论文总字数：18485字

摘 要

随着计算机技术、人工智能等学科的迅猛发展，图像信息数量迅速增加。在庞大的信息数据资源面前，传统的文本检索方式已经远远不能满足人们的检索需求。用户能够在庞大的数据库中精准并快速搜寻到自己感兴趣的图像，这是一个迫切需要解决的课题。众所周知，该技术已是研究计算机视觉领域的热点之一。相比于传统文本检索方式，它把管理者从大量、枯燥的人工管理模式中解放了出来。这一技术也在交通检测、天气预报、医疗图像管理和海底探测等领域得到了广泛的运用，具有远大的发展前景。

本课题的目标是设计出一种颜色分类算法，进而实现对图像的分类。本文选用的是广泛应用于图像处理的高斯模型。该课题的实现蕴含两个关键点：一是提取图像的颜色特征，二是相似度的匹配。本文首先介绍了用于表达颜色特征的颜色模型和颜色模型的转换及颜色量化方法；然后依次介绍了颜色特征提取和相似度的匹配；最后在MATLAB软件平台下实现对图像的分类。根据用户输入的示例图像，与图像数据库中的图像进行特征匹配，找出符合用户需求的若干幅图像，并根据图像之间的匹配程度，按照由小到大的顺序显示给用户，满足用户的不同需求。

关键字：图像检索 颜色特征 相似度匹配 MATLAB

Image Color Classification Based on Gaussian Model

Abstract

With the speedy evolution of computer technology and artificial intellegence, the amount of Internet image information has rapidly increased. In the face of huge information data resources, the traditional text retrieval methods are far from satisfying people's search needs. Users can quickly and accurately retrieve the images they need in the large database. This becomes an urgent problem to be solved. As we all know, content-based image retrieval technology has become one of the research hotspots in the field of computer vision. Compared with traditional text retrieval methods, this technology liberates managers from a large number of boring manual management models. This technology has also been widely used in areas such as traffic detection, weather forecasting, medical image management and sea bottom detection, and has a promising future.

The goal of this topic is to design a color classification algorithm to further classify images. This article chooses a Gaussian model that is widely used in image processing. The realization of this topic contains two key points: one is to extract the color features of the image, and the other is to match the similarity. In this paper, we first introduce the color model and color model conversion methods and color quantization used to express color features. Then we introduce color feature extraction and similarity matching in order. Finally, the image classification is implemented on the MATLAB software platform. According to the example image input by the user, feature matching is performed with the image in the image database to find a number of images that meet the user's needs, and displayed to the user in the order from small to large.

Key words: Image Retrieval；Color Features；Similarity match；MATLAB

目 录

摘 要 I

Abstract II

第一章 绪论 1

1.1课题研究背景 1

1.2基于图像内容的检索研究状况 1

1.3基于颜色特征的图像检索特点及发展方向 2

1.4论文组织结构 4

第二章 颜色模型及转化 5

2.1颜色模型的分类 5

2.1.1RGB颜色模型 5

2.1.2HSV颜色模型 6

2.2颜色模型转化 7

2.3颜色量化 8

2.3.1均匀量化 8

2.3.2非均匀量化 8

2.4本章小结 9

第三章 颜色特征提取方法 10

3.1颜色直方图 10

3.2颜色矩 10

3.3颜色聚合向量 11

3.4本章小结 11

第四章 颜色特征的相似性度量 12

4.1直方图相交法 12

4.2二次式距离法 12

4.3欧式距离法 13

4.4本章小结 13

第五章 图像检索系统的设计与实现 14

5.1系统设计与实现 14

5.1.1开发工具的选取 14

5.1.2系统框架 14

5.1.3用户界面 15

5.2系统功能模块设计 17

5.2.1选择示例图 17

5.2.2特征提取模块 18

5.2.3特征匹配模块 20

5.2.4图像显示模块 21

5.3性能评价标准 23

5.3.1查全率与查准率 23

5.3.2匹配百分比 24

5.4本章小结 25

第六章 全文总结与未来展望 26

6.1全文总结 26

6.2未来展望 26

参考文献 27

致谢 28

第一章 绪论

请支付后下载全文，论文总字数：18485字