A new scanner design using plastic tubing

Built a scanner? Started to build a scanner? Record your progress here. Doesn't need to be a whole scanner - triggers and other parts are fine. Commercial scanners are fine too.

Moderator: peterZ

User avatar
danielcastillo
Posts: 8
Joined: 11 Mar 2015, 21:54
Number of books owned: 0
Country: Chile
Location: Arica
Contact:

Re: A new scanner design using plastic tubing

Post by danielcastillo »

Saludos:

Bueno, la estructura mide 70 X 90 CM. y los tubos de PVC son de 32, los rellené con arena para darle mayor peso....no hay mucha variedad ... todos los tubos son celestes...y por estética mejor es pintarlos... sobre la lámina transpararente es de acrílico... es lo que encontré en mi ciudad (es muy poca la variedad, incluso pensé usar vidrio antireflejo; esa lámina acrílica salió como 27 euros o 28 dólares), veré si encuentro el Velglas. Lo que trataré de mejorar es el soporte para los libros, para que tenga mejor movilidad.

Nos vemos.
User avatar
davidlandin
Posts: 134
Joined: 22 Feb 2012, 15:44
E-book readers owned: kindle
Number of books owned: 0
Country: UK

Re: A new scanner design using plastic tubing

Post by davidlandin »

Greetings Daniel

You have made a wonderful scanner! It looks really excellent. The photos are very clear. I like the handle on the front for the user to push down on the platen. A very good idea. I will include it in my next scanner. Like you, I have also used camera tripods to hold the cameras. I think we need tripods to hold heavy cameras. I fix the front legs of the camera tripods to the scanner frame so the tripods and the scanner become one solid unit.

You said you had a problem with adjustments to the book cradle. I have a new solution which may help you. I have written about it today in my Instructable. It uses a simple wire rack from a kitchen. Have a look at this link
http://www.instructables.com/id/Book-Sc ... -for-the-/

It is good to see that you are using this scanner in the University. I think it will be very gentle for important historical documents, and the scanner will not damage them at all.
crossed flags.jpg
crossed flags.jpg (7.66 KiB) Viewed 11808 times
Saludos desde el país de Gales a Chile


David Landin
User avatar
danielcastillo
Posts: 8
Joined: 11 Mar 2015, 21:54
Number of books owned: 0
Country: Chile
Location: Arica
Contact:

Re: A new scanner design using plastic tubing

Post by danielcastillo »

Saludos David:

Es verdad, el colocar un mango al frente permite presionar la lámina y la hoja del libro de mejor manera, disminuyendo la curva propia de la encuadernación; a la vez, la velocidad de trabajo aumenta por que existe un mayor control y presición de trabajo.
Referente a su forma de mejorar la cuna de libro ¿tiene el link de la página?.... estoy buscando ayuda para mejorar ese aspecto y alguna forma de conectar dos cámaras a un sólo computador.
Sobre el uso de este sistema en libros antiguos, diríamos que me da más confianza que un robot - escáner, ya que los grados de daño del papel (o muy quebradizo, mucho ácido, costuras delicadas, hojas sueltas y rotas, etc.) es mejor trabajarlo casí de manera artesanal .... Nuestra sección de trabajo tenemos libros desde 1612 a 1930 con distintos tipos de papeles y formatos. David, su creación nos está ayudando para digitalizar estas colecciones, las cuales se podrán visualizar en una página web a finales de año, desde ya muy agradecido por tomarse un tiempo para escribir sus apreciaciones de nuestro trabajos.

:arrow:
"Regards David:
True, placing a handle to the front allows pressing the sheet and the sheet in the workbook better, reducing the binding curve itself; simultaneously, the working speed increases there is greater control and precision of work.
Regarding his way to improve the birthplace of book does the link of the page? .... I'm looking for help to improve the appearance and somehow connect two cameras to a single computer.
On the use of this system in old books, we would say that gives me more confidence that a robot - scanner because of damage grades of paper (or very brittle, much acid, delicate seams, loose and torn pages, etc.) is better work it almost handcrafted work .... Our section we have books from 1612-1930 with different media and formats. David, his creation is helping us to digitize these collections, which can be displayed on a web page at the end of the year, from already very grateful for taking the time to write their findings from our work
."

DANIEL CASTILLO
User avatar
danielcastillo
Posts: 8
Joined: 11 Mar 2015, 21:54
Number of books owned: 0
Country: Chile
Location: Arica
Contact:

Re: A new scanner design using plastic tubing

Post by danielcastillo »

Ahora puedo ver su LINK .... el navegador al traducir las conversaciones me oculta los link... :oops:
arpadlabadi
Posts: 4
Joined: 24 May 2015, 09:59
Number of books owned: 500
Country: Hungary

Re: A new scanner design using plastic tubing

Post by arpadlabadi »

Dear David,

My name is Arpad. I am from Hungary.
Thank you for sharing this great design, all the answers on this forum, and the shared plannes of this scanner. Really great job.

