摘要
根据中国车牌的特点,提出一种主要是基于颜色相似度和纹理特征相结合的车牌定位方法,这种方法把采集到的车牌图像转化到HSV颜色空间,通过计算当前颜色与中心颜色之间的距离相似度,结合中值滤波进行车牌的粗定位,再结合车牌纹理特征进行车牌的精定位．
It is presented an effective license plate location algorithm mainly based on color's degree of similarity and texture characteristics. By using this method, we put the images collected to HSV color space, compute the distance and similarity between current color and center color and locate roughly License Plate by using median filter. Then we use texture characteristics to locate license plate accurately.
出处
《云南大学学报（自然科学版）》
CAS
CSCD
北大核心
2006年第S2期94-96,101,共4页
Journal of Yunnan University(Natural Sciences Edition)
关键词
车牌定位
HSV模型
相似度
纹理特征
license plate location
HSV Model
similarity
texture characteristic
