I've got a crazy idea to bounce off ya'all... Like a lot of you, I have a large collection of paper ship plans. Some are common, but quite a few are rare and out of print, or at least extremely hard to get in the US. I need to get these digitized so I don't lose them to fire/flood/etc, not to mention it would make it much easier to make 3D models out of them, but the local blueprint places won't scan them for me without permission from the companies that printed them (most of whom have been out of business for years, if not decades). Low-end large format scanners seem to start around $1500 and the price goes up steeply from there, so buying one of those will not be in my budget any time soon. The other problem I have with using large format scanners is that they can have issues with really long blueprints. I once had one 3' x 10' blueprint scanned when I lived in Texas, but they didn't feed it into the scanner correctly and it went in crooked enough to both jam the scanner and tear and crumple the blueprint. For a while I thought they were going to have to cut my blueprint off the spindle to get it free. That, thankfully, was a cheap and common blueprint, but its definitely the kind of thing I want to avoid with my rare/antique blueprints... Seems to me that if we hooked up a $99 - $200 handheld book scanner (that typically scans 11" wide by up to 3' long at a time, max) to a computer controlled X-Y table (don't really need a Z axis) we could make a large format blueprint scanner that could scan large blueprints in smaller sections, then use photo tools to stitch the sections together to get the full large format scan image. From what I've researched, many of the handheld book scanners have enough built-in memory to handle dozens of scans or more, so we could wait and download the individual section scans after the entire blueprint has been scanned. My largest blueprints are roughly 3.5' x 11' , so I'd need an X-Y table that is roughly 4' x 12'. The good news is that the handheld scanners are light, and there are practically no work forces to speak of (just the weight of the platform, the friction from the scanner on the surface and the forces from moving the platform itself), so the gantries don't have to be beefy at all. I know that having a 4' x 12' X-Y table sounds kind of crazy by itself, but my new house will have a large unfinished basement area (roughly 29' x 15'), and I have a large attic area (roughly 23.5' x 23.5') over my garage I could use for this, too. And I can easily make the table so it can fold up against a wall to take up less room when not in use. I might need some lexan or plexiglass cover pieces to lay over the blueprint to help keep it flat for scanning, but these can be very thin. The control software would need to be able to control moving the scanner to each section and performing the scans. The control software would also need to know to overlap the sections when scanning by some amount so the photo tools can stitch them together. It would be nice if the control software could be told the overall size of the blueprint as an input parameter so that it can minimize the scan time on smaller blueprints, but that's not strictly a requirement. Mind you, I have no experience with the handheld book scanners, so I don't know how well they scan or how finicky they are. But if they are reliable enough for this, I think one could build the entire scanner for about $500 - $600, including the cost of the handheld book scanner, if one used an Arduino or Raspberry Pi to drive the servo motor controller (perhaps a RAMPs or RAMBO board). So, am I totally crazy, or does this sound somewhat feasible?
A good Phone camera can shoot in Panorama mode and automatic stich the images together in one file, worth a shot (eh good pun eh).
I would worry about edge effects when using a camera (not all points are equidistant from the camera). Also, you need all the photos taken from the exact same height or stuff won't line up correctly. So if you are taking about doing it by hand, probably not. If you are talking about replacing the handheld scanner with a phone camera, we'd need a tall Z axis, even if its fixed height, in order for it to work properly, adding weight to the platform. Might be harder to control the phone's camera remotely, as well. But the camera would cover a larger area, and thus need fewer sections and less stitching. If I was going to use a camera, I might be tempted to use a cheap Raspberry Pi camera, as these interface directly to the Raspberry Pi, are very light, and can be easily controlled remotely...
This looks doable. I'd build a scan table with some sort of indexed guide to get consistently spaced scans for stitching in your preferred graphics package. http://hackaday.com/2013/09/27/hacking-a-flatbed-scanner-to-scan-very-large-documents/
Great link, Steve, thanks! I like the idea of using a regular, modified scanner that sells for ~$25 a lot better than using a handheld scanner that sells for $99 - $200. And it gives me a great way of scanning large docs now without waiting to build a large X-Y table and controller. All I need is a modified scanner and enough space on the floor. ;-) The downside is that it can't be automated as easily as a handheld scanner. The scanner's brain in the above example is a PC, so scanning a large blueprint would require moving to the given section, clicking "scan" on the PC, moving to the next section, etc. so you'd have to be there to click 'scan' after every bed movement. With a handheld book scanner, the scan brains are in the scanner itself, I just need to have remote controls to actuate the buttons that say 'start scan', then start moving the scanner platform. Or better yet, just hack the buttons by removing them and adding remote switch control. It would really be nice to just press a button and have the machine scan the entire blueprint w/o user intervention. I must admit that I am starting to think more about a lightweight camera bed rather than a handheld scanner bed. Raspberry Pi cameras are so lightweight, easy to control, and inexpensive that if the camera would give decent results, it's hard to see why I wouldn't go that route. It would be a lot faster than using a scanner, and cheaper.
