性能。

In this paper, a method for Chinese character feature extraction in vehicle-license-plate by using Gabor wavelet transform is presented. Based on the theory of Gabor filters and the statistical information of Chinese character images,a set of Gabor filters was designed to extract texture features of different central frequency and orientation in the character image. Experiments show the method that directly extracts features from gray scale image performs excellently for images in low quality greyscale.