基于MATLAB的车牌识别系统中图像预处理研究毕业论文.doc

毕 业 论 文 基于MATLAB的车牌识别系统中图像预处理研究 PAGE III 基于MATLAB的车牌识别系统中图像预处理研究 专业：电子信息工程 姓名：张黎宾 指导教师：刘文博 摘 要 数字图像处理技术是自1960年以来发展起来的一门新兴学科，随着图像处理技术理论和方法的进一步完善，使得数字图像处理技术在许多领域得到了广泛的应用，并展现出广阔的应用前景。 本文主要是研究车牌识别系统中一个重要的模块——图像预处理。为了使彩色图像减少存储信息，我们可以先对彩色图像进行灰度化，只保留该图像的亮度信息。本文通过传统的方法中值滤波和均值滤波对图像进行去噪，对结果比较分析，虽然中值滤波比均值滤波效果要好，但是图像还是含有较多的噪声。所以本文在小波变换的基础上进行了改进，图像经过小波变换后再进行中值、均值滤波去噪，这样也再次把中值、均值滤波进行对比。通过对效果图的分析和总结，改进后的去噪效果要比之前的更好，同时改进后的中值滤波还是比均值滤波好一些，因此基于小波变换的中值滤波是值得采用的。 本文提出的针对车牌图像的预处理过程对灰度图像可以达到不错的效果，并且在一定的程度上解决了因外界因素而造成的车牌图像对比度较低的问题。 关键词 数字图像处理，预处理，中值滤波 ABSTRACT Digital image processing technology is I emerging subject since 1960,the amelioration of image processing theory and methods gets the Digital image processing technology to be applied widely in many fields and present a broad application prospect in the days to come. This paper is mainly research vehicle plate recognition system a important module——image preprocessing.In order to make the color image to reduce the storage information, we can first on the gray of the color images, just retaining the brightness information of the image. This article through the traditional method of median filter and mean filter for image noise reduction to comparative analysis the result, although the effect of median filtering is better than the mean filter, the images still contain lots of noise . Therefore, this paper makes improvement on the basis of wavelet transform , image after wavelet transform , noise reduction of the median filtering and mean filter. It compared the median filtering and mean filter once more . According to the effect graph analysis and summary, the effect of the improved noise reduction is better than before, at the same time, the improved median filtering is also better than the improved mean filter , so based on wavelet transform is worth the median filter. The idea of vehicle plate recognition system will earn a positive result as far as the pray image is concerned and, to some degree, resolve the problem evoked b