I think it could be done. If you keep it really simple, it would not need to be very complicated. For example, there is no side force like a router has, so you can use gravity to keep the carriage on a track. Some cheap scanners I have taken apart have used this method. A simple piece of aluminum angle with two bearing in a V configuration would work. Next I would skip the y axis. It makes for a little more manual labor, but as long as you overlap the scans the stitching software will take care of the variations from manual placement. I would also not put a precision controller on the system. All you need is a constant velocity movement, not precise location. You could build a very simple controller that just sends a constant set of pulses to a stepper driver, manually turned on and off. The movement mechanism could be fishing line on a spool, easily made to any length.
Very good points! Using gravity to keep it on the tracks makes it easier to pick up, fold up and store away. And makes it cheaper. Yes, precision isn't needed at all, so I could have a much simpler motor controller. Hadn't thought about eliminating the y-axis, but that's an interesting idea. While I like the simplicity of 'fire and forget', in truth moving the bed manually and re-starting the scan 3 or 4 times for a large blueprint wouldn't really be a big deal. Don't think the fishing line on a spool will work if using a scanner bed, though, as you would need to reverse direction to get overlaps between scans. It would work fine with a camera bed, though, as you could always just stop the motor a little sooner to get overlap between photos. Love the ideas so far! This is already shaping up to be cheaper and simpler than I originally thought!
The fishing line spool will move it both forward and backward. The line forms a continuous loop from one end to the other. The line is wrapped around a spool a few times, which is enough friction to move the line. An example of the spools: http://deltaprintr.com/shop/parts-and-accessories/hardware/filament-drive-reels/ This delta printer uses fishing line to move the carrages. At 2:31:00 in the video, it show how the line goes around the spool View: https://www.youtube.com/watch?v=pmXZ_gqv2sg . While most printers have moved away from fishing line because setting it up is a bit tricky, it wouldn't be a problem in this application.
I took mine to fedex office (kinkos) copied them myself and paid.. not problem. Try a self-service place and you might not have an issue... HOWEVER it cost about $70 by the time I was done...
I've tried my local Kinkos/Fed Ex. Their blueprint Nazis consistently come over and stop me. This happened on three different occasions, so I gave up.
Those of us who have paid for the rights to distribute copyrighted plans appreciate that shops are watching to make sure they themselves are not party to theft.
I understand why they do it, and I fully support and respect copyrights. I do not participate in piracy nor do I promote piracy. The problem is that we often need to resize plans, and/or we need digitized plans in order to make 3D models we can use to make hulls, and that's something the copy houses' rules don't allow for. Not to mention that I want to be able to back up my plans in case of fire, etc. All that is legal as long as you don't sell the plans themselves, which I won't.
I'm interested to see how this works out. I have large amount of large scale paper plans (mostly copies from the US National Archives) that I would love to have backed up for safe-keeping and restoration and re-sizing. Print shops are by and large a total pain the rear end, with flaky machines that are often poorly calibrated, indifferent (at best) to aggressively unhelpful staff to outrageous prices. Last time I went to a print shop to re-size plans it took 3 tries to get them scaled correctly and they charged me for the sheets that they messed up as well... to the tune of almost $100.
It will be a back burner project for now, but hopefully I can get started on it this winter, you know, during the off season. I think the camera solution is both easier and cheaper, but I do worry about edge effects with a camera. So I want to do some experiments with a camera before I commit to building a machine for it.
I've been collecting plans for the last 15 years, I can wait a bit longer I tried to experiment with a small handheld "page scanner" a couple years ago. Although it did produce nice scans.... it was near impossible to use without creating massive distortions.... Making it worthless for sheets of plans that are sometimes 8 feet long. I dislike paper plans with a passion, they are often damaged in shipping before they arrive. And then they accumulate damage every time they are unrolled. Some of oldest sets (always stored in tubes) have started to yellow. Digitized (and backed up) plans will last a lifetime, and will last for use of future generations. Although I sympathize with copyright concerns, at some point preservation for the future should also be taken into consideration. Plans you can purchase now may not be available forever......
Would the copyright holders who frequent these forums consider changing the license of plans they have to allow digital derivative works to be created, if they could recoup their investment? Also are enough people willing to crowd fund new, digital plans? I'm tempted to try to draw up new plans for HMS Iron Duke to replace the crummy copy I purchased.
Finish the build before you do new plans for a boat you have not battled. Remember we drive FIGHTING boats (or most of us do) and they come home rather shot up, better plans won't fix that.
Nah. I realized the the ships I have are fundamentally flawed. I'm working on something *MUCH* better : P
Really!?!! Start a thread and post some photos of your ships. Let us take a look, you may be closer to having a battling ship than you realize.
That Iron Duke was voted as most feared twice, so I'm not so sure 'flawed' is the word you are looking for, maybe 'capable,' or 'proven' would be better
