51

Re: New version of FreeLSS committed to Github

Great software updates!  The new preview options allow me to use the Atlas 3D as a stand-alone...exactly what I was wanting.

This scanner just keeps getting better and better.

Thanks for all your hard work.

-Kevin

52

Re: New version of FreeLSS committed to Github

FreeLSS version 1.20 is now available via the ATLAS 3D auto-update system.  It is a minor release that allows for larger objects to be scanned than the default size of 8.5 inches (if the scanner is modified in a way to support larger objects).  This is controlled via a new option to the Setup page called Max Object Size.

- Uriah

53

Re: New version of FreeLSS committed to Github

OK, how do we modify it "in a way to support larger objects"? big_smile

54

Re: New version of FreeLSS committed to Github

Well. I am the reason for larger objects :-)

This is my "TinkerToy" scanner. Can be built using handtools.

http://soliforum.com/i/?6pgHleL.jpg

Scaled up from the Atlas-scanner. Since then I have moved the lasers closer to the camera. Plus added a gearbox between the stepper and the table.

/niels

55

Re: New version of FreeLSS committed to Github

Very nice!  Are you using the stock lasers?

56

Re: New version of FreeLSS committed to Github

The lasers are standard 12mm linelasers from ebay.

I have stopped further use and development on my TinkerToy scanner because the quality of the scans does not fullfill my requirements.

I have tried both the Frelss and the Fabscan software with same negative results.

In the future will use AgiSoft Photoscan. My very first test with photoscan based on 60 photos from different angles all around my object gave me an stl-file that could be sliced and printed without further processing other than aligning the object to the printbed.

So goodbye to laserscanning.

/Niels

57

Re: New version of FreeLSS committed to Github

FreeLSS 1.21 has been made available via the ATLAS 3D update system.  It removes the restriction on only using the first 8 pins of the Pi for motor and laser control.

- Uriah

58

Re: New version of FreeLSS committed to Github

FreeLSS 1.22 has been made available via the ATLAS 3D update system.  It resolves an issue some users were having where their WiFi access points would not be detected.

- Uriah

59 (edited by n2ri 2017-06-26 08:31:57)

Re: New version of FreeLSS committed to Github

link for software update? also will it effect me using the old B RP?

Solidoodle 2 with Deluxe kit cover & glass bed with heater. and 2nd board SD2 used not 3rd and alum platform not installed yet still wood. also need cooling fan installed to board. use Repetier Host couple vers. Slic3r also have all free ware STL programs

60

Re: New version of FreeLSS committed to Github

The scanner updates itself when you click Apply Update on the Check for Updates page.  The software will run on the old B RP as well.

- Uriah

61

Re: New version of FreeLSS committed to Github

FreeLSS version 1.23 is now available via the ATLAS 3D auto-update system. 

Checkerboard Calibration
The major change this release is support for checkerboard based automatic calibration.  This calibration method model handles many previously unhandled conditions including turntable center misalignment, turntable plane correction, the camera being positioned off center, the camera not being positioned completely horizontally, and much better calibration of the lasers.  If the lasers are pointed even remotely close to the center of the turntable,  the checkerboard calibration will do the rest.  This should resolve the "double shell" issue some users were still experiencing.

The algorithm uses two printed checkerboards of differing sizes, one for the front of the calibration item and one for the back.  The PDF file for the calibration pattern can be found at https://s3.amazonaws.com/atlas3d/calibrationPattern.pdf The 3D printable model for the calibration pattern is at https://s3.amazonaws.com/atlas3d/checke … n_item.stl.  The PDF is also compatible with the calibration item that ships with the acrylic frames. These files are also available on Thingiverse https://www.thingiverse.com/thing:1280901 and linked on the Murobo printables page http://www.murobo.com/printables.html.

The manual calibration process has also been updated with integrated diagrams and a step by step procedure.

Please note that checkerboard calibration works on the Pi 1, but is much faster on the Pi 2 and Pi 3.

Start the process by clicking Calibrate like before.
http://soliforum.com/i/?3LIh1fq.png
A new descriptive page explains how to position the calibration item and where to retrieve the necessary files.
http://soliforum.com/i/?fogM42m.png
The progress page.  The updates are currently pretty coarse but will let you know when the process is complete.
http://soliforum.com/i/?wPpywFK.png
After a successful calibration.
http://soliforum.com/i/?Fgxcvu4.png
The Setup page displays the calibration information.  You can also access Manual Calibration from here.
http://soliforum.com/i/?bQBTSwy.png
The manual calibration process has also been updated.
http://soliforum.com/i/?L1qDg9D.png


Separation of Right and Left Results
The results of scans from the left and right lasers are now left in their respective files instead of merging them into one.  This functionality may be integrated again at a later date but right now programs such as MeshLab do a much better job merging the scans than the Pi is currently capable of.

62

Re: New version of FreeLSS committed to Github

Looking forward to giving this a try tomorrow. I have really been struggling with alignment. The separation of the left and right lasers is a great idea. Such an inconvenience to run 2 scans, one for each laser. It will be great getting both files in one pass to merge in meshlab. On a side note... Kscan 3d has some pretty cool auto alignment features but does not support any of the native atlas 3d output files. I been running them through meshlab first then export .obj files for kscan to align for me. Any chance of a future update to Atlas 3d outputting .obj file directly?

