Modified Synthetic Variable Ratio Pansharpening Method

Pansharpening, which is transferring the spatial content of a high-resolution panchromatic (PAN) band into a lowerresolution multispectral (MS) image to produce a spatially enhanced MS image, has always been one of the hottest topics of image processing. Numerous studies have focused on developing approaches to inject the spatial details with minimum spectral distortion. This study utilized the Genetic Algorithms (GA) to improve the performance of the Synthetic Variable Ratio (SVR), which is one of the most conventional pansharpening methods. This method was modified such that the weight of each MS band was estimated by means of a GA to achieve the optimum result. The spectral quality of the image produced by the proposed approach was compared against those of the images obtained from widely-used pansharpening algorithms Principal Component Analysis (PCA), Modified IHS (MIHS), Gram-Schmidt (GS), Nearest Neighbor Diffuse (NND), High-Pass Filtering (HPF) and conventional SVR. The quantitative evaluation of the pansharpening results revealed that the proposed approach resulted in superior spectral quality, compared to the other methods.