I graduated as a medical doctor and I am now working in biomedical research. Although during my studies I studied from printed books (digital ones were not that widespread at that time), I realy like studying from digital media, rather than from printed books. Making annotations in digital books is extremely flexible and powerful, and therefore it works out better for me.
I know that making searchable and annotatable documents throuhg scanning is one of the most challenging tasks because optical caracter recognition softwares require high quality imput in order to work accurately.
I also read in one of your posts that you did not have good results with OCR-s.
Although, may I ask you to share some of your english language scanned pages in a format right from your cameras (preferentially those made with the 16Mp resolution poin-and-shoot ones)? I would like to give it a try to what extent my OCR is able to handle those documents. I think that it would be reasonable to test whether diy scanning can help me, before starting building my own scanner.

Thank you again for this great project, and for your answer in advance.

Arpad

p.s.: By chance, one of your uploaded images is a scanned page of a Hungarian language Bible. Reading your posts in this forum eventually made me understand why you scanned it :-). Great work.
User avatar
davidlandin
Posts: 134
Joined: 22 Feb 2012, 15:44
E-book readers owned: kindle
Number of books owned: 0
Country: UK

Re: A new scanner design using plastic tubing

Post by davidlandin »

Hi Arpad

Thank you for your very encouraging comments. As you requested I have attached a sample page from an English book. I did this just using a hand-held 16mp point-and-shoot Pentax camera using daylight, and I was NOT using the scanner. The reason I didn't use the scanner is that it is being looked after by someone who lives about 200km away from here!

Here is the image:
http://diybookscanner.org/forum/download/file.php?mode=view&id=4978
http://diybookscanner.org/forum/download/file.php?mode=view&id=4978
You will see that this is not at all a good image. It has several problems
[*]It is sideways and needs rotating
[*]It is not flat - the pages are curved because I did not use a platen
[*]There is poor contrast on the page.

OK now, despite what I have said about OCR, I am coming to see that ABBYY Fine Reader is really very good. I submitted the document to ABBYY FineReader 11 Professional Edition and this is the result :
English Abbyy output sample.png
English Abbyy output sample.png (50.84 KiB) Viewed 11535 times
There are a couple of errors in this document, but nothing serious at all. Notice what AbBBYY has done

[*]Rotated the image
[*]Handled the page curvature
[*]Cropped the page to the opposite page remnant is not shown
[*]Enhanced the contrast
[*]Interpreted the text 99.9% correctly
[*]Exported to pdf (It will also export to MS MWord if you want)

Amazing. And I was able to buy ABBYY quite inexpensively (maybe £30 GB Pounds). Also Abbyy has excellent support for other languages. Here is a page image in Hungarian followed by a pdf file made from ABBYY. I set the ABBYY language to Hungarian for this test.

Original photo image
Hungarian page.JPG
Sample of the Abbyy output copied from MS Word
Hungarian ABBYY output sample.png
Hungarian ABBYY output sample.png (79.97 KiB) Viewed 11535 times


OK now - you may ask me why would you want to use a scanner if ABBYY produces such a good result?

The answer is that you need the scanner to produce a fast workflow with the pages flattened by the platen. Even if you use no scanner and just take the photo images, you will always need to do post processing - straightening the image, keystone correction (making the image rectangular) etc - and maybe OCR too.

Well I think ABBYY pro does all these things very accurately, and very quickly. I had not realised until recently how much of the post processing ABBYY will do in terms of image correction. I always thought of it as just an OCR tool. But it clearly does much more than OCR.

David Landin
BruceG
Posts: 99
Joined: 14 May 2014, 23:17
Number of books owned: 500
Country: Australia

Re: A new scanner design using plastic tubing

Post by BruceG »

OmniPage is another OCR program. See attached pdf file for output.
Imported same file as is.
No changes made within Omnipage
Font was already set for other work I am doing
2 errors can be seen (font size & line)

OmniPage also turns page and only selects the left page as does Abby
DIY example.pdf
OmniPage output
(35.54 KiB) Downloaded 482 times
User avatar
davidlandin
Posts: 134
Joined: 22 Feb 2012, 15:44
E-book readers owned: kindle
Number of books owned: 0
Country: UK

Re: A new scanner design using plastic tubing

Post by davidlandin »

Hi Bruce

I've got an older copy of Omnipage, and I've just ordered a newer version to try out. I wonder if you can explain a bit more what you meant when you said:
OmniPage also turns page and only selects the left page as does Abby
Thanks

David Landin
arpadlabadi
Posts: 4
Joined: 24 May 2015, 09:59
Number of books owned: 500
Country: Hungary

Re: A new scanner design using plastic tubing

Post by arpadlabadi »

Dear David,

Thank you for your response and for sharing these samples with me. Also, I would like to appologise for my late response to your post.
And thank you for Bruce for sharing his experience.

