基于图像处理的车牌定位与检测技术本科精品毕业论文.docVIP

本 科 毕 业 论 文 基于图像处理的车牌定位与检测技术 License plate location and detection technology based on image processing 基于图像处理的车牌定位与检测技术 摘要 在当今社会，车牌号是车辆身份认证的主要指标。因此，车牌识别技术在城市智能交通管理中的作用也日益重要，具有巨大的商业价值和社会效益。 本次毕业设计首先对与车牌识别相关的现有技术进行深入研究，在研究的基础上开发出一个基于MATLAB的车牌识别系统，通过编写M文件，对民用车辆的图像处理方法进行了解、分析、对比，且实现了汽车牌照图像预处理、车牌定位、字符分割和字符识别。本文中，图像预处理是将图像灰度化、灰度修正和Sobel算子边缘检测。车牌区域的定位与分割是利用数学形态学进行车牌初定位，再根据定位图像中像素点的统计特性来精确定位车牌。然后扫描车牌区域每列白色像素点，结合车牌字符本身的几何特征进行字符分割，最终完成字符分割。考虑到中国车牌所采用的字符数量不是很多，由此本文的字符识别将采用模板匹配的方法，即对输出的字符图像和模板库里的字符模板进行匹配，通过计算处理后得到的图片与模板字符图片规格化的相关系数，越接近于1则相似度越大，越接近于0，相似度越小，然后进行个数累计并找出最大数值，即为识别结果。 MATLAB仿真结果表明,本文所提出的有关车牌的定位、分割和识别算法对普通民用车牌图像具有一定的适应性,识别效果理想,能够满足实用要求。 关键词： 图像灰度化 边缘检测 车牌定位 字符分割 字符识别 License plate location based on image processing and detection technology Abstract In traffic management, Each motor vehicles have a unique license number that was used to identify a vehicle. Therefore, the license plate recognition system has naturally become an important part of intelligent transportation, and has a wide range of application and development in intelligent traffic management in all countries. Firstly, The graduation project gives a deep research on the existing license plate recognition , based on a study exploring, M document is prepared by using MATLAB software . a variety of image processing method for vehicle are understanded , researched ,compared, and proposed the methods based on license plate image preprocessing, license plate location, character segmentation and character recognition . In this paper , image preprocessing include gray image, filtering and canny edge detection . Positioning and segmentation plate area takes use of mathematical morphology method to determine the location of the license plate , according to a prior knowledge of the license plate information , combined with a color region segmentation method to completed license plate segmentation. Character segmentation is the binary license plate area after a