That is definitely having an effect.  If I reduce the lighting to 58 lux, I get laser distances of approximately 7" and -8". At 47 lux, I get 5" and -6".  Obviously, this results in each laser's output being skewed and slanted when I scan an object.  If I manually calibrate the lasers to 5.25", I get closer, and the geometry is more upright.

I've tried both the original 5MP, and the 8MP cameras.  The only other thing I've done is the firmware update to allow the 8MP camera to operate.

Previously, my Atlas3D had been performing great.  I cannot imagine what I've done.  I was trying to upgrade so that I could better scan 15mm wargame figures; I should have just left well enough alone!

-Kevin

p.s. I just tried 29lux, and the distances were 7 and -8...it seems to just be jumping around.

- Uriah

Here's the corresponding .ply file:

https://www.dropbox.com/s/ogc67jn6krfp3 … 0.ply?dl=1

I disabled noise removal and set the threshold to low.  I didn't move or change anything else.  Here's what I got for calibration info:

Calibrated the turntable and right laser and left laser successfully.
Table Center: (-0.06807,-2.937,-10.34) in.
Table Normal: (0.00426,0.9999,0.01188)
Right Laser Location: (314.6,0,0) in.
Right Laser Normal: (-0.03447,-0.02405,0.9991)
Left Laser Location: (-1.256,0,0) in.
Left Laser Normal: (0.7964,-0.562,0.2234)

-Kevin

- Uriah

knowack, can you post the settings for the preset named "Calibration"?  Also, if possible can you include the log messages that are printed during the calibration step along with the resultant calibration file that shows up alongside the scan files?

Thanks,
Uriah

Uriah,

Sorry for the delay; I've been bouncing back-and-forth between software versions to see if I could figure this out myself.  I have reinstalled the 5MP camera, since it works with both versions, but I plan on going back to the 8MP camera once I get 1.26 working correctly.  On the image of FreeLSS 1.26 that I'm using, the RPi is fully updated, except that the Kernel version is 4.4.50.  I'm sure I'm doing something wrong, but cannot figure out what.

Here's a screengrab of the settings for the Calibration preset:

To my recollection, I haven't changed them from 'factory settings'.

Here's the log of a calibration attempt with a white background (it cannot find the chessboard against a black background):

Calibrated the turntable and right laser and left laser successfully.
boardPlaneCentroid ==> (-0.0120425,-1.13339,-9.73539) in.
boardPlaneNormal ==> (-0.0245098,-0.0209776,0.999479).
63 front right laser points are within range.
94 front left laser points are within range.
boardPlaneCentroid Rotated ==> (-0.0339888,-0.840125,-10.8671) in.
boardPlaneNormal Rotated ==> (-0.0351296,-0.0277334,0.998998).
72 total right laser points are within range.
133 total left laser points are within range.

Table Center: (-0.05861,-2.936,-10.36) in.
Table Normal: (0.003372,0.9999,0.01348)
Right Laser Location: (12.44,0,0) in.
Right Laser Normal: (-0.6266,-0.01651,0.7791)
Left Laser Location: (-19.57,0,0) in.
Left Laser Normal: (0.4411,-0.2279,0.8681)

The white background for this calibration was about 4 inches behind the turntable, with resultant laser distances of 12 and -19 inches (if I'm reading the log correctly).  If I run the calibration repeatedly, without moving anything, I get laser positions of 7 & -3, 20 & -23... When I put the scanner in another location where the white background is about 10 inches behind, it will think the lasers' distances are over 100 inches.  It would seem that the Atlas3D is picking up the laser strike on the background.  I cannot seem to find a suitable background that will allow the Atlas3D to detect the chessboard, AND ignore the laser strike on the background.

Here's some images to show the hardware alignment; I don't know if they're helpful:

Here's the camera aligned to the turntable:

Here are the lasers oriented on the center of the turntable (my background is not squared up, so the lasers look crooked in the vertical; they are not):

Here's the calibration target, sitting over the center of the turntable...just demonstrating laser alignment in the vertical:

I have a monitor and keyboard connected to the Atlas3D.  For actual calibration, I set the target approximately 1 inch in front of the turntable's center, per the instructions.  When trying to calibrate with the black background, the command line says, "Chessboard not found.  Done."  This is the error message I get through the browser interface:

The basic hardware alignment I've shown is what has previously given me great results.  As posted in other threads, I've been able to get good details scanning Matchbox cars, and was working to improve my results from scanning 15mm soldier figures. 

Any insight would be appreciated!

Thanks!

-Kevin

Thanks,
Uriah

...still working on it.

I've tried to manually enter the numbers I used in V1.22, but the results are very poor.

I should mention that this is a full-version of the software, Rpi2, laser-cut acrylic case.

If I revert to v1.22 and manually align the lasers, everything works fine.

I've only had the chessboard successfully detected a few times.  When I do manage to pass that hurdle, the resultant calibration dimensions are badly out of wack...laser distances greater than 100 inches.

The RaspberryPi is fully updated, and I'm using the 8mp camera.  I've tried different backgrounds; here's my current environment:

With this visual, the Atlas3D still cannot detect the chessboard.