Producing WorldView-2 fused images of superior quality by the novel ELSHORA fusion technique

Abstract

In this study, the novel ELSHORA fusion technique was developed for the fusion of the WorldView-2 satellite panchromatic (PAN) and multispectral (MS) images. This fusion technique has the advantage of overcoming the weaknesses of the other existing fusion techniques and producing fused images of superior spectral and spatial quality for all land cover types. This technique uses a modification coefficient for each MS band according to its intersecting area with the PAN band to ensure that only the wavelengths of the MS bands within the PAN band range participate in the definition of the I image, and the I image will be a weighted average of the eight modified MS bands. These modification coefficients will help in the preservation of the original colors as well as achieve spatial and temporal transferability for the ELSHORA fusion technique. This technique also uses an additional coefficient for the NIR band in the agricultural areas to indicate the correct effect of the vegetation, as its reflectance is high in the NIR band. This vegetation coefficient will achieve the performance stability for the ELSHORA fusion technique across the different types of land cover. To evaluate the performance of the ELSHORA fusion technique, it was compared to six standard image fusion techniques: modified IHS, Ehlers fusion, hyper-spherical color space, principal component analysis, Brovey transform, and multiplicative resolution merge. These fusion techniques were utilized to merge the spatial and spectral information of four datasets of WorldView-2 satellite PAN and MS images covering different land cover types: agricultural, urban, and mixed areas. The four datasets were chosen in two different places and acquired at two different times to evaluate the spatiotemporal transferability of the ELSHORA fusion technique. The fused images were compared to the PAN and MS images, as well as to each other, statistically and visually. The results demonstrated the superiority of the ELSHORA fusion technique for all types of land cover. It can effectively generate sharper fused images without color distortion at different times and places.