63

Re: New version of FreeLSS committed to Github

Couple of quick questions...

I'm not trying to be critical, I'm just trying to understand how this works. Dumb questions are the questions not asked right?

• If you print the PDF directly without processing, the individual front and back panels end up 96 x 84 mm.
• The 3d printable model for the calibration pattern is 101.6 x 89.28 mm.
• In the pictures 1 and 3, they appear to be the same size.

A) Should the paper print be bigger?
B) Should the 3d model be smaller?
C) Does the size of the pattern matter for setting the calibration in atlas and does it affect the scale of scanned objects?
D) Should we just set the smaller paper pattern to the top or bottom of the 3d printed pattern holder?
E) Should it be centered on it?
F)How will plastic shown around the edge of the pattern affect the calibration process if it's black plastic? White plastic?

Thanks for the update and any clarification you can provide on these details.

64

Re: New version of FreeLSS committed to Github

Hi Brink,

I just printed the calibration PDF at https://s3.amazonaws.com/atlas3d/calibrationPattern.pdf and the measured dimensions matched what you listed for the 3D printable model.  Can you double check your printer settings? 

A,B) The paper should be the same size as the 3D printed model surface.
C) The size of the squares in the calibration pattern do matter and must be correct.
D) The pattern listed as front should be placed on one side of the holder and the pattern listed as back should be placed on the other.
E) The edges of the holder should match up with to the edges on the checkerboard.
F) The pattern is sized to completely fill the holder but if you are referring to the sides of the holder, that shouldn't matter.

Regards,
Uriah

65 (edited by tonyno 2017-07-28 00:55:08)

Re: New version of FreeLSS committed to Github

Could not detect front chessboard during calibration.

Add added lighting with no change.

What else can I check?

Edit: OK, I got it to work. Forgot about setting up a background. Moving on...

66

Re: New version of FreeLSS committed to Github

hairu526 wrote:

FreeLSS version 1.23 is now available via the ATLAS 3D auto-update system. 


Separation of Right and Left Results
The results of scans from the left and right lasers are now left in their respective files instead of merging them into one.  This functionality may be integrated again at a later date but right now programs such as MeshLab do a much better job merging the scans than the Pi is currently capable of.

So do I still have the option of making a single file from both lasers?  I have no problem with laser alignment, and like using both lasers in order to capture more detail (inside curves, etc.).  The new alignment procedure looks great!

-Kevin

67 (edited by hairu526 2017-07-28 00:45:57)

Re: New version of FreeLSS committed to Github

@tonyno
Can you post an screenshot of the CAMERA page with the checkerboard on it?

68

Re: New version of FreeLSS committed to Github

@knowack You can still make a single scan from both lasers.  The main difference is that the merging would take place in another application such as MeshLab instead of directly on the ATLAS 3D.

69

Re: New version of FreeLSS committed to Github

Here's the calibration info that was generated...

Calibrated the turntable and right laser and left laser successfully.
Table Center: (0.005882,-3.197,-11.67) in.
Table Normal: (-0.01168,0.9998,-0.01299)
Right Laser Location: (458.9,0,0) in.
Right Laser Normal: (-0.02096,0.0363,0.9991)
Left Laser Location: (-0.8393,0,0) in.
Left Laser Normal: (0.9995,0.003165,0.03152)

70

Re: New version of FreeLSS committed to Github

@tonyno
The table looks correct but the lasers are way off.  Make sure that when the lasers are turned on, they hit the checkboard when it's in the front and the back.

71 (edited by tonyno 2017-07-28 02:01:40)

Re: New version of FreeLSS committed to Github

Here's a shot from the preview with both lasers on.

http://soliforum.com/i/?SDa48Xq.jpg

This is with the cal item centered, not 1" forward, just to show laser alignment.

What do you suggest to do next?

72

Re: New version of FreeLSS committed to Github

The lasers look like they should be fine.  Are you using the 5MP camera or the 8MP?  Also, which Pi version are you using?  Lastly, place the calibration item on the table liner so that it doesn't move during the calibration.

73

Re: New version of FreeLSS committed to Github

All parts from original, Atlas release. So, 5MP cam and Pi 1.

The liner was there for calibration.

74

Re: New version of FreeLSS committed to Github

Please do the following.

Turn on both lasers.
Place the calibration item on the table 1" from center.
Ensure that both lasers hit the front of the calibration item.
Enter 180 in the rotation box and click the Rotate button.
Ensure that both lasers hit the back of the calibration item.

If possible, please send me a screenshot of the entire browser window when the calibration item before and after the 180 degree rotation with the lasers on.

If so, please rotate 180 again so it's in the original position, turn off the lasers, and perform calibration.

Thanks,
Uriah

75

Re: New version of FreeLSS committed to Github

http://soliforum.com/i/?Xi5uNKc.jpg

http://soliforum.com/i/?hbpN9QX.jpg

http://soliforum.com/i/?RLvAh22.jpg

Calibrated the turntable and right laser and left laser successfully.
Table Center: (-0.02209,-3.166,-11.37) in.
Table Normal: (-0.006555,1,0.006261)
Right Laser Location: (5.014,0,0) in.
Right Laser Normal: (-0.9056,-0.01441,0.4238)
Left Laser Location: (-4.899,0,0) in.
Left Laser Normal: (0.9103,0.003853,0.4139)