Select to view content in your preferred language

LiDAR to DSM - bad interpolation from very high features

82
0
3 weeks ago
ZachBodenner
MVP Regular Contributor

Hello,

I am trying to create a DSM/CHM from a LiDAR dataset. I'm using the Las Dataset to Raster and going with binning/maximum/natural neighbor for the specs. The area of interest is suburban with a lot of mixed buildings, trees, and complicating features. For this particular question, I'm looking for ways or methods to handle the interpolation that occurs with the GP tool near high features.

Here's a prominent example: high voltage power lines over a lake and surrounding commercial structuresZachBodenner_0-1737132251161.png

The bright are in the left-center of the middle tile are high values written into the DSM next to the power lines. My data has high noise classified, as well as ground, but nothing else at the moment. 

Here's another one, where there are a lot of tree bordering a lake:

ZachBodenner_1-1737132393618.png

the values from the trees are extrapolated across the water where there were no LiDAR returns. 

I had a thought to at least attempt to classify the power lines, but I think there will be some buildings at the same height in other places that would get caught up in a height-based classification.

 

 

Happy mapping,
- Zach
0 Kudos
0 Replies