| Abstract: |
Obtaining reflectance for airborne HSI data requires a good illumination map of the image area. This factor becomes critical when working with images acquired above rugged terrain and/or forest areas. Illumination maps are calculated from solar geometry and two important descriptive parameters of the terrain: slope and aspect. However, calculating slope and aspect is problematic when using DEM raster files with low-to-medium spatial resolution. These traditional methods cause various problems, such as the presence of artefacts, requiring the application of smoothing processes. Furthermore, a regular grid does not take into account the specific geometrical characteristics of sensors with large FOV, in which pixel sizes vary considerably within an image, as is the case for INTA AHS and CASI sensors.
This paper presents a methodology based on the use of raw image geometry for the entire process of obtaining reflectance in rugged terrain. For the calculation of slope and aspect the actual dimensions of each pixel on the ground are considered, disregarding the data from neighbouring pixels. To this end, data obtained from ALS is used, to extract the altimetric data for the actual polygon defining each pixel on the ground. Planes are subsequently adjusted to obtain slope and aspect maps for imagery. In this way, more realistic data is obtained on the morphology of the areas that appear in the image, which may later be used to calculate the reflectance. |