Hello Guest it is March 29, 2024, 07:43:46 AM

Author Topic: Photos of webcam scanner. ( no laughing.. :) )  (Read 67448 times)

0 Members and 1 Guest are viewing this topic.

Offline K3HTK

*
  •  19 19
    • View Profile
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #30 on: April 11, 2008, 10:07:35 AM »
Okay something is goofing up for my post, so here goes again.... ID10T error on my behalf maybe? Ha!

Art,

Here are the files you wanted..... The pointcloud file is supposed to be a roll of masking tape. The others are my setup and a screen shot of the video probe screen. Maybe my camera is too high at 8" above the table? Let me know if you think I should move it closer to the 2 - 3" range.... Thanks again!

-Ed-

Also the cube is 26 mm wide that I have on the table..... FYI
« Last Edit: April 11, 2008, 10:09:11 AM by K3HTK »
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #31 on: April 11, 2008, 10:32:48 AM »
Hey Ed,

I looked at your cloud file using a simple text editor as well as in Mach Cloud.  Your right not much there.  T

I am confused about your setting in the scan y box.  It appears that you are using Metric parameters.  If so, a 3mm scan area is too small for a roll of masking tape.  It needs to be set to a little more than the length of travel your gantry must go so the laser/camera can pass entirely over your scan subject matter.  When you do your scan does the laser move step by step entirely across the roll of tape?  If not, try a Scan Y value of something like 300 (mm) and see what you get.

The other thing I noticed is that the Z data in your cloud file stays pretty steady at 1.000- THAT is a problem but, it might be related to your lack of travel issue previously mentioned.  Before we go on to troubleshoot other issues- verify your scan parameters and get back with us!

HTH,
Sid

Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #32 on: April 11, 2008, 10:47:47 PM »
Slowly getting the hang of things....

Moved the location of the lasr- raised it up higher to get a more defined line.  Changed position of the camera.  Turned off the lights.  Worked with the Ignore setting and things arte starting to come together.

To try and bring things closer to what Art was doing, I tried doing my hand.  It worked pretty well!  One thing I figured out is that once you view the cloud in Mach Cloud, in order to get a good mesh, the 'DECIMINATE" (I just love that word) feature is really helpful.  I found that doing it repeatedly helps to reduce the 'clutter' so to speak.

I have also noticed that Mach Cloud kinda goes funky on me while doing the Ball Pivot mesh thing.  About 40% of the meshes I tried nothing happened.  Or the mesh turned out weird.  I shut down Mcloud and re-launched and sometimes meshes worked again, sometimes they didn't.  So, I just kept shutting it down and relaunching and eventually it works again.

It's kind of irritating because it's not consistent, but on the other hand, it makes me feel like Thomas Edison searching for Tungsten- knowing that when I find it the whole friggin' world is gonna light up!  I guess what I am trying to say is that if things aren't working right, I'm not freaking out.  There is a little 'voodoo' element to all of this right now. But I think I have arrived at the point that I can duplicate results- WHEW!

I will say that after a few bad scans of my hand it took a lot of self control to restrain myself from curling up the 2 fingers on either side of the middle one  :o   

This process really is pretty darn cool!

Regards,
Sid

Offline ART

*
  • *
  •  1,702 1,702
  • Tough as soggy paper.
    • View Profile
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #33 on: April 12, 2008, 09:09:06 AM »
Hi:

 Glad to hear your starting to get results. It really is a bit of an Art more than a science as yet.
Just a note on MachCloud, the meshing is really pretty touchy about point cloud density, finding tangental
vectors for point cloud points is a difficult statistical issue, but I dont think restarting the program is whats making it
work for you, Ive done hundreds of meshes, and I find its the decimation of the cloud thats really the trigger as
well as the ball size used. Ball meshing is basically placing a ball of a set diamter on the point cloud at an arbitray spot, then
rolling it over the point cloud and creating mesh lines whnever the ball hits two other points. Its analogous to rolling a ball
bearing over the surface with some conditions as to the angle of the lines created in reference to other mesh lines in the local
area. Its very complex really though it sounds easy, so point cloud conditions are important.
  Id simply remesh with variosu parameters, and decimate when necessary, decimation tries to erase points that are
non-respresntative of the local areas they reside in as well as averaging the tangental vectors of the neighbours. So its a bit
of an Art more than a science.

  Good luck,
Art
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #34 on: April 13, 2008, 04:43:35 PM »
Well,  most of that makes sense.  Some of that Tangental Vector stuff is a little over my head though but, I think I get the Jist of what you are saying- Thank You!

I think that there is some kind of an issue with my machine being set in inches and you doing your stuff in Metric.  Somehow the conversions don't seem right sometimes.  I'm thinking that emulating what you have done has been working for me thus far so I'm about ready to abandon the Inch system and retune my system for metric units and see if that changes anything.

Having had some real breakthroughs in the past couple of days, I want to try and do some bigger things.  That would involve using the X Scan.  Any tips or is there anything I should  about tryijng this?  I know McCloud (wasn't that a 70's TV show?) asks for the left/right arrow thing to register the striping- any tricks to making that work or should I just 'Tinker'?

