http://www.metaingegneria.com/scanner-laser-3d/
Another thing I’d love to build if I had time…
-D
Plus One on that. It is a nice foundational complement to the makerbots. So many toys, so little time… Jim
3D scanning is what I’ve been dealing with lately at work, and a grad student from Virginia Tech that I’m working with built a structured-light scanner on the same technique (though his is instead from the quite good PDF at http://mesh.brown.edu/byo3d/notes.html). I was considering making my own, if you or anyone else has some interest in this. Structured-light is a pretty fascinating technique.
Also, that link incorrectly says ‘nodal point’ when he means ‘entrance pupil’ when looking for the no-parallax point. That annoys me a little… http://www.janrik.net/PanoPostings/NoParallaxPoint/TheoryOfTheNoParallaxPoint.pdf https://en.wikipedia.org/wiki/Nodal_point#Nodal_points
If someone can be a little more precise about which dslr is required I will acquire one.
d
I have a canon t4i but don’t have a set focal length lens or a remote shutter release
Anything that can be remote-triggered is fine. Pentaxes have probably some of the cheaper bodies, and they can meter on the old prime lenses that are abundant for that lens mount and tend to have pretty low distortion. Whatever we do get, I can run it through lens calibration.
How many megapixels do we need?
-D
Somewhere between 6 and 12 is probably fine. Your resolution is somewhat limited here to how well you can actually resolve the laser line itself.
And will any slr work or do we need one with a specific lens. Im camera dumb.
Any should work; I’ll probably look for a lens myself since the chances of it coming with the right one are pretty slim and I’ll likely just look for an old prime lens.
A couple points that I had seen raised was that consumer SLR’s typically can take about ~60,000 pictures before the mechanical parts that raise the mirror die. This can be mitigated with some SLR’s by putting the camera in ‘mirror lockup’ mode, where it raises the mirror and just keeps it raised.
As for lenses, I know that the Canon EF and EF-S mount’s can get an adapter for old full manual Olympus OM lenses which are a fair bit cheaper on Ebay, especially for some decent fixed focal length prime lenses.
The Pentax DSLRs can also use and meter on a ton of old, fairly cheap glass without needing an adapter; I might also have an old 50mm Pentax prime lens around too. Did you have any Olympus primes on hand?
But the mechanical failure part could be quite a problem if this is a setup that will be taking tons of pictures. I guess this is why Point Grey cameras are often used here. If any point & shoot cameras have good enough glass and resolution (which they often do), and predictable enough focus and exposure (which they often don’t), they’re worth a look. Or, if we manage to find a mirrorless 4/3 one cheaply from Sony, can’t these take Konica-Minolta mount lenses? Decisions, decisions…
If someone guides this camera idiot I’ll be more than happy to go hunting. It doesn’t seem to be a problem to find Pentax DSLRs for $80-150 on ebay for bodies.
-D
Have you seen the 3D scanner that works similar to this that uses your smart phone vs a DLSR? Wouldn’t that be more economical?
Optics on smart phones tend to be awful, and in addition the phones almost never permit any sort of precise control over the focus and exposure via anything automated. The camera APIs in Android don’t even allow it.
The way I see it, just about the best that a smartphone-based approach is going to do is a similar structured-light approach, and the sole benefit that I can see here is that you could potentially process images directly on the phone; however, you are still stuck having to calibrate everything.
If it’s not structured-light based, it’s very likely a feature-based approach that cannot handle a variety of geometries and textures.
Here is the link, It’s been a while since I watched the video but I believe that they were able to focus but I believe they were using an iphone. The app seemed very intuitive.