1

I have downloaded classified LAS data from http://lidar.unh.edu/map/ enter image description here

I am trying to generate accurate DEM and DSM for further analysis. I created a LAS dataset with a projected CRS of NAD_1983_UTM_Zone_19N in meters and the z CRS in meters as well. My question is on which classifications do I choose for the DEM and DSM. Here is the filter options I have in my dataset:

enter image description here

According to unh LiDAR data report class 2 is ground.

enter image description here

What I have done so far:

DEM I chose the class 2 points and used the las dataset to raster.

DSM In the predefined settings I chose first return and used the las dataset to raster tool.

Is this an accurate way of generating these two rasters? Do I not have to take into account the the unassigned class 1, noise class 7, reserved 11, reserved 17, reserved 18?

Additionally (I can ask this as a separate question if it gets requested to), when using the las dataset to raster tool the sampling value is defaulted to 10. I would like to change it to 2 or 3 (would be meters) to match the LiDAR resolution. I also plan on using the generated DEM and DSM for a least cost path analysis and I want them to be higher resolution. Will making the sampling_value 2 or 3 throw off the results or should I leave it at 10?

Improve this question
1

I also would use only ground points (class 2) to create the DEM.

For the digital surface model (DSM), it is acceptable, and may be more efficient from a processing standpoint, to use only first returns. Keep in mind that point cloud data are typically normalized when creating DSMs so that elevation values are elevation above ground rather than elevation above sea level.

It appears that the nominal post spacing is 1m, so applying a sampling_value, where the sampling_type = "CELLSIZE", of 2 or 3 is acceptable

Improve this answer
4
  • great so I was correct on all 3 of my processes
    – ziggy
    Oct 15 '16 at 21:40
  • I am bit confused as to what you are saying with the normalized elevation data
    – ziggy
    Oct 15 '16 at 21:42
  • 1
    Also keep in mind that depending on what your ultimate goal is, changing the "cell assignment type" parameter will slightly modify your results. You may prefer to use "MINIMUM" for your bare ground DEM and "MAXIMUM" for your full feature DSM. Another important note is that it is very likely that your DSM and a resulting normalized canopy height model (CHM) will underestimate the true height of the tops of the trees in most cases; unless a LiDAR pulse return was captured from the true peak of the tree crown.
    – Jae
    Oct 15 '16 at 23:43
  • 1
    @ziggy Here is an excellent post that describes normalizing point cloud data: gis.stackexchange.com/a/113125/8104
    – Aaron ♦
    Oct 16 '16 at 3:33

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.