1、1 毕业设计（论文） 课题 名称 多路径识别算法的研究 院别：院别： 电气工程学院 专业：专业： 电子信息工程 2 摘要 随着各省市高速公路的不断建设， 基本上建成了大小规模不等的联网收费系统， 在联网收费系统建设中，都面临了一个共同的问题车辆行驶路径的识别。路径 识别不仅仅涉及对每一通行车辆如何计算通行费，同时还要考虑将收入的通行费拆 分给哪个收费单位的问题，通行费的拆分直接关系到各条高速公路的合法利益，是 联网收费的关键。所以说解决多路径识别问题在完善高速公路联网收费系统中起着 至关重要的作用。 本文介绍了多路径识别的多种算法， 并着重分析了识别算法中的汽车牌照识别， 根据车牌图像特点，对
2、车牌识别算法关键环节包括数字图像预处理，车牌定位，车 牌字符识别等三部分一一进行了分析研究。 本文采用直方图变换、边缘检测、二值化等方法对车牌图像进行了预处理。对 车牌定位采用了一种基于灰度图像求取卷积能量极值区域的车牌定位方法。采用模 板匹配法对车牌字符进行识别。 并采用了 MATLAB 来完成算法的设计与实现。 研究结 果表明车牌识别率较高，适应性较强，实时性比较好。 关键词：多路径识别；车牌识别；车牌定位；模板匹配；字符识别 3 ABSTRACT As expressway construction unceasingly in provinces, numbers of online
3、charging systems was completed bascially. In the construction of online charging system, we are facing a common problem The identification of vehicle path. Path recognition involves not only how to calculate the tolls to every vehicle, also should consider how to split the incomes to relevant chargi
4、ng unit. Toll split is directly related to the legitimate benifit of every expressway,and is the key to networking toll. So solving the problem of multipath recognition plays an important role in perfecting the expressway networking toll system. This article introduces the multipath recognition algo
5、rithms, And emphatically analysed the recognition algorithm of license plate recognition. Based on image characters of license plate, we have a deeply research on the key technologies of the vehicle license plate recognition system, which include three parts: digital image pre-processing, license pl
6、ate location, license plate characters recognition. In the vehicle images pre-procession, the paper use the methods of Histogram Transformation, Edge detection, binarization. For license plate localization using a kind of calculating convolution based on gray image energy extremum region of license plate locating method.The paper use pattern matching which method has been improved to recognize the characters. And using MATLAB to complete the design and implementation of
