基于形态特征提取的图像匹配搜索技术研究  

Research on image matching search technology based on morphology characterization extraction

作　　者：张华[1] 

机构地区：[1]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430079

出　　处：《物联网技术》2013年第11期16-18,22,共4页Internet of things technologies

摘　　要：传统的图像搜索方法一般是由图像处理软件自动抽取图像的颜色、形状、纹理等特征,并以此建立特征索引库,进而由用户输入要查找的物品图像,从而找出与之具有相近特征的图像。而文中给出了从数学形态学的角度来提取图像的关键形态特征,然后建立海量物品图片的形态细化骨架库,并以此简化图像搜索的关键内容,降低数据库存储量,提高匹配效率以及准确性的具体方法。The traditional image search method is generally performed by image processing software for automatic extraction of image color, shape, texture and other characteristics to establish the feature index database and find out the needed images with similar features according to the sample images users input. A method to extract the key morphological feature according to the mathematical morphology is introduced in this paper, in which the morphological skeleton thinning database of massive goods is established in order to simplify the image search key content, reduce the database storage, and improve the matching efficiency and accuracy.

关 键 词：数学形态 特征提取 骨骼细化 图像搜索 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]