2

I am using ArcGIS and have access to Spatial and 3D Analyst extensions.

http://imgur.com/a/c4k84

I've attached a link to a picture of a 1m raster “canopy height model” which was derived from LiDAR and displays tree heights (in 5m intervals but can obviously be adjusted). I do 3d models w/ a program that uses polygon inputs to derive height classes.

How can I convert this raster to a polygon and capture broad polygon classes?

(I can’t use raster to polygon as I work with 100’s of hectares at a time and an individual polygon for each raster cell would be far too much detail and slow the program down.)

I’m thinking of a heat map (?) type exercise which captures general trends as polygons. I’ve circled several polygons in blue that would be the ideal type of polygons I would like to capture…but ultimately I do need 100% polygon coverage.

I have access to SA and 3D analyst if that helps! In terms of accuracy, I'm not sure. As accurate as I can be without having 1m polygons or anything too small.

  • 2
    It would be best to insert your image into your question, not all users can or will follow links. It sounds like you're trying to generalize your CHM into regions with the final results being polygons... is that correct? I would suggest generalizing your raster before trying anything. How accurate are you intending the polygons to be? – Michael Stimson Jul 18 '17 at 22:12
  • I am using ArcGIS and have access to SA and 3D analyst if that helps! In terms of accuracy...I'm not sure. As accurate as I can be without having 1m polygons or anything too small. – GIS_Dave Jul 18 '17 at 22:53
1

In order not to receive super small polygons after the conversion, try to first smooth / generalise your raster, i.e. using the Filter() or FocalStatistics() out of the Spatial Anaylist Toolbox. You'd have to play around using those in order to find the solution fitting your needs. The FocalStatistics() tool basically does the same as Filter(), except you have more choice on how you'd like to generalise your data. Try different neighbourhoods for instance or run the tool iteratively (use the generated output as the input raster for the next run with FocalStatistics()).

Finally use the Reclassify() tool to group your raster into integer values representing the different classes of canopy heights. Exporting this integer raster to polygons will result in full polygon coverage. If needed you could still perform "clean up" on your polygon data, using Dissolve() or Union() for instance.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.