ConceptsSince I have a tendency to write one-shot software to handle different images, I seldom process any 2 images in exactly the same way. For this reason, this section is not going to be a paint-by-numbers description of the process involved in producing an image from raw data. Instead, I hope to give enough of an overview of the ideas involved to allow you to see what's going on, and perhaps to do it yourself.
Dark FramesIn the Shooting section, I mentioned capturing dark frames when images of a target are taken, but didn't describe what a dark frame is. When an image is captured using a CCD device, there are a couple of things that can cause a pixel to be "lit." The first and hopefully dominant cause is light hitting the detector. In addition to this, though, a fair amount of "noise" is present as well, dependent on the exposure length and the temperature of the CCD. Expensive CCD cameras are cooled in order to reduce this noise, but a QuickCam is not. Fortunately, over the course of a few minutes, this noise is pretty consistent. Since the value of a given pixel is the sum of the effects of the light hitting it and the noise, it is possible to capture an image when no light is present and then subtract this "dark frame" from the images taken of actual targets, effectively removing the noise from them.
For images of bright planets, dark frames aren't really that big a deal, though they should still be used, since every little bit helps. When dealing with long (for the QuickCam, anyway) exposures, however, dark frames are crucial. Here is an example:
Unfortunately, reducing the image sizes for the Web and JPEG compression have smoothed out a large amount of the noise. You can still see that using dark frames can make a tremendous difference.
Stacking Multiple ExposuresAnother important concept in CCD imaging involves "stacking" multiple exposures on top of one another in order to improve the signal-to-noise ratio of an image. This technique is used by the big boys with expensive gear, but is even more important when using a cheap camera and the ETX's relatively unsteady clock drive.
The idea is pretty simple: If you have two images of the same object in which it appears on different places on the CCD chip, and you lay one on top of the other after moving it around until the target is in the same place on both frames, the noise at any given location on the final image will come from two different places on the CCD, and so will have no correlation, while the actual target images will tend to reinforce one another. If you do this with enough frames, the noise will tend to cancel itself out, and the target will be enhanced. Here is an example:
In addition, because the result of all of this is a smoothing of the image, applying sharpening techniques to the stacked image will result in fewer spurious "details" caused by noise. Here is an example of how much sharpening can be applied to some real stacked images These images are probably a little over-processed, but show the point.
|Website © 1999-2000 Emm Graphics|