基于亮度平滑滤波调节(SFIM)的SPOT5影像融合

遥感影像融合不仅要求提高原多光谱影像的空间分辨率,更重要的是保留影像的光谱信息,减少失真。为分析SFIM算法融合SPOT-5影像的光谱保真效果,首先采用IHS、小波和SFIM算法对SPOT-5的全色影像和多光谱影像进行融合,生成新的高分辨率多光谱影像。然后采用均值、相关系数、熵、标准差、梯度五个指标对SFIM不同滤波窗口的融合结果评价;最后采用以上定量评价指标对不同方法的融合结果进行分析评价。结果表明:和传统的IHS、小波融合方法相比,SFIM融合不仅提高了影像的空间分辨率和清晰度,而且较好地保持了原多光谱影像的光谱特征。

The SPOT-5 Image Fusion by Smoothing Filter-based Intensity Modulation(SFIM)

The remote sensing image fusion requests not only to improve original multispectral image spatial resolution,but also to retain the image spectrum information,and reduce distortion.In order to analyze the effect of SFIM algorithm on fusing multispectral images with panchromatic images of SPOT-5,firstly IHS algorithm,wavelet algorithm and SFIM algorithm are used to fuse panchromatic and multispectral images to produce a new high-resolution multispectral one.Then five quantitative indexes such as the average value,the correlation coefficient,the entropy,the standard difference and the gradient are used to evaluate the effect of SFIM fusion with diverse filter windows.At last,the above five indexes are also used to evaluate the fusion effect of those methods.The results show that SFIM fusion has not only raised the spatial resolution and definition of primary images compared with traditional algorithm,for example,IHS,wavelet,but also well reserves the spectrum characteristics of original image.