2

I am attempting to use MCC-LiDAR to reclassify points (.las) originally classified using TIFFS. However, I seem to either be misunderstanding what the parameters really mean or where to get the information I need. Two parameters are needed. First, scale parameter which I have set to 0.75 since the point spacing (which is the same as post spacing?) is 0.75cm. Second, the threshold parameter which I have set to 0.5. Additionally, I set MCC-LiDAR to output an .asc grid at 1m resolution.

The environment is forested with a mixed canopy, in leaf off conditions and the ground environment is karstic. Can anyone recommend what I should input for the parameters? I'm really having a hard time understanding them.

Hillshade -- Azimuth: 135, Altitude: 15, dpi: 100

Hillshade -- Azimuth: 135, Altitude: 15, dpi: 100

Improve this question

1 Answer 1

Reset to default
2

It is not clear what the problem is. Are you getting a high rate of commission error? I would not modify the scale parameter because I found that this is a very stable initiation parameter. The curvature parameter, however, can be dependent on landscape configuration and vegetation structure. I would often start with ~0.07. Have you read the IEEE paper? I explain the curvature parameter and how it works in the model. I would also not bother outputting a raster from the C++ code. My preferred method for generating a DEM is the Topo to Raster (TOPOGRID) tool available in ArcGIS, which is an implementation of the Hutchinson IFD Spline model.

Looking at your DEM jpeg it seems like you have a large number of scatter related errors. These manifest as below ground measurements and are attributed to laser scatter often due to water, snow, etc... The MCC algorithm, by design, does not address these errors. This is something that the vendor should have taken care of. You mentioned that TIFFS processed this data. Could the software have modified the original data? If you need any additional help please email me (you can get my email from my stackexchange profile).

Improve this answer
4
  • I reran MCC-LiDAR using the values you suggested (1.5 & 0.07). After interpolating with an IDW, I ended up with a lower RMSE than I had in prior iterations. Thanks for your help!
    – jwx
    Apr 10, 2013 at 23:47
  • 1
    @jwx, If you have access to ArcGIS use the "Topo to Raster" tool. It is an implementation of the IFD Spline model and will give you much better results than IDW.
    – Jeffrey Evans
    Apr 10, 2013 at 23:58
  • Okay, I will give that a shot once I figure out how to get the data in a usable format for Topo to Raster from LAS format. Previously, when I have tried to run the tool the Shape.Z Field (in the Multipoint data) has not been selectable. I suspect it has to do the the field format, but am uncertain how to translate it to a correct format.
    – jwx
    Apr 11, 2013 at 0:07
  • 1
    You can convert LAS to a shapefile using the "las2shp" function in LASTools (cs.unc.edu/~isenburg/lastools). He has also build a nice ArcToolbox that is a wrapper for the commandline (exe) tools.
    – Jeffrey Evans
    Apr 11, 2013 at 0:14

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.