一种基于颜色和纹理特征的图像研究  

Image Research Based on Color and Texture Features

作　　者：郭丽珍 Guo Lizhen(Automation Engineering Depatment, Shanxi Engineering Vocational College, Taiyuan Shanxi 030009, China)

机构地区：[1]山西工程职业学院自动化工程系,山西太原030009

出　　处：《山西电子技术》2022年第2期93-96,共4页Shanxi Electronic Technology

基　　金：国家级职业教育教师教学创新团队课题研究项目(YB2020020302)。

摘　　要：在K-means算法基础上,结合主颜色直方图检索只考虑图像颜色信息,分块法只描述图像空间信息,本文提出了一种将图像等面积分块法。实验证明提高了查准性和查全率。最后实验演示出基于颜色和纹理特征的三种检索效果图。On the basis of K-means algorithm,combining the main color histogram retrieval only considers the color information of the image and the segmentation only describes the spatial information of the image,a method of image equal area segmentation is proposed.Experiments show that the accuracy and recall rate are improved.Finally,three kinds of retrieval effect images based on color and texture feature are demonstrated.

关 键 词：图像检索 颜色检索 主颜色直方图 K-MEANS 共生矩阵 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]