汽车轮胎号识别中的预处理问题

为便于对汽车轮胎号字符图像进行识别，对其进行一系列预处理．着重研究胎号图像预处理中的增强、降噪及二值化问题．提出一种将邻域均值滤波法和中值滤波法相结合的图像降噪算法，采用该算法能保护图像边缘和细节．提出一种基于经验知识的改进Otsu算法，可缩小Otsu算法搜索阈值范围，减少计算类间方差的次数．结果表明，胎号增强图像比原始图像清晰，降噪后字符图像边缘不模糊，胎号字符二值化速度得到提高，二值化效果也很好．

Pretreatment method in autocar tire number recognition

In order to easily recognize autocar tire number image, several pretreatments are studied including enhancing image, filtering noise and binarizing image. A new noise reduction arithmetic is put forward which combines the adjacent average arithmetic with the middle value fiheration arithmetic. This arithmetic can protect the character image brim. Because the scope of Otsu arithmetic searching threshold is the whole gray field, an improved arithmetic is brought forward which can dwindle the range of choosing threshold and decrease the times of calculating variance between two classes. Experiment shows that enhanced image is clearer than original image, and the brim of noise-reduced image is not faint. The image binarization speed is improved and the effect is satisfied.