lasvoxel metrics

[Susana Gonzalez]

Hi Martin,

We are trying to write a workflow to generate voxel metrics for imputation over large areas, the voxelisation is at 5x5x1 but the summary metric (as a toy example let us say filled voxels/total voxels) is to be computed at a 25 x 25 m resolution (so 5x5 voxels). We usually achieve summary metrics at 25 m resolution using lascanopy. Do you have a suggested workflow for using the outputs of LASvoxel with LAScanopy to get the summary metrics at this resolution? We can do this in intermediate steps with other tools but we would like to start using lasvoxel + lascanopy to generate more metrics to better define our forest structure.

 

Example of the normalised LiDAR pixel (attached)

 

 

We have tried two approaches:

A)

lasvoxel.exe -i %NORMALISE% -step_xy 5 -step_z 1 -odir %VOXEL% -odix _xy5_z1_Count_Int -olaz

Here is what we obtain, coloured by intensity, I think in the intensity field we have the number of points (count) for each particular voxel

From this file we could compute our voxel metrics, the voxels are 5*5*1 but they will need to be aggregated in 25 by 25m, at this stage we can run lascanopy on the intensity field, so this is not a problem.

 

B)

We are trying as well, that I think is where we actually want to go to extract the voxel metrics.

lasvoxel.exe -i %NORMALISE% -step_xy 5 -step_z 1 -compute_IDs_and_voxel_table -store_IDs_in_intensity -odir %VOXEL% -olaz -odix _xy5_z1_vID_Int

where the voxel id is saved in the intensity field.

Coloured by Voxel_ID (intensity)

 

It would be interesting to be available to compute the standard metrics by voxel id and then aggregate them in 25 by 25m, ready for the LiDAR imputation.

 

Tell me your thoughts.

Susana

 

Susana Gonzalez - Forest Engineer, LiDAR Science

 

Interpine Group Ltd

DDI:  +64 7 350 3209 or Australia 0280113645 ext 722
Skype:        sugonar

Address :   99 Sala Street, PO Box 1209, Rotorua 3010, New Zealand
Website :   www.interpine.co.nz

 

Interpine Innovation is Shaping Today’s Forests with the Technology of Tomorrow

He rangahau tenei ra he hangarau apopo

 

[Martin Isenburg]

Hello Susanna,

I am not entirely clear what the question is. Are you asking for additional voxel metrics to be implemented? In a new paper called "Comparison of models describing forest inventory attributes using standardand voxel-based lidar predictors across a range of pulse densities" by Grant D. Pearse, Michael S. Watt, Jonathan P. Dash, Christine Stone, and Gabriele Caccamo that I am sure you are familiar with they list several voxel metrics. Which one of those (a) is difficult or impossible to produce through a combination of lasvoxel and lascanopy and (b) is essential to your experiments. I've appended a screenshot of the appendix of voxel metrics from their paper. 

Your other question is whether "It would be interesting to be available to compute the standard metrics by voxel id and then aggregate them in 25 by 25m, ready for the LiDAR imputation." How would you do this? Currently metrics such as averages, percentiles, bincentiles, standard deviation, skewness, kurtosis only take the y component into account. Is that what you are looking for? The same z-coordinate metrics per voxel rather than per plot? I would expect this to be quite sensitive to the voxelization with a small horizontal/vertical shift in the voxel grid changing results quite a bit ...?

Regards,

Martin

--
Download LAStools at
http://lastools.org
http://rapidlasso.com
Be social with LAStools at
http://facebook.com/LAStools
http://twitter.com/LAStools
http://linkedin.com/groups/LAStools-4408378
Manage your settings at
http://groups.google.com/group/lastools/subscribe

[Simon Papps]

I am also puzzled. It looks as if the process is to use the -compute_IDs_and_voxel_table argument to write 

   voxel_ID,center_x,center_y,center_z,number_of_points 

to stdout, redirect to a text file, and then compute the voxel metrics from this data.

I could do this in R but it is going to be slow.....

Simon