Spatial enhancement of Digital Terrain Model using shape from shading with single satellite imagery

Abstract

Digital Terrain Models (DTMs) play a key role in many geoscience, civil engineering as well as military applications. Despite the fact that there are tremendous amounts of regional and global DTM data available today, there is always a growing demand to denser and more accurate DTMs. On the other hand, the research in Shape from Shading (SFS) using multispectral remotely sensed data is motivated by the wide availability of high resolution multispectral satellite imagery such as Landsat, SPOT, IKONOS, and IRS. Comparisons of the conventional monochromatic and multispectral imageries show the advantages of using multispectral imagery with SFS. In this paper the possibility and applicability of SFS with remotely sensed data to enhance the spatial resolution of DTMs have been investigated. Based on the characteristics of multispectral satellite data, strategies for applying SFS to single satellite imageries are proposed. Experimental results are analyzed and discussed with reference of the true DTMs of the study area. © Springer-Verlag Berlin Heidelberg 2003.