|
Image Processing GuideIf you are taking digital pictures (or even if you're taking photos with film and then scanning them in), there's a lot you can do if you're prepared for the final output to be a screen or a printer. For the purpose of this page, we'll call the picture you took, your light frame. The techniques described here are also used for astrophotography with CCD cameras. The problem of taking digital camera photos at night and taking astro-photos is very similar. Noise ReductionNoise reduction can improve the signal to noise level dramatically. To reduce noise, you need a valid dark frame for each shot you take. Most digital camera noise (CCD) is caused by thermal noise. The colder the CCD, the lower the noise. Luckily most thermal noise is reproduceable with the same image taken with the same exposure. To take a dark frame you do the following:
You now have an image with noise and a picture (the light frame), and an image with the noise only (the dark frame). You subtract the dark frame from the light frame and you get rid of most of the thermal noise. Try it, it makes BIG difference. Further Noise ReductionAnother technique for reducing noise is taking a Bias Frame. The Bias, is the amount of current that is read no matter what the image is for a given pixel on the CCD. This varies from pixel to pixel. Though removing this isn't anywhere as impressive as using a dark frame on it's own, it still reduces an appreciable amount of noise. To take a Bias frame, you need to:
You can now do: light frame - bias frame - dark frame Flat FieldingFlat fielding can remove the effects of vignetting, dust specs and other lens defects. To do this, you need a uniform background that isn't too bright. A clear sky at dusk is the best. Take a normally exposed picture. Then you'll need to convert it into a Mask in your favourite image processing software. You can then overlay the mask on your picture to even out the picture. StackingStacking can make multiple exposures of the same subject similar in exposure to one long exposure. To do this, take multiple pictures and add them together (noise removal should be done on each one first). Stacking is very good for static objects, not too great for a fast moving aurora, but if it's just a static diffuse glow, it's worth a try. Additional InformationThis is just a very general overview on image processing. To find out more and how to do it with your software, check out the links below. More links to come. |
||||||||||