一种小波域内图像的文本信息提取算法  

Study on text information extraction algorithm of image in wavelet domain

作　　者：郑雄波 张晓威 朱磊 

机构地区：[1]哈尔滨工程大学理学院应用数学系,哈尔滨150001

出　　处：《哈尔滨商业大学学报（自然科学版）》2008年第3期281-284,304,共5页Journal of Harbin University of Commerce：Natural Sciences Edition

摘　　要：研究了基于小波变换的图像文本信息提取模型,提出了一种在小波域内提取图像文本信息的算法,该算法首先对图像进行二维小波变换,设置滑动窗扫描高频子带,计算滑动窗内图像的小波纹理特征,采用k-均值聚类算法提取文本区域,再对原图像进行边缘检测,根据文本区域的相关特征精确地定位文本区域,最后将文字与背景分离.从实验结果可以看出该算法能够较准确地从背景图像中提取出不同语种、字体、大小的文本信息.This paper studies text extraction models for images based on wavelet transforms, a text extraction algorithm for images in wavelet domain is proposed. For the algorithm, twodimensional wavelet transform is used to the image firstly and slipping windows are set to scan high-frequency subband, through computing the wavelet texture feature of the image in slipping windows,k - means cluster algorithm is used to extract text area. The edge detection are put on the text area and original image respectively. According to the character of text area, the edge detection are located exactly. The text and background are segmented. The experimental result shows that the algorithm can extract texts information with different languages, fonts, size from the background image exactly.

关 键 词：小波变换 文本提取 K-均值聚类算法 边缘检测 纹理特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]