Scanner as a camera

[Graham Stabler]

My Nikon SMZ-10 stereo microscope came with a 5X projection lens used for photograpy which I replaced with a suitable adapter (including some optics) to allow me to mount my CCD camera however I was thinking about the projection lens today while scanning some papers.

So I fitted it and balanced my canon scanner on a pair of boxes that some cheap Velbon tripods came in above the microscope adjusting things until I got an image at the plane of the glass then I set the scanner going. I got an image though it wasn't particulary good (I think I was getting back reflections from the cover glass) however it made me wonder if trying a high res scanner with the illumination dissabled might be worth a try, the image is fairly large at around 2" accross. It might well not provide any advantage in my case but it seemed a potentially cheap way to add a camera to a system.

Graham

[puzzledpaul]

Someone who works locally has recently won a runners up prize in a comp. with images created / montaged in this manner.

(Am at the bottom of the food chain and just bought his old dslr )

pp

[rjlittlefield]

A zillion pixels from a cheap device -- kind of a "poor man's scanning back", eh?

It's an attractive approach, but I haven't heard much about it and I'd be interested to read any reports from people who have tried it.

I can imagine there are some difficulties with sensitivity and noise since scanners are typically designed to work with bright light. I can imagine some more difficulties with corner falloff because the scanner illumination is probably designed so that every pixel position sees light coming in at the same angle(s), while used as a camera, the angle of incidence will vary from center to corner.

But how bad those are in practice, I have little idea. The last time I tried it, several years ago, the particular scanner I was using needed so much light it was a lost cause for my application. But with a different scanner and a different application, the idea might work fine.

Got any good links?

--Rik

[Epidic]

As Rik states, illumination of the scanner head (both in terms of intensity and incidence) are a problem. I have seen results of a scanner used in the back of an 8x10 camera. It worked, but not really well. Most likely you are going to have to make your own driver. As far as exposure, if you can directly control the scan speed, you can slow it down to increase intergration time. You may also want to control the scanner gamma. One more thing, a scanning head uses a tri-linear raster, not a Bayer pattern. BTW, there are dedicated scanning backs for large-format cameras, so the idea is far from strange - http://www.betterlight.com/.
_________________
Will

[Graham Stabler]

I think I am getting enough light though to be honest I am at the minimum magnification, I'll see if I can easily disconnect the light source and try again, my scanner is only 1200dpi however.

Those betterlight scanning backs are neat and actually pretty simple.

Cheers,

Graham

[Graham Stabler]

OK I blocked a small section of the illumination with some black paper and got identical results so it is not some back reflection effect, what I suspect is that the poor results I am getting are because of the lens array used to get light from the top of the glass to the image sensor, I'm getting a very bright band across the image when the sensor is directly across the axis of the projection lens. The solution could be to just remove the lens array, I had a scrap scanner some time ago that had this technology and I think it was easy to remove. I may invest in a scanner to experiment further.

Graham

[Graham Stabler]

Just to double check I attached a piece of diffuse plastic sheet to the scanner bed, its actually a film used to print on to with a laser printer when making PCB masks and it works well as a screen to project on to. I get a fairly good image although it throws light away and at higher mag is a bit dark.

Graham