模板匹配技术在图像数据库检索中的应用

图像的检索应采用基于图像对象本身(即图像内容)的检索方式,目前流行的各种图像匹配技术主要针对不同类型图像的特征,需要图像先验知识,缺乏广泛性。模板匹配技术理论成熟,但其计算量大缺乏效率而不受重视。因此对传统的模板匹配技术进行分析,采用小波变换技术抽取图像主要结构为逻辑对象,利用改进的模板匹配技术进行检索,实验结果表明可明显减少计算量,提高检索效率。

Application of Template Matching Technology to Image Database Retrieval

Image searching is the most essential function of the image database,to which traditional database searching technologies have their limitations.Image searching should be based on image content,but current various kinds of image matching techniques are mostly suitable for different kinds of images' features,which are lack of extensiveness.Template matching theory is very perfect, but lack of efficiency because of complicated calculation,so is ignored by people.Through experiment this paper analyzes traditional template matching technique,uses wavelet to get major constructer of image as logic object,and conducts matching images with improved template matching technique,thus greatly decrease the amount of calculation.