高光谱影像纹理特征编码分形特征研究

Fractal Feature of Texture Characteristic Coding for Hyperspectral Images

提出了一种基于分形测度分析高光谱影像纹理的技术。根据高光谱影像纹理单元中相邻光谱矢量的相关关系构造方向性纹理特征,得到了一组纹理特征编码值,从纹理特征编码值本身、相同地物纹理特征编码自相似性和使用不同尺度测量纹理特征编码值构成的曲线所表现出来的幂指数关系三个方面,说明高光谱影像纹理特征编码曲线具有分形特征。设计了纹理特征编码曲线构建模型,采用纹理编码曲线的分形维值表征不同的纹理特征,达到区分和判别不同纹理的目的。试验表明,该方法对提取地物纹理特性具有可行性。

In this paper, a fractal analysis technique is proposed to analyze the textures of hyperspectral images. A brief description is given to explain the nonlinear mechanism resulting in fractal of texture feature coding curve. The texture feature coding of same objects are presented to show the self similarity. And some computational results are given to show exponential relation between the total length of texture coding curves and the different measurement unit. A fractal curve model with texture feature coding is designed and the fraetal dimension value is adapted to represent different texture feature of different objects. The experiment result proves that this method is useful for the extraction of the spatial texture with spectral characteristics.