What's the problem with this LAS (or with me using las2dem :-) )

[Michez Adrien]

Hi everyone,

We are working on a LiDAR dataset ( > 50 pts / m²) and we got some strange results using lastools ... And got 'regular' results using FUSION.

We intended to compute CHM but the problem is arising even when running a simple las2dem.

As you can see with attached images, the DEM computed with lastools by running a simple las2dem (las2dem -i "original_LAS_file.las" -elevation -oasc) is really not good-looking when compared to the one we got with FUSION. You can find the original las file HERE and the DEM generated with las2dem and fusion HERE

Has someone an idea ? What do we miss ?

Thanks in advance for your help !,

Adrien

[Martin Isenburg]

Hello Adrien,

you make no friends on this mailing list if you provide a link to 95.9 MB uncompressed LAS files instead of a link to a 15.5 MB compressed LAZ. 

(-: 

Downloading your LAS sample took about 18 minutes via my Internet link here in the Philippines whereas I would have had the file (including integrity check) after 3 or 4 minutes if it was LASzipped ... )-:

100,588,185 original_LAS_file_536d585c127e1b0f4351_.las
16,243,933 original_LAS_file_536d585c127e1b0f4351_.laz
   2 File(s)    116,832,118 bytes

You can see a similar (albeit less pronounced) effect between the las2dem (via TIN interpolation) and lasgrid (highest point per raster) here:

http://rapidlasso.com/2014/11/04/rasterizing-perfect-canopy-height-models-from-lidar/

What you are seeing are "pits" that come from TIN interpolation. But you seem to make things worse by using all points instead of only the first returns. So for comparison first I would try:

las2dem -i original_LAS_file.laz -first_only -step 1 -odix _tin_d00 -obil

Given you have 50 pulses/square meters I you can use the much simpler which is probably what FUSION is using:

lasgrid -i original_LAS_file.laz -highest -step 1 -odix _grd_d00 -obil

Maybe try to replace each return with a 5 / 10 / 15 or 20 centimeter disk:

lasgrid -i original_LAS_file.laz -subcircle 0.025 -highest -step 1 -odix _grd_d05 -obil
lasgrid -i original_LAS_file.laz -subcircle 0.050 -highest -step 1 -odix _grd_d10 -obil
lasgrid -i original_LAS_file.laz -subcircle 0.075 -highest -step 1 -odix _grd_d15 -obil
lasgrid -i original_LAS_file.laz -subcircle 0.010 -highest -step 1 -odix _grd_d20 -obil

It seems you should be able to do much much finer CHMs for single tree detection given your high pulse density (or low pulse spacing) as can be seen in the lasinfo report below. So as you drop your '-step 1' down to '-step 0.5', '-step 0.2', or even '-step 0.1' you will definitely want to switch to the pit-free algorithm by (Khosravipour et. al, 2014) ...

Martin @rapidlasso

PS: I also noticed that you use inefficient ASC rasters. Switch to BIL, TIF, or IMG and read/write all your raster products up to ten times faster.

D:\LAStools\bin>lasinfo original_LAS_file_536d585c127e1b0f4351_.las -cd

lasinfo for original_LAS_file_536d585c127e1b0f4351_.las

reporting all LAS header entries:

  file signature:             'LASF'

  file source ID:             0

  global_encoding:            1

  project ID GUID data 1-4:   00000000-0000-0000-0000-000000000000

  version major.minor:        1.2

  system identifier:          ''

  generating software:        'TerraScan'

  file creation day/year:     31/2014

  header size:                227

  offset to point data:       229

  number var. length records: 0

  point data format:          1

  point data record length:   28

  number of point records:    3592427

  number of points by return: 1766837 930247 547829 249898 78195

  scale factor x y z:         0.01 0.01 0.01

  offset x y z:               0 0 0

  min x y z:                  889000.00 6388400.56 1239.98

  max x y z:                  889499.99 6388499.99 1545.97

the header is followed by 2 user-defined bytes

reporting minimum and maximum for all LAS point record entries ...

