
I took this panoramic photo of my house at night one weekend. I'm always disappointed at the level of quality produced by my camera at night-time. Zooming in to the darkest part of the sky produces this image:

It kind of reminded me of the Bayer pattern, used in the pixel layout of most digital cameras. The idea is that it's cheaper to produce a camera that doesn't need 3 detectors per pixel (one for red, green and blue). Instead, one sensor is used per pixel in the following pattern:

So the top left pixel only measures the amount of green falling on it. The red and blue values are computed by interpolating between the neighbouring pixels that did measure red/blue.
I wonder if the effect seen in the dark areas of the image is a direct result of the structure of the imaging elements of the camera. It could be worth investing in a 3-CCD camera, one that uses 3 sensors per pixel, as it might reduce these artifacts. It would be most relevant when the images were to be printed out full-size.
No comments:
Post a Comment