CloudCompare forum

When I transformed a scan using a 4x4 matrix, it would be in the right place until I saved it as an las file. Then when I brought the las back in to CC, it was offset by more than a km. Turns out the problem seems to have been caused when I save the cloud to an LAS file using "optimal resolution", if I save as LAS using "custom resolution" the LAS file is in the right place.

Is this a bug or am I not using CC correctly?

That doesn't look normal indeed... Which version have you used?

And can you maybe share with us the file and the matrix so that we can try to reproduce it on our side? (admin[at]cloudcompare.org). Or at least give us more information (such as the original global shift vector (if any), the matrix you apply, the center and dimensions of the bounding-box, etc.)

Daniel,

The transformation matrix is:

0.98461723 0.17201975 -0.00005255 -457.39103595
-0.17201975 0.98461723 -0.00002421 360.43177936
0.00004760 0.00003289 0.99953083 -480.00684696
0.00000000 0.00000000 0.00000000 1.00000000

Local Global Shift -0.000 -0.000 -0.000
World Global Shift -288000.000 -170000.000 -0.000

The problem doesn't seem to be related to the data, matrices, or shifts. It only happens when I save an las file using 'optimal' selection.

I'm running 2.11.3

Oh, have you tried 2.12? Because several things have been fixed in this new version regarding Global Shift.

I tried to simply change the LAS resolution, and couldn't reproduce the issue (at least with 2.12). And I tried the 2 LAS filters available on Windows.