  X            88900000   88949999

  Y           638840056  638849999

  Z              123998     154597

  intensity           3       1683

  return_number       1          7

  number_of_returns   1          7

  edge_of_flight_line 0          1

  scan_direction_flag 1          1

  classification      2          4

  scan_angle_rank   -37         29

  user_data           0          0

  point_source_ID     3          7

  gps_time 71229684.436821 71230700.197493

number of last returns: 1769966

covered area in square units/kilounits: 33064/0.03

point density: all returns 108.65 last only 53.53 (per square units)

      spacing: all returns 0.10 last only 0.14 (in units)

WARNING: there are 16727 points with return number 6

WARNING: there are 2694 points with return number 7

overview over number of returns of given pulse: 839623 764007 892787 686373 306956 85279 17402

histogram of classification of points:

          672606  ground (2)

         2919821  medium vegetation (4)

> --
> 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

[David Herries]

Adrien

 

Do you think what FUSION is actually giving you is right?  You wanted a DEM right not a DSM or CHM yet?

 

Always great to view outputs in 3D to see what you’ve asked it to do, it actually did!   (PS.  Also looks like Martin has just replied to give you the detail, so here’s just a summary in pictures)

 

Your DEM_FUSION.asc

 

Your DEM_las2dem.asc

 

Both of these have used all the points in the LAS file as you have not asked to use the ground and model point only (Class 2 and 8)

 

Using LASTOOLS with basic default settings, but only using ground and model key points.

las2dem -i "original_LAS_file_536d585c127e1b0f4351_.las" -keep_classification 2 -keep_classification 8 -elevation –oasc

 

so I’m guessing this is what you wanted.

 

 

And now what you are looking at for your CHM is to put the vegetation back on there again..

 

 

Have fun

 

David Herries     Interpine Group Ltd

Mobile:      021 43 5623   DDI:  +64 7 350 3209 or Australia 0280113645 ext 721

--
Download LAStools at

[Michez Adrien]

Hi everybody,

Thanks for the informations ! It's helping ... But not (yet) working (properly) :-)

I've updated LASTools today and tried the pit free  chm pipeline from arcgis toolbox... I still got some unwanted result, CHMs look smoothed.
But I got something better using simple lasgrid on max elevation at 0.5m

You can find the original LAZ file HERE and the normalized point cloud HERE.

Below, this is what I had using the pit free CHM pipeline with a step of 0.5m and and beam width of 0.2m.
I tried the same, changing the step (0.25m) and the beam width (0.1m). Still looking "smoothed"

0.5m
    
0.25m


Below, see what I got using lasgrid  -i normalized.laz -step 0.5 -elevation -highest -odir -o lasgrid_0.5.tif



Not too bad but we hope to do better with this very dense point cloud ! I tried the step 0.25m :



The result of lasgrid at step 0.5m looks OK but we should work with finer steps. What's the problem here ? How can we have these stripes with a such a dense point cloud  (50 pts / m²) ?

Thanks for your help !

Adrien

[Mike Windham]

HI Michez,

Neat dense data you have here. I took the liberty of publishing this from our new web browser solution.   In the link below, it's interesting how we can see the different species of trees....if you open the setting panel in the top right you can change to color by intensity and even more pronounced are the differences.  There are a couple screen shoots attached, for anyone with out a google chrome or firefox browser.    

https://youcontrol.newspin.com/pointcloud/567/show#/567

Mike Windham | Inventor, US Marine & CEO | New Spin

755 Research Pwy. Ste. 540 Oklahoma City, OK 73104

p. 405-200-1880 Ext. 11  f. 800-360-6949
e. mi...@newspin.com  w. newspin.com

---

The information contained in this transmission may contain privileged and confidential information, including client information protected by federal and state privacy laws. It is intended only for the use of the person(s) named above. If you are not the intended recipient, you are hereby notified that any review, dissemination, distribution, or duplication of this communication is strictly prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.

---

[Michez Adrien]

Hi everyone,

I didn't get answer so I try again :-)

Sorry to insist but do you guys think it's logical to have CHM like this with such a dense point cloud (>50 pts / m²) ?

