Effect of wavelet based image fusion techniques with principal component analysis (PCA) and singular value decomposition (SVD) in supervised classification

With more promotion in satellite image processing techniques and the accessibility of various resolution images, fusion is necessary to combine panchromatic and multispectral images for further applications. Recent researches show that wavelet based image fusion algorithms provide high spectral quality in the fused images, but less spatial information in fused images due to critical down sampling .To increase spatial and spectral resolution, we have implemented wavelet based image fusion algorithms along with singular value decomposition(SVD) and principal component analysis (PCA) and its influences on supervised classification. The quality of the fused images is evaluated by quantitative and qualitative measurements. Qualitative evaluation is confirmed by edge detection methods. Quantitative results proved in terms of with reference and no reference image quality metrics. Supervised classification is used to check whether the spectral distortion caused by wavelet based fusion methods and the classification accuracy is measured by Kappa index (K). Results shows wavelet based image fusion combined with Eigen value methods such as SVD and PCA improves the classification accuracy as compared to actual multispectral images. Best classification results are achieved by framelet transform with SVD based fusion.