Writing out the labeling of boxes, bottles, and solid packaging is an annoyance. Think for a moment of the great leap ahead brought about by the label printer -- print a label on a flat adhesive strip and then wrap it around a bottle or can.
The downside is reserved for those who have to write down and classify received merchandise. Still needs be done with a flat sheet of paper, pen, and formfile or index system. Writers' cramp is a horror!
Is there some way to put together a darkbox for the top of a copier or scanner that would facilitate the printing of an image "in the round" upon current standard copier paper? I used the descriptor "Cat's Cradle" because the cashier scanners I've seen spin out a web of bar code reading lightbeams that reminded me of a web strung between fingers that we used to call a 'cat's cradle'.-- reensure, Feb 17 2002 Actually, Rods, if it's a video camera it can make use of the motion (as you're flipping the object over) to discover the object's shape (I've seen this demonstrated; hold up an object in front of the camera, wave it around, and out comes a 3D wireframe model. More usefully, you can fly an airplane over a piece of land and get a topographic map). Knowing the object's shape will help it "unwrap" the label from any curved surfaces, etc., before running it through the OCR. This is all somewhat compute-intensive, but we all have multi-GHz multi-processor computers in our toasters these days anyway.-- wiml, Feb 18 2002 Ditto what [wiml] said about motion processing as part of the unwrapping process. A 3D scanner would probably work best where the scan software controlled the position of a turntable on which the object rested.
An alternate approach might employ the techniques used in holography. Coherent light (a laser) is split into two paths with a fixed splitter. One path goes directly to film (or in this case a high definition video camera). The other path involves reflecting off the object being scanned and then going to the film (video camera). The surface shape of the object is encoded in the phase differences between the two light paths when they recombine. Optical interference produces the brightest spots where the paths are exactly in phase and the dimmest spots where the paths are exactly 180 degrees out of phase. Use a second camera to capture just the reflected light without interference, and perform the reconstruction by computation rather than by optics.
// but we all have multi-GHz multi-processor computers in our toasters these days anyway // But how many of us have a (Newtek Video) Toaster in our multi-GHz muli-processor computers? Hmm?
<aside>Gee I'd love to see the people who brought us the Amiga computer do that kind of thing again with today's technology</aside>-- BigBrother, Feb 18 2002 Suspend your target object in a suitably-positioned clamp, and attach your scanner optics to a moving head, like a CAT scanner. Combine with motion-control software.-- angel, Feb 18 2002 Yes, angel, camera + motion control = the best solution (at least, I think so). In fact just an stepper-controlled index table (model-mover) and a stationary camera would probably do it.-- bristolz, Feb 18 2002 BigBrother:"<aside>Gee I'd love to see the people who brought us the Amiga computer do that kind of thing again with today's technology</aside>"
A computer that's fairly powerful, does lots of pretty things but has few applications written for it and that has a rabid cult following? Isn't that the Apple?
<NOTE: The above is humor. I have no interest in starting the holy wars...>-- StarChaser, Feb 18 2002 random, halfbakery