基于matlab的车牌分割及数字识别系统_学位论文.doc

PAGE  学号：南湖学院毕业设计（论文）题 目： 基于MATLAB的车牌分割及数字识别算法设计作 者届 别系 别机械与电子工程系专 业电子信息工程指导老师职 称完成时间2013.05PAGE II PAGE I PAGE II PAGE I 摘 要 车牌识别技术是智能道路交通管理的重要内容，其识别的准确性和可靠性直接影响到交通管理系统的性能。车牌识别技术包括车牌图像获取、车牌定位、车牌分割、车牌校正、车牌字符分割、车牌字符归一化和车牌字符识别，本文重点针对车牌图像的分割和车牌数字的识别进行了算法研究和设计，同时也对其它步骤进行了探讨。论文首先对获取的车牌图像进行了预处理，包括车牌图像增强、车牌区域提取、车牌几何校正以及车牌字符分割和归一化，然后设计了BP网络算法，最后在MATLAB平台上设计实现了以上各种算法。实验结果表明，本文中的分割算法能准确的获取字符区域并实现对字符的分割和归一化，经过训练后的BP网络能稳定、可靠的实现对分割后字符的识别，实验结果达到预期要求。 关键词：MATLAB；车牌分割；图像预处理；数字识别 PAGE II PAGE I PAGE II PAGE II Abstract License plate recognition technology is an important content of intelligent traffic management, accuracy and reliability identification directly affects the performance of traffic management system. License plate recognition technology include license plate image acquisition, license plate location, license plate segmentation, license plate correction, license plate character segmentation, normalization of license plate character segmentation and license plate character recognition. This paper mainly design and research on the algorithm for license plate image segmentation and license plate number recognition, but also for other steps. Firstly, the thesis get the license plate image preprocessing, including license plate image enhancement, plate region extraction, plate geometric correction and license plate character segmentation and normalization, and then designs the BP network algorithm, finally use the MATLAB platform to design algorithms to implement above. Experimental results show that the segmentation algorithm in this thesis can obtain character area and realize the character segmentation and normalization accurately, after training the BP network can implement the recognition for character segmentation stably and reliably, the experimental reaches expectation. Keywords: MATLAB; License plate segmentation; Image preprocessing; Digital identification PAGE