Photographing Archival Documents and Maps ?

b@nD

Banned
Joined
Mar 22, 2012
Messages
754
Reaction score
1
At sites that host scans of old books one often notices that a camera is used ( Photograph and then OCR ).
This would be better than a normal flat bed scanner especially if one did not have a specialized book scanner that uses a platen
What sort of equipment would one need to do this ?

I am thinking along the lines of a FF DSLR plus a good macro lens plus a tripod
and
timer shutter release. ( so you do not have to fiddle with the camera once set up and to speed things up)
Good clean light and a nice surface to work on. ( flash is frowned upon )

I know that some higher end DSLR's allow you to tether a laptop so you would be able to see exactly how the photos ( scans ) were turning out.

What specific items / models / brands would you recommend for this task ?

DSLR ?
LENS ?
TRIPOD ? ( this is needed to set up everything so that the camera is stable and does not move during proceedings )


Suggestions / advice appreciated
 
How big are these maps? A FF dslr and a macro lens might be a bit of an overkill. Even an old canon 400D (consumer level body) will let you tether to your computer and the kit lens would afford you quite a bit of latitude when it comes to document size. A remote trigger would make life simpler than a timer though you could use a smart phone or just trigger it from the computer.
 
It also depends on the quality required e.g. to copy antique maps professionally one would use a vacuum table and a large format camera.

For anything less pretty much anything will work ...

BTW A true macro lens - with flat field focus - is a good idea.
 
A rectilinear wide angle would be more practical than a macro lens.
 
A rectilinear wide angle would be more practical than a macro lens.

A rectilinear lens can still introduce pin-cushioning and distortion (e.g. stretching) as one moves towards the edges of the frame.

This won't happen with a flat-field lens, which is designed to provide distortion-free edge-to-edge sharpness.
 
A rectilinear lens can still introduce pin-cushioning and distortion (e.g. stretching) as one moves towards the edges of the frame.

This won't happen with a flat-field lens, which is designed to provide distortion-free edge-to-edge sharpness.
Then it's not particularly rectilinear, is it?

An EF lens used on an aps-c body would reduce/remove edge distortion that might be present with a FF camera.
 
Google 'planetary scanner' for some ideas. If you have large maps you could possibly have the camera(s) mounted on some form of rail system where you advance the camera(s) in increments and then stitch the resulting images together.
 
Last edited:
I would use the Canon EF-S 60mm f2.8 macro on any APS-C Canon camera (or equivalent Nikon). Most antique maps I have seen have lots fine detail that will be lost in one wide angle shot so I recommend mounting the camera on inverted tripod column (column faces down between the legs) and then take photos of each section of the map. Freeware stitching software such as Microsoft ICE and Hugin do amazing jobs. I think the most important this is to make sure the paper/canvas surface is always completely flat else you will get variations in lighting.

For lighting, two desklamps with normal 100w tungsten bulbs facing towards the center point (i.e. directly under the camera) at a 45 deg angle will help cut out shadows.

Shoot in manual mode, tungsten white balance, manual focus and at the lenses optimum sharpness aperture, probably around f5.6 - f7.1 for the 60mm f2.8. Maintaining the same light levels throughout the shoot is critical else brightness variations on the individual images will make the stitch look like a checkerboard... So shoot in a dark room, with only the desk lamps on and you should be fine.

For stitching, from my own landscape work experience, try to overlap by at least 10% and most 25%. Too much overlap will introduce errors into the stitch.
 
I would use the Canon EF-S 60mm f2.8 macro on any APS-C Canon camera (or equivalent Nikon). Most antique maps I have seen have lots fine detail that will be lost in one wide angle shot so I recommend mounting the camera on inverted tripod column (column faces down between the legs) and then take photos of each section of the map. Freeware stitching software such as Microsoft ICE and Hugin do amazing jobs. I think the most important this is to make sure the paper/canvas surface is always completely flat else you will get variations in lighting.
Just as important is to make sure you physically move the map or the tripod rather than to just refocus on a new section. Hanging the map on a wall might be the only option for larger maps.
 
Answers & Questions // Questions and Answers

How big are these maps? A FF dslr and a macro lens might be a bit of an overkill.
Even an old canon 400D (consumer level body) will let you tether to your computer and the kit lens would afford you quite a bit of latitude when it comes to document size.
A remote trigger would make life simpler than a timer though you could use a smart phone or just trigger it from the computer.
Thanks for all the feedback

A whole variation
Some are BIG -- wall size more than a metre square
Many state archives are still in the dark ages when it comes to digitizing their holdings
Large fragile maps are a particular concern ( wide carriage plotter scanner presents problems )

Please could you explain in more detail how you would tether the DSLR to a computer.
Really sketchy on how this works ( have a 5D Mark 1 and DELL laptop )
Have one of these Timer Remote Controller TC-80N3

Agree on a tripod where you can invert the tripod column.

Will have to see how much I am allowed to do ?
This is for a personal project -- not involved in any formal capacity in any archive business.
 
Examples

It also depends on the quality required e.g. to copy antique maps professionally one would use a vacuum table and a large format camera.
Any links to these ?
I know I could go and Google it -- but would not really know what I was looking at :cry:
 
Rectilinear // Flat Field -- Plat Poot :--))

A rectilinear lens can still introduce pin-cushioning and distortion (e.g. stretching) as one moves towards the edges of the frame.

This won't happen with a flat-field lens, which is designed to provide distortion-free edge-to-edge sharpness.
Sorry -- bit lost here as well :o
 
Please could you explain in more detail how you would tether the DSLR to a computer.
Really sketchy on how this works ( have a 5D Mark 1 and DELL laptop )
Have one of these Timer Remote Controller TC-80N3
Canons come with some pretty good software including an app called EOS Utility which is the one you want. :) Start it up, connect your camera to the PC via USB and off you go.
 
Sorry -- bit lost here as well :o
http://en.wikipedia.org/wiki/Petzval_field_curvature
Oversimplification:
Ordinary lenses, when taking a photo of a flat object (e.g. afore-mentioned map), the image formed on the sensor is not perfectly flat, so the edges are minutely out of focus compared to the center.

For normal every day photos it's not a problem, but when you're working with macro shots where DoF is razor thin, it becomes noticable, so macro lenses are designed to have the image formed flat and completely parallel to the sensor, from center to corner.

That's what I understood from the article, I could be completely wrong :D
 
Some decent quality clear plastic to protect the documents and a handheld scanner for the higher-detail/larger documents? Less effort than having to move the thing you're 'scanning' about to take multiple photos of it to stitch those in.
 
http://en.wikipedia.org/wiki/Petzval_field_curvature
Oversimplification:
Ordinary lenses, when taking a photo of a flat object (e.g. afore-mentioned map), the image formed on the sensor is not perfectly flat, so the edges are minutely out of focus compared to the center.

For normal every day photos it's not a problem, but when you're working with macro shots where DoF is razor thin, it becomes noticable, so macro lenses are designed to have the image formed flat and completely parallel to the sensor, from center to corner.

That's what I understood from the article, I could be completely wrong :D

Spot on !
 
Top
Sign up to the MyBroadband newsletter
X