一种基于连通域和跳变的车牌定位方法

在车牌识别系统的实现过程中，车牌定位是最基础的一步。本文介绍一种基于连通域和灰度跳变的车牌定位方法。该算法首先对灰度化后的车牌图像进行“阴阳牌”的判定，对判定后的“阴阳牌”和普通牌分别进行处理；根据灰度跳变特征对车牌重新进行更细致的行定位以及列定位，得到完整车牌。实验表明这种车牌定位算法具有定位准确、光照干扰小的特点。

A Method of License Plate Location Based on Connected Components and Gray Level Jump

In the License Plate Recognition (LPR) system, license plate location is the basic step. A method of license plate location is introduced based on connected components and gray level jump of plate. Firstly, it is detected that whether the license plate has shadow region. Then special binaryzation method is used for shadow license plate. Finally, license plate location is completed using gray level jump. Experiments show that this algorithm is precise for license plate location and robust on light disturbance.