Page 1 of 1

Commercial Full Auto Book Scanner

Posted: 23 Mar 2017, 21:00
by jck57
I'm trying to decide whether or not to build a full auto scanner to sell. It would be a more streamlined table top version of this : viewtopic.php?f=14&t=2776&start=40. The scanner's robotic actuators would still be powered by compressed air so it would require a separate compressor. I've been working on the design with an eye for dependability and low cost. Although my prototype was rather big and clunky, it performed pretty well. Not very fast but it's good with various book sizes and paper qualities. The double sheet detection system stops the machine before it scans the same page twice or skips a page. Then, after automatically adjusting the page picking mechanism, the machine resumes scanning.

My question for all of you: Is there a market for such a machine? Commercial automatic book scanners start at close to $100,000. My goal has been to come up with a design that could retail for less than $10,000 minus cameras and air compressor.

What do you honestly think? Am I wasting my time with this project?

Thank you for your consideration.

Re: Commercial Full Auto Book Scanner

Posted: 24 Mar 2017, 21:00
by dpc
Some issues that I can think of off the top of my head:

1) The type of client wanting a fully automated scanner likely wants the entire package (with support), and that means software capable of processing the images coming from the scanner and spitting out at a minimum a visually acceptable PDF. They aren't going to want to rely on a hodge-podge of free software packages to produce the end product.

2) Compressors are typically noisy and must be maintained. Is there the possibility of using electric linear actuators instead?

3) Manual labor to turn pages is cheap. It's a simple operation that a 4-year old can do. Google scanned millions of books and didn't resort to using automated scanners. With their smart people and deep pockets I'm guessing they could have developed an automated scanning solution if they thought that was the way to go. Instead they chose to invest in post-processing software to flatten pages and remove images of thumbs from the pages as they were held open.

4) Throughput does matter. It would be odd for someone to purchase an automated scanner if they don't have a lot of books to scan. The first thing they are going to ask themselves is how long will it take to get their entire collection of books scanned? This doesn't mean you can sacrifice accuracy of the scans or risk damaging the pages to speed up the process - these things are expected. So perhaps the solution to more throughput is utilizing multiple versions of your automated scanner in parallel? How would ten of your $10K scanners compare to one of the $100K units? Would it reliably produce more throughput than the $100K scanner? Would it require more maintenance (operating costs)? A plus of your design would be that you could probably share the same compressed air supply across multiple units. Perhaps there other ways to share the capabilities of multiple scanning units to increase throughput without it costing $10K for each additional unit?

5) What does the scanner do when it can't separate stuck pages or encounters some other problem? Does it stop and require manual intervention or does it continue and somehow notify the operator that there needs to be some manual scanning performed? If so, does your scanner have a manual mode where it could be used to scan the problem pages quickly?

One option might be to pitch your scanner to companies that already have more expensive automated scanners so that they can partner it with their existing post-processing software package. This would allow them to enter a different segment of the market without much of an investment. Even someone like ATIZ, who sells manual scanning solutions, might be interested. They have a nice software package that would complement your system.

Re: Commercial Full Auto Book Scanner

Posted: 24 Mar 2017, 22:34
by jck57
Thanks dpc for your excellent reply.

Yes, post-processing software is important. I have no experience or ability in this area. Either the user would have to resort to free software or somebody else would have to produce the software to go along with the machine.

Quiet air compressors are available and affordable. I'm still somewhat interested in electric actuators but from a purely utilitarian aspect, nothing can beat pneumatics. If you buy an air compressor you can also fill up your tires and drive staples!

I have also considered that Google chose to use manual scanning.

I don't have the exact numbers handy but I think 2500 pages per hour is about top speed in commercial full auto machines. My machine can probably do 1500. I am confident about page turning error parity. The bottleneck is not scanning speed but post processing.

My scanner can self-adjust to turn pages after sensing a no-page or double-page fault. The machine locks out after 10 attempts to turn the same page while readjusting after every fault . This is what would happen with two pages stuck together.

For several years now I've been working on various approaches to the problem of automatic book scanning. It's become a pastime and a passion. A day rarely goes by without thinking, scribbling, and revising ideas. All that doesn't mean that even a perfectly functioing and affordable machine would meet a genuine need in the real world. But in any case, I can't stop putzing around with it.