I downloaded the trial version of ABBYY FineReader 12 Professional, and made some trial as well with those images that you shared. The result is realy fascinating with both the English document and with the scan of the Hungarian Bible scan.
I also took handheld pictures from one of my textbook's randomly choosen chapter pages (page.jpg), and one of the glossary pages (glossary.jpg). I used my canon EOS 450D DSLR camera (it has a sensor of 12 Mpx), and I paied attention mostly to avoid blur resulting from handshake (I used strong illumination, wide apperture, and 400 ISO - with this setting the shutter was rather fast, and the noise level still considerably low). As you can see, especially in the case of page.jpg, the picture is not of the best quality in a sense that it needs to be rotated, the page is not flat especially at the saddle, and additionally, there is a considerable reflection of the lightsource there making the local contrast even lower. So I just opened the JPEG files as they were right from the camera in ABBYY (importantly only one file at a time - I will speak of its reason later), and just asked ABBYY to convert it to PDF. That's all.... And ABBYY made a great work: Automactically rotated the page, and made a very accurate OCR even at the curved and reflecting part of the saddle of the file (page.jpg). And even the text in the smallest font size were recognised correctly. What I also liked is that ABBYY made a considerable compression of the files while preserving a relatively good quality (output files: page.pdf and glossary.pdf, 150 KB and 69 KB, respectively).
I have to admit, I did not study how to use ABBYY, I just started using immediately after opening it. It was rather intuitive, and I am sure that further fine-tuning of the OCR and pdf conversion is also possible.

Here I send the jpeg and the corresponging pdf files.

ABBYY did a great work. However, one drawback of ABBYY FineReader 12 Professional (£99 GB Pounds on amazon.co.uk) is that it does not support batch OCR/conversion of multiple files. I think this feature is very important for those scanning books, and not only archiving single documents. So in order to get this feature - as far as I understood - one should buy the ABBYY FineReader 12 Corporate Edition (a single license costs £149 GB).

Definitelly, I do not intend to make an advertisement for ABBYY. Without knowing much about OCR softwares I just tried this one, because David mentioned ABBYY 11 in his post. Importantly, my trial was definitely not an in depth trial of the software - I merely made a couple ad hoc conversions in order to convince myself about the effeiciency of OCR from those documents made with diy scanners :-). Second, I did not make any research at all to check other comercially available OCR softwares. I am pretty sure that ABBYY may not be the ultimate solution (although worked fairly decently), and other producers may offer softwares of similar quality for this price tag (or maybe even lower).

So what I would conclude is that it is quite a good idea to make a diy scanner if someone would like to scan books (and especially a lot of them) even if the final goal is to achieve accurately annotatable, editable and searchable documents - it can done.
David also raised the question that if OCR softwares are so good, than why to build a book scanner instead of shooting the books from - let say - a tripod instead. Well, my opinion as well is that a good diy scanner makes scanning much faster and the final result will be way better than without it. So for me, aming to scan my textbooks from the university, if a diy scanner increases the OCR-s accuracy only with 0,5%, that could make a difference whether I overlook a topic or not. - And also, I think it is fun to change ideas, identifying and solving problems while building such effective tools from minimal budget. That is what diy is about afterall.

David, thank you for your answer again, and for this great construct. I will give it some try as soon as I can.

Arpad
Attachments
glossary.pdf
(68.44 KiB) Downloaded 376 times
glossary.jpg
page.pdf
(149.78 KiB) Downloaded 382 times
page.jpg
arpadlabadi
Posts: 4
Joined: 24 May 2015, 09:59
Number of books owned: 500
Country: Hungary

Re: A new scanner design using plastic tubing

Post by arpadlabadi »

...It seems, that the corresponding jpg files were not uploaded for some reason. I give it try again. And, yes, I know that the lack of white ballance adjustments makes this files rather strange. Well, that is what should be done in camera preferably, and the OCR software does not adress this part of post processing (how it should?). Anyway, fortunately it did not disturbe the OCR process at all and therefore seems to be an aesthetical issue for me.

Arpad
Attachments
page.jpg
glossary.jpg
Post Reply