Two of my favorite hobbies are computers and photography, and tonight I was able to do both of them. I have been learning how to use image averaging for the purposes of noise reduction and increasing the dynamic range of images. (Kenobi: "I suddenly feel a disturbance in the Force like the sound of many eyeballs suddenly glazing over in unison.") Digital cameras are wonderful creations, but they have some limitations, especially in low-light situations. A picture will tend to have more noise in it if a faster ISO speed is used, or if the shutter is open for a longer duration. One way to compensate is to take several identical photos, and then average them out. Since noise is random, the image will remain clear, but the noise pixels will fade out.
To give you an example of this, here are a few pictures. This first image is a single frame of a nighttime scene (the view out my living room window):
This second image is the result of combining eight frames and averaging them out:
At such a small size, the pictures look fairly similar, although the tree leaves in the second picture are a little bit clearer. So in order to show what I'm talking about, here is an enlarged segment of the original picture:
If you look closely at this image, you can see the noise pixels in the tree trunk. As a comparison, here is an enlarged section of the "averaged" image:
The tree trunk has a significantly improved appearance, and the window screen is very defined. The foliage near the base of the tree doesn't look so odd either.
The main reason for using this technique is providing increased clarity to images that are taken under less than ideal conditions. For example, night scenes can be enlarged and still look remarkably clear. The downside is that multiple
identical images must be taken, which requires a good tripod. A multi-frame mode on the camera is also useful. But this technique could come in very handy in photographing landscapes at night.
For those of you who haven't completely glazed over yet, and are interested in how this is done, there is a good article on the
Cambridge in Colour web site.