Abstract: Location and binarization are vital in license plate recognition. Only single step is taken in traditional way so as to meet the demand of recognition time. But because of the flaw of arithmetic, it is not very adaptive to some particular images, which leads to the failure of recognition. In order to solve this problem and satisfy the need of recognition rate meanwhile, this paper gives a new way composed of two steps based on modification, which can improve the effect of binarization and location to offer an accurate input for the next process.

Key words: License plate recognition (LPR), Location, Binarization

摘要： 车牌图像定位和二值化是车牌识别系统中一个关键的部分。传统的处理方法是使用单一的算法，以便能够满足实时性的要求。但是算法本身的缺陷性使其不能适应一些具有独特特点的车牌，不能正确地识别出结果。基于提高算法的适应范围考虑，在满足识别速率要求的前提下，该文提出了基于修正思想的“去粗取精”处理模式，可以大大改善定位和二值化处理的效果，为后续的处理提供了一个良好效果的输入。

关键词: 车牌识别, 定位, 二值化

CLC Number: 

• TP391.4