基于DSP的车牌识别系统的应用研究.doc

基于DSP的车牌识别系统的应用研究 摘要 伴随着世界各国汽车数量急剧增加，城市交通状况日益引起人们的重视，如何有效地进行交通管理，已成为越来越多人关注的焦点，解决这些问题的关键就是建立智能交通系统。车牌识别是智能交通系统的重要组成部分，它在交通控制与监视中有着多种用途，目前已广泛应用于各种领域。 本文将TMS320C2812作为核心器件用于车牌自动识别系统中，完成车牌图像的采集、数字图像的处理、提取车牌信息并针对提取的特征对字符进行识别。首先分析了车牌识别系统实现的背景以及其实现意义。然后对实现车牌识别的硬件环境作简要介绍。接着对车牌识别过程中所涉及的边缘检测、字符分割、大小归一化等一系列数字图象处理技术进行进一步的详细分析。之后，对几种字符特征提取算法进行了对比分析，最后选取最适合的网格特征提取法，以此为基础进行模扳匹配，最终识别出车牌号码。 关键词：车牌提取；图象处理；车牌识别；DSP； Abstract With the increment of vehicle all over the world, the situation of city traffic has attracted the attention of people. How to control the traffic effectively has become the problem which more and more people pay close attention to. The way to solve this problem is to establish the ITS—Intelligent Transportation System .Vehicle license plate recognition system is the crucial part of the ITS. It is widely used in vehicle monitoring and traffic control. This experiment attempts to use the new generation DSP—Digital Signal Processor to implement the recognition task. The DSP chip TMS320C2812 is used to process the picture of the vehicle, distill information of the license plate and recognize the vehicle license plate. The first part of the thesis is about the background and meaning of the vehicle license plate recognition system. Consequently we analyze the environment of the experiment. In the next part we analysis the fundamental theory and technique of the image processing, including the collection of picture of vehicle, distill of the license plate, segmentation algorithm of character. Then we put forward several methods to distill the feature of the characters. On the basis of the above-mentioned research, we make the template matching and recognize the characters. Keywords: Vehicle License Plate Location; Image Processing; Character Recognition; DSP; 1、引言 　改革开放以来，我国的交通运输业迅速发展。但伴随着其发展，也出现了一系列问题，如交通堵塞、交通事故和环境污染等等。虽然可以靠建设更多的道路设施来满足交通运输增长的需求，但在资源、环境矛盾越来越突出的今天，道路设施的增长将受到限制，这就需要依靠提供除了设施以外的技术方法来满足这一要求。交通的迅速发展使得全世界的研究者不断采用先进的电子和计算机视觉技术来检视超速车辆、掌握车辆行驶，或者用于收费站，以提高车辆的通