See how smoothed are the CHM produced with LASTools arcgis pit free chm pipeline :

50cm (bean midth 0.2m)                                                                                                                25cm (beam width 0.1m)
    

We have a lot of pits when when run lasgrid on a normalized point cloud at step below 50 cm (lasgrid  -i normalized.laz -step 0.5 -elevation -highest -odir -o lasgrid_0.5.tif )

step 50cm                                                                                                                                               step 25cm
    


Are we wrong when thinking we could reach better results with our data ? You can find the original LAZ file HERE and the normalized point cloud HERE. You can also vizualize it HERE (thanks Mike Windham ! )

Regards,

Adrien

[Martin Isenburg]

Hello Adrian,

not sure I understand what your issue is with the CHMs you create. Attached a script that creates many different ones.

Is none of them sufficiently "good" for your purposes? Why do you find the pit-free CHMs too smooth? Maybe use smaller circles. Even without circles (just using points) the pit-free is a clear winner, no (see attached, first three at 50cm, second three at 25cm resolution)?

How low was your plane flying? How wide was the laser beam? I would use circles that are at smaller than the laser beam width. Maybe half the width ...

Rename the script from many_different_chms.but to many_different_chms.bat. I cannot email a *.bat file.

Martin @rapidlasso

[Michez Adrien]

Hi,

Thanks for this answer. I think we have rather the same results but I don't think the CHM from the pit free algorithm is a clear winner.

If you look at the higher trees in red, you'll see that the crowns are less visible in the pfa CHM ( STEP50_liege_norm_pfa_d00.png VS STEP50_liege_norm_grd_d00.png): the lower parts of the crown seem to have been smoothed. For me, the gridded approach presents the best contrast ...

Is it just me ? Am I looking in a wrong way ?

I've read (from you) that the pit free algorithm allows better tree crown delineation and canopy maxima extraction but "visually", I'm not sure about it. Later, I'll try to test it with a canopy maxima detection algorithm... I guess somebody in the list has already performed such tests ?

Regards,

Adrien

[Lavoie, David]

Hello all,

 

I have been looking at and working through the CHM steps that have been posted on here lately.  I get the results that others have seen but have a few issues/questions….all the way through, one outputs .bil formats.  These files have valid heights for the above ground points.  In the final step, and in all examples of final output I have seen, the output is in .png format with values no longer in height above ground but as RGB values.  Although a nice visual representation of the CHM what steps does one need to take in order to output a grid with actual height values.  I have tried changing the format in the final merging step to other grid formats but end up with very bizarre results….0 heights registering 255 and other values ranging upwards from there to values in the 10’s of millions….as if all values from each pixel in all the input grids are being multiplied with one another.  Any ideas for me?

 

Safety Starts with Me: Live It 3-6-5

_________________________________
David Lavoie
Data Processing Manager
Airborne Imaging, A Clean Harbors Company

_________________________________

Mike Windham | Inventor, US Marine & CEO | New Spin

 

--

[Martin Isenburg]

Hello,

to get a height raster with actual geight values (and not a mapping to '-false' colors) you need to modify the last call to lasgrid as follows. Instead of

lasgrid -i %NEW_OUTPUT_FOLDER%\tmp_dir\*.bil -merged ^

        -step %STEP% ^

        -highest ^

        -false -set_min_max %MIN_MAX% ^

        -ll %LOWER_LEFT% -ncols %NCOLS% -nrows %NROWS% ^

        -odir %NEW_OUTPUT_FOLDER% -o %NORMALIZED_INPUT%.png -ocut 4 -odix "_pfa_d20" -opng

use

lasgrid -i %NEW_OUTPUT_FOLDER%\tmp_dir\*.bil -merged ^

        -step %STEP% ^

        -highest ^

        -ll %LOWER_LEFT% -ncols %NCOLS% -nrows %NROWS% ^

        -odir %NEW_OUTPUT_FOLDER% -o %NORMALIZED_INPUT%.bil -ocut 4 -odix "_pfa_d20" -obil

This is in respect to the script that I have attached once more to this email for reference. It is called "many_different_chms.bat.txt" and you should rename it to "many_different_chms.bat" in order to run it. What essentially needs to happen is to remove the '-false' from the command line (this is mapping from heights to false colors) and to change to another export raster format (here from PNG to BIL).

Regards,

Martin @rapidlasso