GearHeads Forum

Just starting out. I didn't see the answer to my question but maybe I just don't know how to search for it.
Trying to use the stereo photometry tool. When I try to output my results (STL or DFX) the screen image that I have created disappears and I'm given the options to output the current display or all individually. Objects to export is blank.

What am I missing?

BTW there is nothing in my project tree.

Thanks for any help.

Hi:

  When your screen 3d looks ok, you need to press the "Add stl" button. This will add the stl to
your project screen for export. If it doesnt create one, you havent changed the x,y output size, if the
system see's that hasnt been changed, it flashes a quick message at the top telling you the output res
needs to be rational.. its default is the camera resolution..

Art

Alber;

From what I remember of stereo Photometry, 4 steps must be followed in sequential order for it to work. 1)pick you camera and res, 2) load bmp image and set the acquisition speed, 3) calculate normal's, 4) create 3D. From the impression I get of the problem you're experiencing, is it possible you left the output size set to X0,Y0.Z0 before export?

Cheers
Bob
 

Thanks. At first, I was not sure what you meant by changing the output to something reasonable. Then I hit Add STL several times and realized that it said change it from pixels to inches or mm. I just deleted a zero and it was happy. I'm not sure if it thinks that it is inches or mm.

The captured images don't look very good. I thought that it was exposure so I changed the Aquisition speed.

I switched to a TV screen that is 38" diagonal. I taped the webcam to the center of the screen and turn the lights off.

Aquisition speed questions. It does not seem to make much difference here is 25ms and here is 250ms.

Aquisition speed does seem to increase the confidence but I'm gettings numbers in the 20-50 range.

I also have to change the Y fisheye correction and the gradient scale to get anything like my face. One benefit is that the rest of my head looks like the Greek god Hermes

Thank you in advance for any advice.

Hi Alber:

 Not unusual to get output similar to what you show. I found I had to play with the camera settings like
white balance and such to tame it down. I also recommend clipping it, this removes the drawn out background.
 the output varies a lot though depending on camera used, monitor, darkness..etc.
Play with the settings, its really a matter of finding the sweet spot.

    Monitors arent great at providing light for that type of 3d determination. Im working presently on a new
method, but you need a projector for structured light 3d to work. It may appear in a future vexx.

  The head in the photo below was machined into HDF as an inward 3d head. Its eyes follow you as you walk past
it and its stl came form the webcam module. Its not a perfect 3d of myself, but its close enough to be eerie..

Art
 

Great. Thanks. I was doing a lot of fiddling now I know I not totally on the wrong track. 

What does clip do?

Will one of those cheap portable projectors that say HDMI but that only describes the cable. They are only VGA resolution. Not bitter. Not at all. Not even, well a bit pissed. If that will work I can say that I intended to use it that way all along.

Hi:

  Clip will remove that back portion, think of it like slicing off the front. The rear most always gets stretched too far back
due to the poor  light you really get from monitors. That module is a fascinating look at how so little information from
minor light changes can be used to determine 3d coordinates, unfortunately it takes a lot of playing with conditions
to get a good one.

  In the method Im playing with now, Im sending structured light patterns to my projector through the second
monitor jack on my video card. Ive also added in the capability to control a Canon camera. There are many related
methods to doing a structured light projection method for 3d, from binary coding, gray code, fringe field and mixtures
of the three. Ive purchased a Canon rebel camera from ebay for 88 dollars that should be controllable from the
software.  I dont know how this method will work, but Im curious enough to code it up and see. My patterns look
good from this cheap projector, but a webcam seems to be terrible at doing a photo, hence the Canon camera...

  I hope to send very fast flashes of various images, take fast photos between each projection change,
and process it to get hi res 3d's of faces for laser 3d engraving. If it all works Ill add a button in Vexx to
allow anyone with a Canon camera and a projector of some sort to run it, though how many people fit
that category is highly questionable. Most Canon cameras will speak to software API's so Vexx has an
amazing amount of control over the camera.. if it all works as advertised anyway..

  As you have interest in 3d imaging, Ill keep you posted on that projects progress.



Art