I want to calculate a viewshed based on ASTER DEM data. Unfortunately, the data contain a lot of artefacts in this region, most of which are caused by poor coverage. Even (or especially) in flat areas, the DEM shows a rather cratery landscape with lots of small holes.
Since the ASTER data come with a quality assessment file, I wonder whether I could use this to optimise the dem for my viewshed. The num-file indicates the number of readings, which the elevation of the respective pixel in the DEM is based on. Where there were no readings, the value was taken from SRTM-data (which seem to cause a lot of these craters). This is indicated in the NUM-file by value -1.
Would it be possible to interpolate the DEM and give the grid cells with more readings a higher weight than those with few readings or SRTM-values? Some sort of weighted neighbourhood statistics?!
