超谱图像小波包融合方法研究

Research on Wavelet Packet Fusion Method of Hyperspectral Image

针对超谱图像的特征选择问题,为了有效降低超谱图像数据维数,最大限度地保留超谱图像的分类特征,结合多源数据融合的理论,提出了一种基于小波包多分辨率分解的数据融合方法,给出了具体实现过程,并通过仿真实验研究了特征图像维数及分解层次对融合效果的影响。

Aiming at the feature selection of hyperspectral image, in order to decrease the image data dimension effectively, keep the classification feature of the hyperspectral image farthest, consult the theory of multi-source data fusion, a kind of data fusion method based on wavelet packet multi analyzing is illustrated , put forward the detailed realization procedure, and researches on the affect of the dimensions of feature image and analyzing levels to the fusion effect.