Inaccurate Point Cloud of Road - 0.5m Thick Data

[philip]

Hi

We recently conducted a fixed wing UAV survey in the UK using RTK. 2cm GSD.

The resultant point cloud produced was 0.5m thick for the road areas, rather than being a nice thin sharp line of data. This can be seen when viewing a profile from kerb to kerb across a section of the road. 

Is there any way to take the las file and 'sharpen' it? Attached screenshot of the 'mogul' effect.

Thanks!

[philip]

Heres the LAZ file: https://drive.google.com/file/d/177vWFpxiq4h-jiTqxHIY6kaJZDQEj33G/view?usp=sharing

[Sam Hackett]

Hi Philip,

Your point cloud data is derived from photogrammetry, you probably are already aware or this, just noting for other readers as UAV doesn't specify photogrammetry.

The phenomenon you are seeing is common for uniform surfaces with a low amount texture.  You will see the surfaces improve where there are features such as road markings which provide 'texture' for the photogrammetry.  While we think of a bitumen road as having ample texture, this is only helpful if the photogrammetry has adequate GSD to uniquely identify parts of the texture.  Without adequate GSD the software gets confused with which pixel groups to match where and consequently gets it wrong a significant amount of the time.

If you want a survey surface product for the roads I would suggest manually extracting the road markings as 3D lines and then excluding the road point cloud from the surface.  I would not trust the accuracy of the surfaces to be survey grade without ground checks spread over the survey extent.

The phenomenon is visible on some other surfaces also like some of the roofs.

In future if you want higher quality road surfaces you may need to prescribe this in the scope.  There are techniques to help this; including obliques and reducing GSD are two things that will help. 

You may find the processing software can produce a DSM point cloud derivative which may be cleaner than the densified point cloud, however a 'clean' surface does not guarantee it to be correct, just easier to work with.

You guys must have some different Airways rules than down here in New Zealand.

Merry Christmas

Sam

[philip]

Hi Sam

Yea I think you are right - the issue arose when the client requested to see the raw point cloud and then compare the data with a laser scan of the same area...

Thanks and Merry Xmas to you as well.

Regards, Phil

[Martin Isenburg]

Hello,

using the (attached) variety of the batch script from a recent blog post (*) and without spending much time on tweaking parameters I was able to generate this (also attached) hillshaded 25 cm DTM from your photogrammetric point cloud with not too many 'mogul' effects left. Further tweaking (and maybe adding another step to compute a "mean" ground" surface could further improve results.

(*) https://rapidlasso.com/2018/12/27/scripting-lastools-to-create-a-clean-dtm-from-noisy-photogrammetric-point-cloud/

Regards,

Martin @rapidlasso

--
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