基于颜色特征的图像数据库检索系统开发a3pn8chv.doc

基于颜色特征的图像数据库检索系统开发摘 要 目前图像检索技术主要分为两大类：基于文本的图像检索(Text-based Image Retrieval）和基于内容的图像检索(Content-based Image Retrieval）。 本课题主要研究基于内容的图像检索技术，其中侧重于颜色特征的研究和学习。基于内容的图像检索是指直接根据图像内容的各种特征进行检索，它的研究目标是提供在没有人参与的情况下能自动识别和检索与目标图像的特征相同或相似的图像。通过对该领域的研究可以深刻地理解图像重要特征的算法。 基于内容的图像检索是一种利用图像的视觉特征(颜色、纹理、形状等)进行图像检索的技术。本论文所开发的图像检索的体系结构可以划分为两个部分:颜色特征提取和图像检索。颜色特征提取主要负责提取图像的各种低级视觉特征建立图像的特征向量，图像检索主要负责对查询样本图像和图像库中的图像进行相似性计算以便找到目标图像。 在本文中将主要论述基于颜色特征的图像检索系统的开发过程，以及颜色特征的理论知识。最后的实验结果和分析证明了本文所开发图像检索系统是有效的和高效的。 关键词：图像检索，颜色空间，直方图，特征提取 ABSTRACT Image retrieval technique mainly is divided into two major types currently: text-based image retrieval and content-based image retrieval. This topic mainly studies a technique according to content-based image retrieval, particular emphasis on in the research and study of color feature. Content-based image retrieval is that direct carry on an retrieval according to various characteristic of picture contents. The research target is provide automatically identify and retrieve the target image of characteristic homology or the picture of likeness under the situation that no one participate. By way of the research in this field, I can deeply comprehend the arithmetic of image important characteristic. CBIR is an image retrieval technique, which synthesizes various visual features in digital image, such as color, textual, and shapes features. Generally, the image retrieval system I developed is divided into two different sub-systems: feature extraction system and image retrieval system. The former was mainly devised to acquire low-level visual features and to construct image feature vector; while the later to make enquiry to sample image database and to retrieval object images by calculating the similarity functions. In this article, mainly discuss the development process of the system according to the image retrieval of the color characteristic, and the theories of the color characteristic. The experiment of the end test result proved that this image retrieval syst