Charge-Coupled Device

All CCDs work on the same principle. The CCD surface is a grid of pixels (pixel is a contraction for "picture element"). Small CCDs may have a grid of 256 x 256 pixels, while large CCDs may have 4,096 x 4,096 pixel grids. Although many CCD pixel grids are square, this is not always the case; scanners and photocopiers, for example, have a single line of pixels that passes over the picture or page of text being imaged. The pixels are tiny; some CCDs have pixels only 9 microns across, while others may have 27-micron pixels. The scale and resolution of the image a camera is able to form on the CCD depends both on the pixel size and the grid size. Regardless of the pixel or grid size, however, each pixel on the CCD has the ability to convert the light striking it into an electric signal. The voltage accumulated by each pixel during an exposure is directly proportional to the amount of light striking it. When the CCD is exposed to light for a length of time, an image of whatever is being observed—whether a distant galaxy or cars in a parking lot—forms on the CCD as an array of differing electric voltages.

After an image has been recorded on the CCD, the device can be "read out," meaning that the voltages are extracted from the CCD for storage on a computer. The analogy that is almost universally used to describe this process is the "bucket brigade" analogy. Picture each pixel on the CCD as a bucket with a certain amount of water in it. When the CCD is read out, the water in each row of buckets is emptied into the adjacent row. The water in the first row goes into a special row of storage buckets, the water in each bucket in the second row goes into its neighbor bucket in the first row, and so on across the whole CCD. Then, the amount of water in each of these buckets is emptied, measured, and stored in a computer's memory. This process is repeated until all of the rows have been shifted into the storage buckets, emptied, and measured. If you now replace the water with electric voltages, and replace the measurement of water with the digital measurement of the analog electric signal, you have the basic process by which an image is extracted from the CCD. The actual process of reading out the CCD is performed by fairly complicated and exquisitely synchronized electronics that move all the electric charges between the "buckets," convert the analog voltages into digital numbers, and make the data available for storage on a computer.

Once the pixel outputs have been measured and stored on a computer, they can be used in a variety of ways. For simple line drawings, the image processing software may render the data from the CCD in black and white. For pictures, a 256-level grayscale may be appropriate. In either case, a grid of numbers, corresponding to the original light intensity, is present and can be analyzed in any way the person studying the image desires.

From the description above, it may seem that CCDs cannot be used for color imaging, since they respond only to light intensity. In fact, color CCDs are available, although they are used in video equipment such as camcorders and almost never in astronomy. If an astronomer wanted to create a color image using a CCD, the old practice of taking three images through three different color filters is still the usual way to go. True color CCDs have pixels with built-in filters, alternating red, green, and blue. They can produce real-time color images, but they are undesirable for scientific work because they introduce significant difficulties into the data analysis process, as well as reducing the effective resolution of the CCD by a factor of three.