Methods To Sense The 3D Surface/Structure Of A Book
Moderator: peterZ
Re: Methods To Sense The 3D Surface/Structure Of A Book
Or the lasers will need to be projected from the side, instead of the camera... side camera means more than one...
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Methods To Sense The 3D Surface/Structure Of A Book
If you look at the rig I built, the lasers are being projected in from the side at a 45 degree angle. That solves the problem of the camera and laser being coaxial. It also converts vertical displacement to horizontal displacement, which is easily sensed.
Anonymous, I do not agree that we can get away with the book images alone, unless we do stereo and there is a high correspondence of features among them. Many, many people have tried that and I believe that we have the very best one (ST) right here on this forum. I'm convinced that we need to augment the images with a little something extra, whether that be controlled lighting or a laser beam.
My goal is to get scan speeds up to page-turning levels. Scans as fast as I can turn the page. Minimal hardware. I think we have the technology to do this now and it's just a matter of assembling the pieces plus a bit of software development. I'm a shitty programmer but I'm going to be doing my best to make this happen.
Anonymous, I do not agree that we can get away with the book images alone, unless we do stereo and there is a high correspondence of features among them. Many, many people have tried that and I believe that we have the very best one (ST) right here on this forum. I'm convinced that we need to augment the images with a little something extra, whether that be controlled lighting or a laser beam.
My goal is to get scan speeds up to page-turning levels. Scans as fast as I can turn the page. Minimal hardware. I think we have the technology to do this now and it's just a matter of assembling the pieces plus a bit of software development. I'm a shitty programmer but I'm going to be doing my best to make this happen.
Re: Methods To Sense The 3D Surface/Structure Of A Book
I assume you guys have seen this - http://hackaday.com/2010/03/19/flip-boo ... itization/ - ?
It's what turned me on to book scanning in the first place, simply awe-inspiring. With our budgets (and programming skills!), we can only dream of such speed and accuracy for now
It's what turned me on to book scanning in the first place, simply awe-inspiring. With our budgets (and programming skills!), we can only dream of such speed and accuracy for now
Re: Methods To Sense The 3D Surface/Structure Of A Book
The link to and discussion about that video was a few posts ago on the previous page.
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Methods To Sense The 3D Surface/Structure Of A Book
Well, it is super fast, but it is not that accurate (yet). I agree that it is very inspiring/motivating.ahmad wrote:we can only dream of such speed and accuracy for now
I really feel like we are just a step or two from a scanner that can capture as fast as we can turn the pages. If the laser lines or lighting tricks work out (got my BuckPucks today for the LED arrays), we can go as fast as the camera can capture, which is pretty fast. For my D200, for example, that's 7 frames a second, much faster than I can reasonably turn pages.
I really think this is the future of DIY scanning. We'll always need V-platen scanners, but we could also use some very fast scanners for bulk work.
Re: Methods To Sense The 3D Surface/Structure Of A Book
If this works out I'll probably have enough excuses to buy the EOS 5D Mark II at 21.1 mp... more than enough to shoot both pages of a textbook at >300dpi... heh.. plus I can use it for so much more than book scanning... just the $2500 price tag for the body makes me cringe a bit, but having another practical use for it besides taking great pictures.. heh...
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Methods To Sense The 3D Surface/Structure Of A Book
Totally. And that camera was announced in 2008 - soon the $2500 price tag will be a thing of the past. Heck, you can get 18mpix for $799 right now: http://www.amazon.com/Canon-T2i-Digital ... B0035FZJHQ
Re: Methods To Sense The 3D Surface/Structure Of A Book
Yea but the 5D Mark II is sooo much better than the Rebels. For more than just MP size...
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Methods To Sense The 3D Surface/Structure Of A Book
Totally. My D200 is all-metal, weather-sealed, pro-everything, and you can feel it. It was worth the premium over the cheaper models, back in 2006.
Re: Methods To Sense The 3D Surface/Structure Of A Book
Hey, that 550D isn't all that bad
I have one, and it's amazing. It does feel a bit flimsy, though, as it weighs considerably less than the 5D and 7D, but it still has amazing quality. 12,800 ISO, 1/4000 s exposure, 1080P video, 18 MP (5184x3456 image resolution), 3.7 shots/second, and MagicLantern. It was a huge upgrade from my flimsy 6 MP FujiFilm point-and-shoot, but it's not really comparable to full-frame DSLR's (but it's good for book scanning!)...
I have one, and it's amazing. It does feel a bit flimsy, though, as it weighs considerably less than the 5D and 7D, but it still has amazing quality. 12,800 ISO, 1/4000 s exposure, 1080P video, 18 MP (5184x3456 image resolution), 3.7 shots/second, and MagicLantern. It was a huge upgrade from my flimsy 6 MP FujiFilm point-and-shoot, but it's not really comparable to full-frame DSLR's (but it's good for book scanning!)...