PDAL exception: writers.las

[chort]

Hi Daniel!
First, thanks for this excellent piece of software, I've been using and abusing it a lot and it beats most bespoke SW in sheer functionality!

I've been trying to get my head around an error that I get while running ICP through command line using command (on v2.11.3 64-bit):
"c:\Program Files\CloudCompare\CloudCompare.exe" -AUTO_SAVE OFF -O -GLOBAL_SHIFT AUTO input_1.las -O -GLOBAL_SHIFT AUTO input_2.las -ICP -REFERENCE_IS_FIRST -OVERLAP 70 -C_EXPORT_FMT LAS -SAVE_CLOUDS FILE "out_1.las out_2.las"

The error I've been getting when trying to save the result is:
PDAL exception: writers.las: Unable to convert scaled value(-399999999998) to int32 for dimension 'Y' when writing LAS/LAZ file out_1.las

The funny thing is, input datasets on which this crashes are part of a series of several pairs of input datasets, all of which are geographically within few hundred meters, clipped from same inputs using same software (Safe software FME Workbench). Most work fine, but I have two pairs that throw this error.

I checked the LAS file metadata using PDAL info, but I cannot find anything terribly wrong with its offsets. All files, working and not working look very similar.

Do you maybe have any fresh insights?

Many thanks!

[daniel]

That's weird (and I would be interested to get the files if possible ;), but it's maybe due to the input 'LAS scale' that is not compliant with the new cloud coordinates after registration. As by default we re-use the 'LAS scale' from the input file...

[chort]

I've just posted two zip files on your admin e-mail address. Please let me know if you'd like some more, I've got plenty ;)

[daniel]

Thanks!

I've answered by email and fixed 1 bug and improved 1 thing! Hopefully it should solve all your issues ;)

[chort]

Wow, that was fast, thanks! When will the fixes be included in the latest Alpha, so I can give it a spin? ;)

[daniel]

Sorry, I forgot to say it: the new alpha version is online ;)