Survey on pixel level image fusion techniques

Abstract

Most existing fusion techniques such as Principal Component Analysis (PCA), Intensity-Hue-Saturation (IHS), Wavelet Transform (WT) yields color distortion in other words it is also known as spectral distortion problem. When the high resolution panchromatic image unite with low resolution multispectral image results in fused image of high spatial resolution multispectral image. The above said fusion techniques require additional transformation to increase spatial information and also decrease spectral distortion in fused image. Another efficient fusion method named as linear pixel level fusion overcomes this color distortion. The main advantage of this method is to produce color image same as that of natural image by adjusting the correlation properties with the help of the user and no additional transformation needed. The fused image produced using linear pixel level fusion method is suitable for human visual perception. In this paper correlation coefficient of different fusion methods are analyzed and adding additional transform it increases and reduced color distortion is obtained.