基于.颜色特征的图像检索.doc

WORD文档下载可编辑 PAGE 专业技术资料分享 WORD文档下载可编辑 专业技术资料分享 摘要 随着多媒体及网络技术的发展，互联网中的图像等多媒体信息的数量在迅速增加。如何在庞大的多媒体数据资源海洋中搜索到自己所需要的图像，随之成为一个函待解决的问题。近年来基于内容的图像检索技术己经成为了研究的热点，该技术具有传统文本检索方式所无法比拟的优越性，目前已被应用于多媒体通信、生物识别、知识产权保护等方面，具有广阔的发展前景。 本课题设计了基于颜色特征的图像检索系统，该系统主要包括两个关键步骤：颜色特征的提取和相似度匹配。本文中首先采用应用广泛的HSV颜色空间来表示图像的颜色特征，对颜色分量进行非等间隔量化形成特征矢量；然后用欧式距离度量函数进行图像的相似性匹配；最后在MATLAB平台下编程实现了基于颜色特征的检索系统。根据用户输入的样本图像与图像库中的图像进行特征匹配，找出与样本图像距离比较小的若干幅图像，并按照图像之间的距离由小到大的顺序显示给用户。 关键词： 图像检索；直方图；颜色特征；MATLAB ABSTRACT With the development of multimedia technology, In Internet the number of images and other multimedia information in the rapidly increasing. How to locate the useful image in the vast multi-media is a great technology problem. In recent years, Content-based image retrieval has become a research focus, the technology has the superiority which the traditional text retrieval methods can not match. It has been widely used in multimedia communications, biometrics, intellectual property protection and so on, and has broad prospeets for development. This topic is designed based Color Image Retrieval System, which includes two key steps: color feature extraction and similarity matching. In this article was first used widely HSV color space to represent color feature image. We carried out color components quantification of non-interval to form feature vector; Then use the Euclidean distance metric function to match the similarity function of the image. Finally, this system should be implemented under the platform of the MATLAB by programming. Under user input sample image to conduct feature matching with image which in library images. In this system, It also can identify the images that have a short distance to the sample images, And in accordance with the distance between the images displayed to the user in the order from small to big. Keywords： Image Retrieval；Histogram Algorithm； Color Features；MATLAB 目录 TOC \o 1-3 \h \z \u HYPERL