Spatial and radiometric characterization of multi-spectrum satellite images through multifractal analysis

Abstract

Abstract. Several studies point out that vegetation indexes can been used to estimate root zone soil moisture and earth surface images, obtained by high resolution satellites, give presently huge information on these indexes based in several wavelengths data. Because of the potential capacity for systematic observations at various scales, remote sensing technology extends possible data archives from present time to over several decades back. For this advantage, enormous efforts have been made by researchers and application specialists to delineate vegetation indexes from local scale to global scale by applying remote sensing imagery. In this work, four bands images have been considered, involved in these vegetation indexes, taken by satellites IKONOS-2 and LANDSAT-7 of the same geographic location to study the effect of both spatial (pixel size) and radiometric (number of bits coding the image) resolution on these wavelength bands. In order to do so, a multifractal analysis of these multi-spectral images was applied in each of these bands. The results showed that spatial resolution has a similar scaling effect in the four bands, but radiometric resolution has a larger influence in Blue and Green bands than in Red and Near InfraRed bands. This information should be taken in to account when vegetation indexes based on different satellite sensors are obtained.

Método de segmentación basado en la estructura fractal del mapa de singularidades : aplicación a imágenes de uso agrícola

“…Todos los EMF resultaron curvas parabólicas convexas con distintas simetrías dependiendo de la estación del año del mapa de NDVI estudiado. En un sistema fractal el EMF consiste en único punto, por tanto los resultados obtenidos apoyan la hipótesis de comportamiento multifractal de los mapas de NDVI, en acuerdo con otros trabajos previos (Lovejoy et al, 2008;Alonso et al, 2017 Respecto a la evolución de a lo largo de un año, la simetría cambia desde "a derechas", en invierno y primavera, hacia "a izquierdas", en verano y otoño.…”

“…agrícola 84 calculados para cada pixel y la distribución espacial de los mapas de NDVI fueron considerados como una medida o distribución de probabilidad normalizada en R 2 . Para asegurar la validez del análisis multifractal todos los valores negativos del NDVI fueron convertidos a cero, tal como se ha realizado en otros trabajos(Alonso et al, 2017).…”