Thanks a ton for you assistance and input Art- it really has helped me.  I hope I can help others the same way!  Well, except for all of that software coding stuff  ;)

Regards,
Sid

Offline ART

*
  • *
  •  1,702 1,702
  • Tough as soggy paper.
    • View Profile
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #35 on: April 13, 2008, 07:30:18 PM »
Sid:

  When doing striping, Set the X scan to the total amount of stepover, and the X step to somethign liek 200mm's.
It will then scan in the Y, come back, move over 200 and scan the Y again, and repeat till the entire X range is used.
When loading into MachCloud, ( The TV show was McCloud. :) ) , when it says to use arrows, the second stripe shoudl appear,
at that point you arrow back and forth to try to match up the points as best you can, then hit enter. This will repeat for all Stripes. Inaccuracies will make it very hard to match things up. Results will vary widely. :)

Art

Offline K3HTK

*
  •  19 19
    • View Profile
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #36 on: April 13, 2008, 11:43:35 PM »
Hey Sid,

I am sorry... I am using inches here so I wasn't sure if put inches or mm in for values. I tried calibrating with inches and mm for my cube. Basically my cube is 1 inch so I had 26 mm in for calibration. Y was setup for 3 inches and it is moving 3 inches so sorry for the confusion as I was confused on what the heck I should use for this add-on. Ha! Mach 3 is setup for inches though.

As far as the roll of tape and travel goes, I was scanning about half the roll. I can try scanning something else. Maybe a can or something.

I am not clear on the Z data issue? Can you help me out on understanding why the Z data being steady around 1.000 is bad? I am not sure I understand the issue..... I am a noob so I apologize.... 

Thanks Sid!

-Ed-
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #37 on: April 14, 2008, 12:28:25 AM »
Hey Ed,

No Apologies necessary.  I think we are all learning here. 

OK, so the way I understand it- if your machine's native units is set up in inches, then all of the numbers you input should be in inches.  Art's machine is set up in Metric which is why in all of his examples you see mm.  That being said, if your machine is in inches, use inches to calibrate your cube (i.e if your cube is 1 inch, enter 1.00), specify your Y & X Scan distance and the step in inches and if necessary your Ignore Z parameter in inches.

I tried a roll of tape in the beginning as well and didn't get results at first.  Now, I think it was due to other issues but, nevertheless I decided to start with something really simple- my trapazoid wood block.  It worked and now that's going to be my 'lucky block'  ;)

As far as the Z issue is concerned, try and think about what the scanner is doing as the exact opposite of what happens when you cut a part using your machine.  When you cut a part, the computer is reading the G-Code and generating/dispersing commands telling your motors where to go.  The scanning process is just the opposite.  When you scan the computer is COLLECTING axis information via the camera. When the scan takes place the computer is recording data. It records its position in the X, Y but what we are really looking for here is what is happening in the Z.

The Y and/or X positions are specific and predictable (because we told the computer where to move them and how far).  But the Z is where the magic is.  As the laser passes over the object you are scanning the camera sees this object as being higher (i.e. away from the table in an upward direction).  This is the Z value that makes the scan work. 

Example...  If you place you 1 " cube and scanned that whenever the laser is shining on the top of the cube, the computer should register 1.0.  If you told you machine to scan a 1.5" area in the Y axis and a .25" step the point cloud file should look something llike this:
 X        Y      Z
0000  0000  0000  (no cube yet 
0000  0025  1000  (start of cube
0000  0050  1000  (more cube
0000  0075  1000  (more cube
0000  1000  1000  (end of cube
0000  1250  0000  (no more cube
0000  1500  0000  (no more cube


When you at a point cloud file using notepad or wordpad.  You see 3 sets of numbers seperated by a space.  These are you X, Y & Z coordinates.  Theoretically you could add the X, Y & Z before each respective vallue and throw a G1 command at the beginning of the line, you would have a cut/toolpath file.  Of course you would have thousands of points to cut and I think it wouldn't be something that we are all use to in terms of smoothness and speed.

When I looked at your point cloud file and saw that the last set of numbers on every line where all aproximately 0, it showed that you had a scan of a flat piece of something. Try recalibrating your cube using the 1.00 inch value and then rescaning.  I think your getting pretty close to getting some clouds.

I hope that all makes sense- it's hard to explain but you will get the hang of it.

Good Luck,
Sid

 
« Last Edit: April 14, 2008, 02:10:58 AM by sshneider »

Offline Chip

*
  • *
  •  2,055 2,055
  • Gainesville Florida USA
    • View Profile
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #38 on: April 14, 2008, 01:22:53 AM »
Hi, Sid

Looks like your doing pretty good with this, Hears the pic of Ports Pins, Mill Options for X,Y,Z, Coordinates letters addition.

Not sure MachCloud will load them though.

Thanks, Chip
Re: Photos of webcam scanner. ( no laughing.. :) )
« Reply #39 on: April 14, 2008, 02:06:52 AM »
Thanks Chip!  I remember that you told me about that before and I tried it but it didn't work for me.  I think that check box will cause one of the other probe based digitizing plugin/wizard to add the letters when it writes the file.  I'm not a 100% sure on this though.  I am awaiting a touch probe and will check out this feature when it arrives. 

I am pretty excited about my progress since we spoke last week.  Thanks again for your help & advice.

Regards,
Sid