Implementación en grass-gis de técnicas de fusión de imágenes digitales

Abstract

In the field of remote sensing, there are different ways to integrate images from different spectral and spatial resolution in order to obtain fused images that enhance the capabilities which present separately. Several of these techniques have been implemented, mostly on proprietary software, which limits free software users access to these geo-informatics tools. This study had a threefold aim with the first being to implement three image fusion techniques in free software (GRASS-GIS): High Pass Filter, Principal Component Analysis and Gram-Schmidt. The second aim was to apply these techniques to merge multispectral and panchromatic images of four satellite platforms (QuickBird, IKONOS, Landsat 7 and Landsat 8) and an airborne platform (Natmur-08 Project). The third and final aim was to assess the results qualitatively by a visual comparison and quantitatively by two quality indices implemented in GRASS in this research: Universal Image Quality Index and ERGAS Index. The initial results demonstrate that the application of fusion techniques, both GRASS and proprietary software, produce very similar results. With regard to the application of fusion algorithms implemented in GRASS, qualitative evaluation reports better assessment to images with higher spatial ratio. The results of the quantitative assessment indicate greater dispersion in terms of the type of image used, but greater consistency between the valuation provided by each index for the same image analyzed.