Question Flatbed scanner for astronomical photography

[giodec]

Hi all. I am an amateur astronomer since 1980, with an interest in recycling / modifying existing low cost technologies to convert them for the study of astronomy.


I am thinking about using a CCD flatbed scanner (in the case an Epson Perfection 1650) in slide copying mode (i.e. without the scanner light on and removing the lid) to get high resolution images of the Sun , Moon and planets.

I want to do this building a support for the scanner in order to project and focus the image of the heavenly bodies , obtained through a telescope, over a ground glass fixed over the flatbed scanner glass. The whole "camera" is to be fixed to a telescope in order to let it follow the celestial object along its apparent movement over the sky.

In the case of the Sun (filtered) and the Moon there should be enough light to let the scanner register a good image in one scan only . On the other hand in case of the major planets it could be necessary to shoot more scans and then stack them one over the other in order to get a visible high resolution image, as the images of planets like Mars, Jupiter and Saturn , when enlarged 200 or 300 x, are relatively dim.


I would need an help from an expert in scanner control software in order to have the possibility to automatically repeat the scan over the same image , and if possible to modify other embedded parameters which cannot be modified through the standard control software of the scanner (I use Vuescan on Win 10 64 bit OS).

There are no electronic, mechanical or optical problems I cannot manage in this project, I only need help on the software side. Any advice is welcome. Best regards Giovanni , Italia

[Note: Moderator edit to break up one solid paragraph of text.]

 

[Eximo]

Interesting idea, but I think the answer here is that you would be better off using an actual camera with an extremely long exposure setting. Most flatbed scanners are pretty low resolution compared to cheap and readily available camera CCDs. Expensive high res scanners can do about 21 MPixels, which is on-par with low end DLSR cameras. They should have settings to allow for very long exposure times, and all the software/hardware is self contained.

I know it goes against what you are trying to do, but it would be a heck of a software project.

I can think that it could only work by forcing the scan cycle to run continuously, save each file, and then use software to combine them. Photoshop should have a smart objects and stacking tutorials that might do what you are after. Not sure how effective it would be.

But some other mathematical way to continuously add the light data collected is possible, but it might just blow out after a few cycles if done simply.

 

[giodec]

Thank you Eximo for your detailed answer. Well my scanner has a maximum resolution of 1600x3200 ppi, with a scanning surface a little bigger than an A4 sheet; so considering a (reduced) scanning area of 8x11 inches at 5Mp per square inch we have an humongous total of 450.56 Megapixel, Please correct me if I am wrong.
Apart the not too bad resolution and low cost of older models , there are other interesting aspects of getting a scanned image instead of a single shot in astronomy. For example using a fine grating it should be possible to selectively scan the h alpha line of the solar spectrum in order to get a very nice h alpha image of the Sun. Another use of the scanner could be to get an hi res image of solar spots by simply projecting the unfiltered image of the Sun taken through the telescope over the flatbed glass covered with a fine grain ground glass.
Think about giving it a try...

 

[Eximo]

Thank you Eximo for your detailed answer. Well my scanner has a maximum resolution of 1600x3200 ppi, with a scanning surface a little bigger than an A4 sheet; so considering a (reduced) scanning area of 8x11 inches at 5Mp per square inch we have an humongous total of 450.56 Megapixel, Please correct me if I am wrong.
Apart the not too bad resolution and low cost of older models , there are other interesting aspects of getting a scanned image instead of a single shot in astronomy. For example using a fine grating it should be possible to selectively scan the h alpha line of the solar spectrum in order to get a very nice h alpha image of the Sun. Another use of the scanner could be to get an hi res image of solar spots by simply projecting the unfiltered image of the Sun taken through the telescope over the flatbed glass covered with a fine grain ground glass.
Think about giving it a try...

I don't believe it works that way. The scanner works by stitching an image together, exposing parts of the image to capture. Take a look at what a flatbed scanner CCD actually looks like. A thin horizontal line. As the scanner moves along the target document it grabs a very narrow section. The megapixel value really only relates to the amount of data it can store in memory and assemble before offloading to external storage.

In your scenario you would have to set your telescope up with a narrow aspect similar to the normal scanning area exposed and move along at the same rate as the typical scanning cycle. Getting that to line up and be repeatable would be an engineering miracle. Or you would just have to have the sensor be fixed and make sure the image you present to the sensor moves at the expected movement rate it would have had.

If you are looking at the same spot of the sky, no matter your configuration, you only have the one sensor area to work with. You could do the same thing and stitch an image together, but you could probably do that easier with a camera once again.

The more I look at this the more it just makes more sense to do the same concept with a regular camera CCD. It will have more sensor area and mean that movement is less important. An interesting project, but I imagine if you went to a manufacturing supplier and just started picking up raw CCDs and figuring out how to run them you would be better off then starting with a scanner.

 

[giodec]

"In your scenario you would have to set your telescope up with a narrow aspect similar to the normal scanning area exposed and move along at the same rate as the typical scanning cycle. Getting that to line up and be repeatable would be an engineering miracle. Or you would just have to have the sensor be fixed and make sure the image you present to the sensor moves at the expected movement rate it would have had."


No , it is not intended to work that way. If you look on Youtube there are several videoclips about converting a scanner in a large format camera; they project the image of the subject to be photographed taken with some kind of lens capable of giving a virtual image of a subject (someone uses a simple biconvex lens, others some kind of photo lens) over a ground glass which substitutes the original scanner's transparent glass. The image is then scanned by the sensor and integrated via the scanner's proprietary software to the PC. The resulting image has (in theory) a lot more details than an image taken by a standard digital camera (whose sensor is 24x36 mm. at best) . The principle is exactly the same of using a large format camera as a Linhof or else instead of a 35 mm. camera.