From: Sunburned S. <sun...@gm...> - 2007-06-30 15:07:13
|
I don't know if this is relevant in this conversation, but one reason for the inaccuracies in DEMS are the problems and expense in collecting elevation data. This is especially difficult in urban areas or heavily forested areas where the "actual ground surface" is obscured. No matter how good your "line-of-sight" algorithm is it will only be as good as your underlying terrain model. :] The Sunburned Surveyor On 6/26/07, Andrea Antonello <and...@gm...> wrote: > Hi Rob, thanks for your reply. > > > I've yet to see many robust implementations of visibility analysis -I > > used to use this as the case study for how to lie with GIS when I > > lectured on the subject. > > > > variability within the limits of precision of DEMS (or whatever else you > > use to map the terrain) cause huge error margins in the computed > > visibility. Any accuracy issues (actual errors outside the limits of > > precision) are disastrous.. > > > > I saw one example using Monte Carlo techniques (to introduce repeated > > random changes within the limits of precision) that concluded that the > > actual viewshed cannot be measured to within about 50% of the real > > value. I.e. if you think you have 100 ha of viewshed, it may be 50 or 200! > > I was thinking about that also. I talked to people that want to do > analysis using srtm 90 m dems. I uderstand your issues. > > > Still, it makes pretty pictures to please politicians I suppose, but > > please be careful about implementing a poor aproach (i.e. if the inbuilt > > tools created a min and max and average position for a line of sight, > > this would promote better analysis, and make the tool more important. I > > dont know how hard this is to achieve for an entire viewshed - suspect > > the algorithms would be quite tricky, but for a single line of sight its > > pretty simple. > > I would not want to implement a thing from scratch. I know GRASS has a > los analysis tool, but I couldn't find one that does it simply between > two points. So my question was more about: is there something in > geotools already that does something similar, since I would like to port > such a tool into udig? > > However, if you make some accurate tests, you probably can give me an > advice about GRASS's module? I was considering to look into the code to > link it to udig. > > At that point I'm not sure if I'm not getting OT... > Sorry > > Andrea > > > > > > > Rob > > > > Andrea Antonello wrote: > >> Hi developers, > >> is there some class that does line of sight (visibility analysis) taking > >> into consideration the earth curvature? > >> I think I saw something, but I can't find it any more. > >> > >> Thanks, > >> Andrea > >> > >> ------------------------------------------------------------------------- > >> This SF.net email is sponsored by DB2 Express > >> Download DB2 Express C - the FREE version of DB2 express and take > >> control of your XML. No limits. Just data. Click to get it now. > >> http://sourceforge.net/powerbar/db2/ > >> _______________________________________________ > >> Geotools-devel mailing list > >> Geo...@li... > >> https://lists.sourceforge.net/lists/listinfo/geotools-devel > >> > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Geotools-devel mailing list > Geo...@li... > https://lists.sourceforge.net/lists/listinfo/geotools-devel > |