Spectral quality evaluation of pixel-fused data for improved classification of remote sensing images

Various methods proposed for image fusion satellite images are examined from the viewpoint of accuracies with which the color information and spatial context of the original image are reproduced in the fused product image. Image fusion is a useful tool in integrating a high resolution panchromatic image (PI) with a low resolution multispectral image (Mis) to produce a high resolution multispectral image and better understanding of the observed earth surface. In this study, five typical fusion methods of Gram-Schmidt (GS), Ehler, modified intensity-hue-saturation, high pass filter, and wavelet-principal component analysis (PCA) are compared. The spectral quality assessment of the products using these different methods is implemented by image quality metrics. The accuracy of classification result is assessed by means of the support vector machine based on radial basis function kernel. Our analysis indicates that as a whole, the Ehler and wavelet-PCA methods show good performances, followed by GS. Also, the examination of confusion matrix shows that both Ehler and wavelet-PCA yield better accuracies in the classification results.