How to "scan" oversized documents with DSLR?

Question

I tried a variety of search terms and came up empty, so here we go!

I want to digitize a collection of maps. All I have at home is a dinky A4 scanner. There are no oversize scanners to be found in my city.

However, I do have a DSLR, tripod, and small variety of lenses!

This is an actual example: Cassino '44 game map (relax, it's my photo)

That mega-map is composed of four maps, each measuring 85x54cm.

To digitize it with the DSLR I assume one would shoot directly center-downwards from a tripod... as for the rest:

• How much should I try to "get" in one photo? (i.e. - reasonable expectation for coverage? multiple shots of each map?)
• What lense focal length(s) will minimize distortion?
• What f-stop makes sense?
• How best to light the image?
• Can Photoshop handle the stitching heavy lifting? Or should I use a specialized tool like Hugin?

and so on.

Answer 1

This will be a game of trade offs. The first thing I would recommend is that mounting it on a wall may be easier than putting it on the floor since it may simplify lighting. Lighting the map clearly and evenly will be the most important aspect as it will be very noticeable in the end result if it isn't evenly lit. Using two or more strong, well diffused lights of the same type should probably allow for decent coverage of the map, but more light is better than less as long as it is even and the same color.

As far as taking the image, you have two options. You can use a fairly wide lens and capture the whole thing at once, which will produce a natural enough image, but will likely have some amount of distortion from the difficulty in correcting for the angle of view since the corners are coming in to the lens at more of an angle.

The alternative to this is to use a long focal length and take multiple images in a grid and piece them together in to a larger image. This also has the advantage of producing a more detailed image since the resolution will be much higher, but it is more time consuming and still requires that the camera be repositioned to be aiming straight on or distortion will still occur based on the map not being perpendicular to the axis of the camera lens.

Aperture will depend on how straight on you can get things and how far in you zoom. In practical terms, the depth that will be in focus is based on distance from the lens and thus curved, so when shooting something perfectly flat that is filling the view on a wide angle, it's going to move slightly out of focus on the corners if the DOF is too shallow. Adjust the aperture to account as necessary for this. If shooting with the latter method, it should be less of an issue since the entire portion of the image you care about will be on the same plane.

Answer 2

I am not an expert, but I spent some time thinking about this when I had to 'scan' a big diploma... In the end I just took a snapshot because I was in a hurry, but I hope you can find the following advices reasonable:

I would use Hugin to correct the perspective distortion, it should be easy if you have the corners of the maps to make perfect rectangles. If your maps have overlapping areas, then it can do the stitching for you too.

About the focal length, the further away you are (and the longer the lens), the less distortion on the edges... Of course Hugin can correct that too, but that would loss some quality (because of the interpolations done to remap the geometry).

The F-Stop does not matter a lot if the maps are perfectly aligned with the focal plane as for the Depth of Field, but vignetting and sharpness are important, so you could safely go for the optimal settings for the given lens (usually 2 stops down the maximum aperture).

The best way to light would be an even diffuse light, try to not use the flash to avoid reflections if the paper is shiny. Natural light with some reflectors (even white cardboard) could be nice. Do not forget to calibrate correctly the white balance.