基于小波包和分形相结合的纹理分类  

Classification of Texture Based on Wavelet Package and Fractal

作　　者：杨敏[1] 杨峰[1] 刘臻 郭韡[1] 郑岚[1] 

机构地区：[1]中国矿业大学(北京)机电学院,北京100083 [2]山东联通枣庄总公司,山东枣庄277000

出　　处：《现代电子技术》2008年第9期161-162,176,共3页Modern Electronics Technique

摘　　要：纹理是图像内容的重要特征之一,纹理图像分类中最重要的步骤是纹理特征的提取。纹理图像利用空间-频率域分析法时,纹理特征主要集中在中、高频段,纹理图像边缘包含着丰富的纹理特征信息。利用小波包和分形相结合的方法提取纹理特征值进行纹理图像分类,实验结果表明此算法是合理有效的,纹理分类的正确识别率较高。Extracting texture features is the most important measure, the texture image uses space - frequency domain anal-ysis method, feature of texture is mainly in middle and high frequency band. Texture image is classified by combining wavelet package with fractal and extracting texture features. Experiments prove this algorithm to be true and effective, the algorithm improves the ratio of correctly classify texture.

关 键 词：纹理特征提取 小波包 分形 纹理分类 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]