Warning of inconsistant echo range (0m)

[cgretler]

Dear Team of AMAPVox,
Thank you so much for this powerful tool. I have a question regarding a warning in some of my computation. I am working with AMAPVox version 2.0.0 and LAZ-files from ALS-acquisitions. During the voxelisation process, I often get the following warning of inconsistent echo ranges:

2023-06-23 07:53:54 WARN  - Inconsistant echo ranges (either interlaid blank echo or distances not monotonically increasing).
Shot 1633470, 4 echoes
  origin (2646227.8955, 1266448.9913, 1282.372) direction (-0.3215, -0.1541, -0.9343)
  Echo ranges 1018.7898m 1020.697m 0m 1024.7287m 
  Echo classes 3 3 0 2 
Shot discarded.

Depending on the file I am running, it is possible that very many shots are getting discarded due to this which is unfortunate. I got the same warning from files of different ALS-acquisitions from different sources. I tried to find out the reason for this, as I assumed a faulty LAZ-file. The echo that causes the problem is of class 0 (never classified) and its range is 0m. If I understood correctly, this would represent an echo directly at the sensor. In my case it would be directly where the airplane is located at that time.

I looked further into it and wasn’t able to find points of this class 0. Neither LAStools nor CloudCompare identified points of class 0 in these files. There were also no echoes visible that are far from the ground indicating to be close to the airplane altitude.
Have you ever encountered something similar, know where this echo could come from or how to fix this? Thank you very much for your help.

[philippeverley]

Hi there,

AMAPVox is based on shot geometry reconstruction: it follows a shot line across the voxel space and update voxel space state variables on the way. Interlaid blank echo might suggest something went wrong during acquisition and we chose to warn and stop the voxelization rather than to discard quietly faulty echoes.

LAStools or CloudCompare might not even load the 0-class points, it would explain why you do not see them with both tools ? Did you try with R package lidR ?

What I'd do:

• Open your LAS/LAZ with lidR;
• Discard every 0-class points;
• Re-order / re-number echo ranks (optional, AMAPVox will complain if echo ranks are missing, but you can disable the warning in the GUI);
• Write a corrected LAS/LAZ file.

[cgretler]

Hi @philippeverley,

thank you very much for your reply. I will try it with lidR and discard every 0-class point, if possible.

[cgretler]

Hi @philippeverley ,

I tried what you proposed. Unfortunately, I was not able to identify any points with class 0 or even points without any assigned classification. I tried LASTools, CloudCompare, lidR and laspy.

I clipped a LAZ file so I could upload it here. In this small test area, 5 shots are getting discarded due to the problem. Maybe you can find these class 0 points? As this is such a small area, only few shots are discarded, but the problem is still there. I am still wondering why and how AMAPVox finds these points and discards the shots.

I thank you in advance for your help.

test_area.zip
trajectory.csv

[philippeverley]

Thanks @cgretler for sharing your data. Surprisingly with AMAPVox 2.0.2-linux I gave it a try and did not get any of the warnings you mentioned in the first post. Here is the voxelized scene:

Logs and configuration file attached.
cgretler-log.txt
cgretler-config.txt (xml not supported by github, changed to txt)