Looking for advice: removing large low/high “patches” before ground classification in LAStools

[Róbert Bors]

Hello everyone!

I am struggling with a point cloud problem that I cannot solve on my own, so I would be very grateful if you could give me some advice.

The data collection did not take place under ideal conditions (tight deadlines, no possibility of re-flying), and unfortunately there was fog during the LiDAR survey. As a result, a large number of points ended up below and above the actual surface, forming continuous patches rather than isolated noisy points. (Screenshots attached.)

These patches are the main problem:

• they are not just single outliers, but larger "layers" of points below and above the road corridor.
• Lasnoise cannot simply classify them as noise because they are locally dense and interconnected.
• When I run the lasground_new program, it still uses some of these patches as ground points, which makes it impossible to build a reasonable surface (DTM/DSM) afterwards.

I have tried several combinations of the tiling, lasground_new (including wilderness and extra_fine), lasheight, lasnoise, and lasthin commands, and I have also experimented with LASlayers, but I have not yet found a reliable method to remove these spots before starting the final ground classification.

What I'm looking for:

• A workflow in LAStools that automatically removes or reclassifies these low/high spots,
• ideally something scriptable (batch / LASlook command line),
• and not based on manual selection and deletion, because the corridor is long and the number of spots is large.

Does anyone have a recommended method (or example command sequence) for handling this type of "fog noise" / large low and high spots in LiDAR data before lasground_new?

Thank you in advance for any ideas and experiences you share with me.

Best regards

Robert

[Karl]

Hello Robert!

Out of curiosity because I used to work on the impact of fog on LiDAR sensors, have you looked at the metadata information of these specific points ? 

I am thinking of intensity or multi-echo information, maybe it could lead to an appropriate filtering method.

Can you provide a pointcloud sample ?

Karl

[Róbert Bors]

Hello Karl!

I haven't gone into that much detail, but I'm happy to send you the sample that I'm stuck on.

You can download it from this link.

https://GLYR.quickconnect.to/d/s/15taLhMjbsY3OuebDBWh4rmDkO4ctYyp/hLxU-lc72Z0wJnHG8WFOqXY3H-bnjdQ9-HbvA_qBgwQw

I would appreciate any help you can give me, because I'm really stuck here and it's important that I be able to automate the process.

Thank you in advance for your help!

Best regards,

Robert

[GeoViz]

Robert, I deal with datasets like this from time to time. Do you have the trajectory data available? 

Matt

--
Download LAStools at
https://rapidlasso.de
Manage your settings at
https://groups.google.com/g/lastools/membership
---
You received this message because you are subscribed to the Google Groups "LAStools - efficient tools for LiDAR processing" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lastools+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/lastools/07bfb4dc-f286-48ac-9cea-f1a422bebe8an%40googlegroups.com.

[Karl]

Hi again,

I have looked at the data and I can see what you mean with the points below the ground.

This can indeed originate from multi-scattering effects in fog conditions.

These points are very low in terms of z values, more than 100m below the normal ground points from what I can see.

My advice for a quick solution would be to develop to filter based on the z-values.

You can plot a histogram of z values to see the spread of the data, it should be quite peaked around 200.

And, remove all points which have a z value too far from the mean.

Concerning points above the actual surface, I am not seeing any.

Good luck :)

Karl