Re: [LAStools] Creation of DSM beneath vegetation in Eastern Norway

[Terje Mathisen]

Ben Robson wrote:
> Hi,
> I'm sure this question has been asked elsewhere, but I had a good luck
> and couldn't find it. I am not very familiar with LiDAR, I just know the
> basics, I will soon have some data from Eastern Norway and I'd like to

Ben, I'm based in Oslo, I'd be happy to help you!

Send me direct email if you'd like to talk.

> create an elevation model stripping away the vegetation in order to see
> deformation to the ground caused by icebergs during a large flood

The ground scouring happened mostly during our last glaciation period
about 10K years ago, right?

It is extremely obvious in many areas close to the sea, as well as up in
the mountains: Both areas have very little to no vegetation cover.

> thousands of years ago. I saw Martin demonstrating this at the LiDAR
> workshop in Salzburg at the end of last year, but when I looked on the
> lastools homepage I found both lascanopy and lasground, which is more
> suitable? And is there any sort of step by step for beginners?

lasground will do what you like, I'd suggest something like this:

lasground -i *.las -keep_class 2 -odir ground_only -olaz

You can see what the result looks like with

lasview ground_only\*.laz

>
> Anyway, thank you very much, and Martin, your session at the workshop
> was really great!

Regards,
Terje
>
> Ben
>
> --
> Download LAStools at
> http://lastools.org/
> Visit the LAStools group at
> http://groups.google.com/group/lastools/
> Be social with LAStools at
> http://www.facebook.com/LAStools
> http://www.twitter.com/LAStools


--
- <Terje.M...@tmsw.no>
"almost all programming can be viewed as an exercise in caching"

[Antonio Ruiz]

Hi,

Ground classification takes into account the height of the points in a neighbourhood and the buffer should be enough for this neighbourhood to avoid artifacts at the edges of the tiles.

According the lasground README:

"The tool also produces excellent results for town or cities
  but buildings larger than the step size can be problematic.
  The default step size is 5 meters, which is good for forest
  or mountains. For towns or flat terrains '-town' the step
  size is increased to 10 meters. For cities or warehouses
  '-city' the step size is increased to 25 meters. For very
  large cities use '-metro' and the step size is increased
  to 50 meters You can also set it directly with '-step 35'."

The buffer should be at least the step size. If you have rivers or very large buildings, something larger would be better. I would set 50 or 100 m but I'm experienced with software from another company.  

Regards

2013/1/18 Ben Robson <benrob...@gmail.com>

Hi,

Thank you both for your help. Yes it is the DTM I am indeed after. I will have a play around with these suggestions then, I understand what a buffer command is but I don't get how it is advantageous in this situation? 

Thank you for your help!

Ben

On Thursday, January 17, 2013 10:56:34 AM UTC+1, Michel Wolters wrote:

Hi,

Just to avoid confusion, I personally use these abbreviations:

DEM: A digital elevation model (a general non-specific term)

DSM: A digital surface model, heights including objects on the surface. (A 3 meter tree on a 5 meter hill would then give 8 meters in the DSM).

DTM: A digital terrain model, heights of the terrain not including the objects on the surface. (The same situation would then give 5 meters in the DTM).

DOM: A digital object model, heights of only the objects on the surface. (the same situation would then give 3 meters in the DOM).

What I think you want after reading your e-mail is a DTM.

To get a DTM you need to use LASground, as LAScanopy produces forest metrics (see readme). Example commands would be:

lastile -i (folder)\*.las -merged -odir (output directory)\ -rescale 0.01 0.01 0.01 -olaz -o tile -tile_size 125 -buffer 30 -full_bb

lasground -i (folder)\*.laz -odir (output directory)\ -fine

blast2dem -i (folder)\*.laz -step 0.5 -o (output directory)\(filename).tif –keep_class 2 -merged

With LAStile, you determine the size of your tiles (if you have a large area, you may need to retile your data in order for your computer to process it more easily). You may also need to rescale your data (see the LAStile readme). For the best results, I advise you to make a buffer around your tiles (that's what the buffer 30 and -full_bb commands are for). You should determine yourself what buffer size would be most applicable for you (play around with parameters and look at results).

With LASground, you should use a filter setting which is applicable to your region (see LASground readme):


  For very steep hills you can intensify the search for initial
  ground points with '-fine' or '-extra_fine' and similarly for
  flat terrains you can simplify the search with '-coarse' or
  '-extra_coarse' but try the default setting first. 

With blast2dem you will need an appropriate step size (from blast2dem readme):

  The most important parameter '-step n' specifies the
  n x n area that of LiDAR points that are gridded on one raster
  (or pixel)

The setting -Keep_Class 2 keeps the LiDAR points classified as ground in the LASground application, -merged makes from all tiles one file. Leaving this out results in many tiles the size of the input LiDAR tiles.

All the parameters depend on what you want with the data and what is in your area.

Hope this helps,


Michel

On Thursday, January 17, 2013 10:23:28 AM UTC+1, Ben Robson wrote:

Hi,

I'm sure this question has been asked elsewhere, but I had a good luck and couldn't find it. I am not very familiar with LiDAR, I just know the basics, I will soon have some data from Eastern Norway and I'd like to create an elevation model stripping away the vegetation in order to see deformation to the ground caused by icebergs during a large flood thousands of years ago. I saw Martin demonstrating this at the LiDAR workshop in Salzburg at the end of last year, but when I looked on the lastools homepage I found both lascanopy and lasground, which is more suitable? And is there any sort of step by step for beginners? 

Anyway, thank you very much, and Martin, your session at the workshop was really great!

Ben

[Martin Isenburg]

Hello Ben,

glad to hear you enjoyed the workshop. It seems like Michel, Terje, and Antonio more or less answered your questions already. One alternative pipeline I'd like to point out (based on Michael example) is this one here where you use multiple cores (assuming here that you have 8) all the way until the final merge:

:: tile input into compressed 250 by 250 tiles with 25 meter buffer

lastile -i folder1\*.las -rescale 0.01 0.01 0.01 ^
        -tile_size 250 -buffer 25 ^
        -o folder2\tile -olaz

:: ground classify tiles with ultra fine option on 7 cores

lasground -i folder2\*.laz ^
          -ultra_fine ^
          -odir folder3\ -olaz ^
          -cores 7

:: rasterize ground points without edge effects on 7 cores

las2dem -i folder3\*.laz –keep_class 2 ^
        -extra_pass -step 0.5 -use_tile_bb ^
        -o folder4\ -obil ^
        -cores 7

:: use lasgrid for a final many BIL to single TIF raster merge

lasgrid -i folder3\*.bil -merged ^
        -step 0.5 ^
        -o dem.tif

All your tiles need to have less than 15-20 million points each or the process may run out of memory. Look into the new ./example_batch_scripts folder to get ideas and syntax for setting up your own LAStools processing pipeline.

Regards,

Martin @rapidlasso

--
http://rapidlasso.com - fast tools to catch reality