1 Topic by adaptedesigns (edited by adaptedesigns 2015-06-11 21:44:25)

  • adaptedesigns
  • Member
  • Offline
  • Registered: 2015-05-15
  • Posts: 15

Topic: Scan and print

Hi,

Using the 1.6 version of freelss and after a long calibration process, I can get this kind of results.
Parameters : 5M pixels still mode, 1200 frames, 6400 steps for revolution (my stepper is a 0.9°)
For the calibration I use, as Pirvan, a checkerboard pattern to be sure that lasers were perfectly aligned with the camera red pattern. Also use a personnal calibrator adapted from the model designed for the Atlas scanner.
Then I scan 50% of a tube and check under meslabs the diameter and adjust precisly, for each laser, the distance from camera center : a long but, in my opinion, imperative process.

Original :
http://soliforum.com/i/?7YcI3Z1.jpg

Freelss preview :
http://soliforum.com/i/?SAKO9PB.jpg

Meshlabs brut ply file
http://soliforum.com/i/?Rgz5YRE.jpg


Poisson reconstruction :
Process : poisson-disk sampling 250000/compute normals for point set 1000/surface reconstruction : poisson 12-7-1-1.
http://soliforum.com/i/?gKMyIZc.jpg


Netfabb basic cleaning and repair stl file :
http://soliforum.com/i/?qKx0h2e.jpg


Then printing (Support and raft ON, 0,20 layers, Infill 0%) :
http://soliforum.com/i/?Z9L0VOj.jpg
http://soliforum.com/i/?i1wWRWe.jpg
http://soliforum.com/i/?3uMFuKi.jpg
http://soliforum.com/i/?YHuEzUO.jpg
http://soliforum.com/i/?WMQSBnU.jpg
http://soliforum.com/i/?SeLF9ZA.jpg

Copy VS original :
http://soliforum.com/i/?MZcK5Um.jpg
http://soliforum.com/i/?uxtCHZQ.jpg

Loosing some details (eyes and lips for example) but I must work harder to command Meshlabs.

Freelss 1.7
Raspberry Pi 2
Piclop personnal design (check Website)
Direct RJ45 connection

adaptedesigns's Website

2 Reply by pirvan

  • pirvan
  • pirvan
  • Member
  • Offline
  • From: San Francisco
  • Registered: 2013-03-27
  • Posts: 1,797

Re: Scan and print

Very nice looking piece.

You mentioned that you spent a lot of time calibrating the lasers distance from the camera.  Obviously there isn't a heck of lot you can do physically, so what exactly did you do to calibrate them?  Did you input the actual numbers, or did you end up using artificial number by doing trial and error scans?

Obviously you've read my posts smile, so you know the problems I'm having with the left and right lasers not agreeing with each other.  Did you have similar problems?

The problem is more obvious when you have something that is not primarily cylindrical, but has limbs or something like it protruding outward.

BTW, do you mind sharing the laser distance calibration numbers?  Are they the same (left and right)?

One last thing, I noticed some imperfections on the lower left cheek of the model.  Have yo use the Z brush inside MeshLab?  You can use it to manually smooth areas that need some work.

To print or, 3D print, that is the question...
SD3 printer w/too many mods,  Printrbot Simple Maker Ed.,  FormLabs Form 1+
AnyCubic Photon, Shining 3D EinScan-S & Atlas 3D scanners...
...and too much time on my hands.

3 Reply by adaptedesigns

  • adaptedesigns
  • Member
  • Offline
  • Registered: 2015-05-15
  • Posts: 15

Re: Scan and print

Thanks Pirvan,

I cannot use the numbers provided because my scanner is a personnal developpement (check my website). They are specific to my design.
Yes, to calibrate, as I say, I've scanned a tube and check, for each laser, the size of the point cloud with Meshlabs and adjust the numbers to get the same size as measured for the real tube. This method allow me to find the exact numbers for each laser.

As you notice, I get the same problems with not cylindrical objects.

Thank you for this tips in Meshlabs, I'm still discovering this soft. For my scanned buddha head, there are no other treatement that the one i've mentionned.

Thanks for sharing your experiments.

Freelss 1.7
Raspberry Pi 2
Piclop personnal design (check Website)
Direct RJ45 connection

adaptedesigns's Website