|
From: Shahak N. <sn...@ya...> - 2015-04-26 03:43:43
|
I'm trying to find a way to accurately (or at least semi-accurately) compute a geodetic (i.e. WGS84/spheroidal) buffer around any given geometry anywhere on the planet. I've come across the tutorials and postings like http://docs.geotools.org/latest/userguide/library/jts/operation.html, http://gis.stackexchange.com/questions/33509/geotools-buffer-using-different-crs, and http://osgeo-org.1560.x6.nabble.com/Creating-a-buffer-around-a-point-td4323303.html. First problem is that they seem to hard-code a given CRS to supplement WGS84 for the MathTransform; I'm unclear on how to find the right/best CRS if I'm not limited to, say, Belgium (Lambert 72). More importantly, though, I'm not even convinced the approach provided in these examples is entirely sound, since it still utilizes JTS buffer algorithms under-the-hood, which are computed on a 2D Cartesian plane. So, although you might be converting to/from meters and then reprojecting, the JTS operation itself will be applying a fixed distance (in decimal degree units) for the buffering operations, which is obviously wrong. Any tips/ideas? |