本文提出了一种基于特征融合的纹理图像分类方法，它结合了纹理图像的特

点和框架小波变换方法，处理过程中充分考虑了图像各尺度间的依存关系以及不

同频带中所包含的图像纹理信息，利用支撑矢量机作为分类器，对标准纹理库中

的图像进行了仿真实验。

-This paper presents a texture feature fusion based image classification method, which combines the special texture

Point and framework for wavelet transform method, the process in full consideration of the image dependencies between the various scales and different frequency bands in the information contained in the image texture, the use of support vector machine as classifier, the texture library of standard images for simulation experiments .