车牌定位是车牌识别系统需要解决的首要问题。基于FPGA的硬件特点, 提出了一种简单、有效的车牌定位算法。在车牌区域图像的预处理阶段, 提出了一种新的基于色彩分量的灰度化方法, 该方法不仅可以消除小型民用车车牌图片的背景和车身信息, 还可以较好的保留车牌区域信息, 从而降低定位难度, 简化定位步骤, 提高小型民用车车牌定位的效率。并设计了一种实时车牌定位模块的软核IP, 该IP Core通过了代码覆盖率和功能覆盖率的测试。实际工程项目测试表明, 该IP Core定位准确、反应速度快。

The license plate location is the most important issue in the license plate recognition system. According to the characteristics of the FPGA, a simple and effective license plate location algorithm is proposed. In the preprocessing stage of the area of license, a new method of graying based on the color component is proposed. The new method can not only eliminate the background and the body information of the small civilian vehicles’ license plate picture, but also preserve the information of license plate preferably. The difficulty of location is reduced, the location’s steps is simplified, and the efficiency of the small civilian vehicles’ license plate location is improved. According to the new algorithm proposed, the soft IP Core of real-time license plate location is designed. It is tested by the code coverage and functional coverage. It has accurate locating and fast response by the actual project test.