基于IHS变换的Gram-Schmidt改进融合算法研究

Research of improved Gram-Schmidt image fusion algorithm based on IHS transform

针对Gram-Schmidt(GS)变换算法在融合高分二号卫星(GF-2)多光谱(MSS)和全色(PAN)影像时产生的光谱失真和细节模糊问题,提出一种基于Intensity-Hue-Saturation(IHS)变换和GS变换的影像融合改进算法(IGS)。该算法首先将原始多光谱影像变换至IHS空间,取I分量作为GS变换第一分量进行GS变换;然后将I分量与全色波段进行直方图匹配,并选择3*3滑动窗口以标准差为判定条件模拟出新的全色波段;最后用模拟全色波段替换I分量进行GS逆变换得到融合结果。将IGS融合结果与IHS,GS以及Principal Component Analysis(PCA)变换融合结果进行对比,选择主客观两种方法对融合结果进行质量评价,结果表明IGS算法在光谱保留和提高影像细节方面要优于其它3种算法,能够更好地用于GF-2影像的融合,为后续的应用研究提供参考。

Aiming at GS transform algorithm produces the spectral distortion and the detail blur problems in the fusion of multispectral image （MSS） and panchromatic image （PAN） of the GF 2 satellite, an improved algorithm（IGS） based on Gram Schmidt （GS） transform and Intensity Hue Saturation （IHS） transform is proposed. Firstly, the original MSS is transformed into IHS space, and the I component is taken as the first component in GS transform. Then the l component is matched with the panchromatic, by selecting the 3 ~＋ 3 sliding window to create a new panchromatic band with the standard deviation. Finally,the I component is replaced by the simulated panchromatic band, and the inverse Gram Schmidt transformation is performed on the new set of transform bands to produce the enhanced spatial resolution multispectral digital image. The improved algorithm is compared with IHS, GS and Principal Component Analysis （PCA）. Moreover, the results are evaluated in both objective and subjective. The result shows that the improved algorithm is better than the other three algorithms in spectral preservation and image detail enhancement, which can be better used for the fusion of GF 2 images, and provides a reference for further application research.