Towards a better hybrid pansharpening algorithm for high resolution satellite imagery

Most pansharpened images from existing algorithms are apt to present a tradeoff relationship between the spectral preservation and the spatial enhancement. In this paper, a hybrid pansharpening algorithm using neural networks based on primary and secondary high-frequency information injection method is used to efficiently improve the spatial quality of the pansharpened image is being developed. The injected high-frequency information in the proposed algorithm is the differential data of panchromatic and intensity images and the Laplacian filtered image of high frequency information are obtained with the help of regression method. The extracted high frequencies are injected by the multispectral image using the local adaptive fusion parameter and post processing of the fusion parameter. The proposed algorithm gives better spatial quality when compared to available fusion algorithms with high spectral information. MATLAB 13 [b] version is used to build a GUI to apply and to present the results of the image fusion algorithms. Subjective (visual) and objective evaluation of the fused images have been performed to evaluate the success of the approaches. The objective evaluation methods include Correlation Coefficient (CC), Root Mean Squared Error (RMSE), Relative Global Dimensional Synthesis Error (ERGAS), Relative Average Spectral Error (RASE), Degree Of Distortion (DD) and Average Gradient (AG).