Computing for the Density

[Charmaine Cruz]

Hi Martin and everyone,

I would like to ask a question about lastools.

My task is to compute for the density of LiDAR points for every flight.

From my previous report, I used lasgrid to compute for the actual values of  density  (step is 5) and the output is number of points per 25 sqm, and it ranges from 1-255. It seems that I get higher value of density than expected.

What is the importance of considering the value of scan angle in computing for the density?

Thank you very much.. :)

--
Engr. Charmaine A. Cruz

Research Associate

Disaster Risk and Exposure Assessment for Mitigation (DREAM) Program

National Engineering Center

Alfredo Junio Hall

University of the Philippines Campus

Diliman, Quezon City 1110

+639159383637

[Martin Isenburg]

Hello Charm,

great question. Since you are expecting around 2 pulses per square meter and since you are using a step of 5 which equals 5m by 5m = 25 m^2 and since you are using the last only (or first only) retirn filter to count only one return per pulse you should expect to get values around 50 in the grid cells.

You will get - possibly much - smaller values whenever the strip does not fully cover the grid cell. It is possible that just one pulse hits one corner of a grid cell.

You will get - possibly much - larger values when the sampling on the ground gets condensed (but also streched out) because of strong pitch variations of a small aircraft in turbulences. But the more "systematic" larger values come from hardware that use a zig-zag scanner to scan. As the oszillating mirror slows down to change direction at the scan line edges you tend to get much much denser sampling there ... somewhat unrepresentative for the average sampling so it is important for zig zag scanners not to count the density of points along the edges when computing averages. One way of "not" counting them is to use a filter on the scan angle. For example

lasinfo -i sample.laz ^

           -compute_density ^

           -drop_abs_scan_angle_above 18

or

lasgrid -i sample.laz ^

           -density -step 5 ^

           -drop_scan_angle -20 20 ^

           -o density_25sqrmtr.asc

could do the trick.

Another issue ... if you density ranges up to 255 then higher values may be cut off and you should use '-density_16bit' instead of '-density' and use wider counters (16 bit instead of 8 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
 
 

[Charmaine Cruz]

Hi Martin,

Thank you very much for your response. :) 

Problem solved! I realized that I got higher density value for my last report because I did not filter points based on its scan angle. 

Thank you Martin,.. :)

[Martin Isenburg]

Hello,

a little follow-up to Charm's question from a few weeks ago. With LAStools you can easily create an intuitive visual illustration of the actual point density distribution in LiDAR flight lines (see attached example). First I counted the number of points per 25 square meters using lasgrid and output the result an BIL grid

lasgrid -i sample.laz ^
           -last_only ^
           -density_16bit -step 5 ^
           -o density_25.bil

I then visualized the BIL grid with lasview (yes, lasview - like most other LAStools - has read support for BIL and ASC) and played with the scaling and the coloring to produce the visualization (see attached).

lasview -i density_25.bil -color_by_elevation2

It is obvious that for this zigzag scanner the area along the edges of the scan line have a much much higher density than average. The single red point the attached visualization represents a particular 5 by 5 cell that had 388 points falling into it meaning a point density of over 15 point per square meter (i.e. 388 / 25). Most of the 25 square meter cells are blue and have between 45 and 50 points per square meter. Interesting are the subtle blue wobbles in flight direction that are caused by the changes in the airplane's pitch which compresses and stretches the scan line spacing.

The attached histogram (cut off at 127) shows the distribution of point counts per 25 square meter cell for the entire strip. The high values on the right have to do with the high zigzag density at the edges. But the low values on the of the peak worry me a bit. I was thinking that those may all be from the edge of the flight line where the 25 square meter cells have little to almost no overlap with the strip, but the visualization of 25 square meter cells that contain just a single last return (see last image) shows that the cells are all over the place. 

lasview -i density_25.bil -color_by_elevation2 -clip_z_above 1.1

Why? Water bodies? In any case, it seems that most cells receive around 48 points so I would say this flight strip delivers a pulse density of nearly 2 pulses per square meter.

Regards,

Martin @rapidlasso

--

http;//rapidlasso.com - fast tools to visualize LiDARs