基于VC++的车牌识别系统与开发教材.docx

毕业设计（论文）中文摘要 基于VC++的车牌识别系统研究与开发 摘 要：车辆牌照识别（LPR）系统是一个专用的计算机视觉系统，它能够自动地摄取车辆图像和识别车牌号码，可应用在公路自动收费、停车场管理、失窃车辆侦察、门卫系统、智能交通系统等不同场合。LPR系统的广泛应用将有助于加快我国交通管理自动化的进程。 本文对车牌识别系统中的图像预处理、车牌定位、字符分割和字符识别等环节涉及到的技术、算法以及系统整体设计作了全面的论述，并与已有方案进行了比较，对部分关键算法进行了设计和改进。 图像预处理：采用灰度变换、对比度增强、图像平滑、边缘检测、二值化等方法对图像进行了处理，提高了图像的质量，强化了图像区域。 车牌定位：这是车牌识别系统中至关重要的一步，本文采取了基于边缘点图，根据车牌区域的竖向纹理特征对车牌进行定位，具有较高的车牌定位准确率。 字符分割：首先对得到的车牌进行倾斜度调整，然后利用连通域法分割得到单独的字符，并对字符进行了归一化处理，便于下一步字符识别。 字符识别：选取了13个特征点，采用模板匹配的方法，对汉字、字母、数字字符进行快速识别，取得了较高的识别率和可靠性。 用采集到的40多幅汽车图片作为实验样本数据，用纯软件的方法实现了车牌字符的自动识别，达到了较高的识别率。 关键词：图像预处理 车牌定位 竖向纹理 字符分割 字符识别  毕业设计（论文）外文摘要 License Plate Recognition System Research and Development Based On VC++ Abstract: As a special computer vision system, the License Plate Recognition System (LPR) can capture a vehicle automatically and identify the plate numbers in the image. It can be used in the road tolling system, access control system, stolen vehicle identification system and intelligent transportation system and so on, contributing a lot to the automation of the transportation management. In this paper, some methods and algorithms are introduced in such stages as preprocessing of images, license plate location, character segmentation and character recognition. At the same time, I compare the methods with the existing. Moreover, I design and improve parts of critical algorithm. Image preprocessing: The image is processed by several algorithms including gray level stretching, contrast stretching, image smoothing, edge detection, image binarization. Through the processing, the noise on the image is eliminated effectively and vertical texture is clearer. License plate location: This aspect is an extremely important step in this system and is also emphased in this paper. One kind of location method of license plate image, basing on edge point image and vertical texture feature is put forward in this paper, which can get a satisfactory recognition effect. Character segmentation: First, inclination rectification