Ion Sola Torralba
An engineer at the Public University of Navarre (NUP/UPNA) is proposing the use of synthetic images generated by means of a simplified simulation model that enables the quality of the topographic correction performed to be assessed in a straightforward, objective way.
In his research he developed a methodology to assess the topographic correction of images captured by satellite; it would enable erroneous interpretations to be avoided when the images are used for mapping land uses and cover (forests, crops, urbanised areas, etc.), spotting changes, or estimating biophysical parameters (percentage of land occupied by vegetation, foliage areas, chlorophyll content, etc.).
Land relief is responsible for significant variations in sunlight between hillsides of varying steepness and orientation. These variations in sunlight lead to distortions in the images captured by satellite and which may be erroneously interpreted in mountainous areas in particular.
An engineer at the NUP/UPNA has analysed the existing topographic correction methods to establish how to improve the quality of satellite images. To do this he took into consideration the main factors that influence the image capturing process, such as the actual features of the satellite itself; the date and time of capture, which determine the solar geometry; relief of the earth’s surface; and the spatial distribution of the covered surfaces in the area being studied.
Two main categories
In recent decades numerous topographic correction methods have been proposed; basically, they can be grouped into two main categories: the empirical ones, which do not require any auxiliary data; and the semi-empirical ones, which require the calculation of the angle of solar incidence at each point and, therefore, need a digital elevation model. The latter is a visual, mathematical representation of the altitude values of the terrain with respect to average sea level, and allows the relief shapes and elements or objects present in it (buildings, plants, trees, etc.) to be characterised.
According to the researcher Ion Sola-Torralba, “although there are numerous topographic correction methods, there are no clear rules as the which mathematical algorithm or tool should be used for each specific case depending on the type of terrain and the conditions in which the image is acquired. What is more, determining which method is the best for a particular case is no simple task, as there is no single, standard assessment criterion that can be applied”, he added.
These difficulties notwithstanding, this research has shown that these semi-empirical methods allow, in a straightforward, effective way, many of the distortions introduced by the topography to be corrected, although they cannot be corrected completely under severe conditions such as extreme sunlight. “Despite the fact that they displayed better overall behaviour, none of the semi-empirical methods tested managed to correct the topographical effect on hillsides in the shade completely,” he pointed out.
In his view, “it might be advisable for research efforts to be geared towards improving the behaviour of topographic correction methods under severe conditions in order to come up with quality products produced by satellite images”.