Each one of those pixels is a number. Its value represents the brightness of that exact location on the telescope's focal plane. For my Cookbook camera, there are 4096 possible levels represented as 10-bits (ones and zeros) for each of 378 x 252 pixels, not many at all by today's standards. For color imaging with a webcam, each pixel is a triplet of numbers, commonly representing how much in each of three colors, red, green and blue.

Digital image processing seems magical to me. Individual short exposures are added to be the equivalent of one much longer. Noise can be subtracted out, while unevenness in sensitivity and illumination can be divided away. Whole images can be rotated, stretched, aligned and combined to reveal movement (asteroids or a new comet?) or changes in brightness (variable stars? novae?). Images are just raw data, and visualization techniques can be used to reveal insights otherwise invisible. Aesthetics can be served as well. Techniques like those pioneered in the dark room by David Malin, unsharp masking and tri-color composites, are now available at a mouse click.

Over the first ten years of the digital revolution in astroimaging, the most remarkable advances have been in software. Not that the equipment hasn’t evolved—it has, from tiny CCD's to mega-imagers, video chips, webcams and digicams, but without software to crunch and display, it's still just a bunch of numbers.

At right are some of my better Cookbook shots of deep sky objects depicted here in a different way. They are plotted in 3-D as if brightness represents height or topography. After each had gone through basic image calibration and cosmetic enhancement, further processing provided a three-dimensional representation that can be colored, shaded and illuminated as if it were a real, 3-D object. Peaks are bright spots (mostly stars), while dark regions are valleys and low areas. Finally, this is all collapsed back into numbers for delivery to your browser’s screen. After all this processing and calculation, can you recognize any of them? (All three are famous Messier objects: place your mouse over the colored image for a look at the original.)

To MATLAB, an image is just a matrix of numbers. Moondark is written by Doug Miller, published on the web, and printed in the Delmarva Star Gazers' Star Gazer News. This document was last revised on 26 October 2002. All text and images copyright © 2002 Douglas C. Miller, All Rights Reserved. This material may not be reproduced in any form without prior permission.