by frtps » Wed Nov 14, 2018 1:55 pm
I wonder how widespread this problem is. Some SRTM-1 height data for Australia I found on the USGS website has a large bump in the central district of Sydney, presumably from the tall buildings there. What happens in the centre of US cities with tall buildings? What do Manhattan and Chicago look like? I am guessing that the heights in the elevation data are average heights in the 1x1arcsec square, and so rapid changes in height (i.e. cliffs) are lost. So, I have been working on incorporating cliffs from OSM into the elevation data, and discarding any heights that are too close to the cliff edge, with some nice results which I'll share once I've tracked down some lingering issues. Perhaps if OSM contains sufficiently accurate building heights and areas the elevation data could be similarly auto-corrected by interpolating points from areas with no tall buildings.