This is just a thought I had and wanted to run it by yinz to see if it could work in theory.
The idea is that by hacking a 2d flatbed scanner so that the beam doesnt move from side to side, is rotated 90 degrees, and will keep scanning until you tell it otherwise, could you make it into a 3d scanner?
Assuming you get that physical hack above to work, then you would then turn the scanner on its side, place a box painted all black inside over top of the scanner to simulate a dark room (using flat black paint not glossy). Then get a device that can rotate slowly and evenly (heck legos could work) and run a poll through that and into the box with a plate on the top (essentailly a turn table). You would then place your object onto the plate and as the scanner scans, you have the plate rotate and the beam would then scan the entire object on all sides minus top and bottom which could easily be filled in later in something like blender.
The only issue I am really seeing with this would be getting a program to turn the scanned date into a 3d model as opposed to a 2d scan.
Now is only theoretical but does it sound plausible?
ive looked at alot of homebrew 3d scanners and they all have one thing in common, that damn digital camera. I mean I have one, but setting up the angle and calibrating blah blah is a pain and scans just dont seem to come out nice. The idea was that the scanner itself would do it all.
I think what would be ideal would that a program would take all those scans and subtract them from a cube, kinda like a reverse heightmap, so essentially it would kinda work like a real artist does when carving stone or ice. But im not sure if there is such a thing as subtractive modeling.
The alternate route would be to compile the scans into a 3d image much like other (freaking exspensive) software does.
2d-scanners are pretty unsuited for 3d-scanning purpose. since they can only scan a thin line.
for doing 3d-scanning you need to scan the distance between your object surface and your device.
with a single line for scanning the only possibilitys i see is to get this distance via triangulation, which would mean.project a point onto your model. put a black piece of paper with a hole in it between your object and the scanner and find calculate the distance based on where the point is localted on your scanned line (using trigonometry). (simmlar to autofocus devices in digicams)
other ways are to do the traditional laser-line scanning (the line can be easily made from a laserpointer which points onto a small mirror mounted on a small motor.) (easy to build yourself)
also common are some sort of projected bars which get photographed. this way you can scan large-sized objects within seconds.(harder to make yourself , usualy quite expensive)
onther intresting idea was to use cheap distance-sensors mounted on servo-motors. some child build a 3d-scanner which was capable of scanning his own room. very slow but cheap and worked
long answer short. flatbat scanners are unsuited to build 3d-scanners.
aside from hardware hacks you would need to write a special scanner driver in worst case. so you better stick with some images, taking the 5 minute trouble to calibrate it once after all. there is a reason why 3d-scanners are quite expensive
