基于分形信号的高光谱影像增强方法

Fractal Signature based Hyperspectral Image Enhancement

目前常用的高光谱影像增强方法大多继承了多光谱影像的增强处理方法,这类方法没有充分利用光谱信息,而基于混合像元分解的图像增强方法存在端元的选取问题。基于影像的自相似特征,探索运用分形信号进行遥感影像增强的可能性。以3景Hyperion高光谱影像数据为基础,把基于地毯的方法进行修正后用于计算高光谱影像中每一像元的分形信号。结果表明,与原始高光谱影像相比,分形信号影像可以更好地突出地物特征,从而达到影像增强的目的,原始曲线形态特征、初始尺度的选择以及采样点数目对分形信号和分形特征尺度均有影响。

Most of the hyperspectral image enhancement methods,which are applied to multispectral image processing,have not fully used the spectral information of hyperspectral data.The unmixing based enhancement method is limited by endmember selecting.A fractal signature based approach to image enhancement is presented in this paper.According to modified blanket method,the upper and lower fractal signature curve corresponding to original spectral curve of each pixel can be computed,and the formed fractal signature image at each scale is used for image enhancement.Three scenes of Hyperion image are experimented in the work,the corresponding upper and lower fractal signature images are computed.It can be seen from the fractal signature curves and images that ground targets are characterized differently by signature value and feature scale.The experiment result shows that the fractal signature image can outstand targets more saliently than the original data.Thus the proposed approach can be applied to hyperspectral image enhancement.The experimental result shows that the original curve,the selection of initial scale and number of sample has a significant impact on the fractal signature,and a brief discussion on these issue is presented.