for making a model for Urban Drainage I am doing watershed delineation for an urban area. I am using a 1 meter resolution LIDAR DEM. the elevations of building surfaces are misleading . How can i reduce the impact of such problem.
Thanks allot in advance for your help and support.
this seems to be difficult to do, from what you describe. If you reduce the elevation of the building areas by their height (assuming you have building height info and building polygon info), these areas will become flat, which also do not help for delineating catchments based on DEM information.
Better look for a DEM without the buildings represented in it.
Not sure if this comment help.
Regards,
Joao
Hi Fazli,
if you don't have the point cloud (.las file), I imagine there little you can do, as Joao said.
Sorry if I "answer" with another question, but why in urban watershed delineation buildings have not to be taken into account?
True is that you probably don't get a perfect roof shape for buildings and that can cause problems when processing flow accumulation algorithm and stream flows.
Probably is the word urban befor watershed tha misleads me! I would take buildings, roads and human assets into account to see how the human infrastructure (city) superimposed on the natural landscape convoys water.
Let me know I am curious!
Nic
A way to do that on a Raster DEM would be to construct a 3D polyline around the building base to create a special DEM you will operate as an update of your LIDAR DEM to "shave" the buildings off. Usually, one takes 2D polylines of the footprint, that you project on the dem surface to create 3D polylines, then you create the new update surface model you use on your LIDAR DEM to flatten the building space.
However, as mentionned by João P. Leitão, it seems strange to try and remove buildings in a watershed analysis, One usually removes vegetation surface instead (to avoid hedges acting as walls for example).
If you have the raw point clouds (las files), there is two possibility :
- you might have some data in your points that distinguish between ground and not ground
- you can use http://www.cs.unc.edu/~isenburg/lastools/ to find buildings (lasground/) then classify the data (lasclassify/) then remove buildings and regenerate your DEM (las2dem/)
If you don't have the las files, maybe you could try creating a las point cloud from your DEM data, then using lastoolsas mentionned previously ?
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India