Peripherals: Video cameras and scanners

Video cameras and scanners

Video cameras are versatile devices, being able to capture images of any type, including solid objects. Scanners are limited to images on paper, but they are able to scan each spot on the paper with much greater accuracy than cameras, and so are more widely used for this type of input. Scanners provide a low-cost way of inputting material that's been typed or printed on paper into a computer system. (I say 'low cost' because rekeying an A4 page costs between £4 and £5.) More was said on this earlier in this chapter on page 58.

Most scanners incorporate a special sort of camera made up of charged-coupled devices (CCDs). Each CCD receives light from the image, and, provided the light is strong enough, will generate an electrical charge. This means that light areas or 'dots' of the image are represented by charged cells, and dark areas by uncharged cells. As the paper containing the image moves past the camera during the scanning process, these charges create electrical impulses which are fed into the computer, where they are interpreted by the scanning software as parts of the image.

The resolution of the typical scanner is 300 dots perincli, which means that it splits each square inch of the image up into a matrix of 300 x 300 tiny areas. This is better than the resolution of most computer screens, and the same as the resolution of most laser printers (see later in this chapter). Scanners typically cost between £1,000 and £2,000.

The technology behind this is well understood and quite straightforward. Nowadays, scanners are widely sed to get drawings, diagrams, and photographs into computer systems for incorporation into documents and books which are made up electronically prior to printing. Much more difficult is the recognition of the image by the computer, so that it is able to act upon what it 'sees'. Just as the speech-recognition system described in the previous section worked by breaking down the spoken word into its phonetic elements, so image­ recognition systems work by breaking down the image into component parts, identifying each and analysing their position relative to each other.

Image-recognition systems have been used for some years in industrial robots, but these devices have a limited 'vocab­ulary' of components, and the recognition process is highly complex and so relatively slow. It seems likely that the application of the transputer and parallel processing tech­niques will revolutionize image processing and image recog­nition in the future. This may lead not only to a new generation of industrial robots but also to new computer systems and applications which are able to perform tasks such as converting hand-drawn sketches into neat designs, and recognizing faces in a crowd.

Other input devices Other input devices include:

• Kimball tags, used for stock control in some clothing stores. Attached to each article of clothing is one of these tags, recording data on the article as a pattern of punched holes. When the article is sold, the tag is removed and the data scanned into the store's computer.

• Bar codes, used for stock control in food stores. Here, the data is stored as a pattern of thin and thick lines printed on the product's packaging; the thin lines repre­ sent binary Os, the thick lines ls. At the checkout a light pen or similar device shines light across the pattern, the reflected light being translated into electrical pulses for computer input.

• Magnetic ink character readers (MICRs), used in bank­ing. Cheques are identified by numbers printed with special magnetic ink, which can be read by an MICR and converted to binary digital form for computer input.

Storage devices

Because computer memory can hold only a limited amount of data and programs, and because, in the case of semicon­ductor RAM, it loses that data when the power is turned off, some form of long-term mass storage is essential. At the present time, magnetic disk is the main mass storage medium, though other types of mass storage are also used, in particular the optical disc described later.

Like the ordinary audio disc, a magnetic disk stores information on circular tracks marked out on its surface. As the disk rotates, an arm moves a read/write head to the required track, and 'reads' (i.e. retrieves) data from or 'writes' (i.e. stores) data to the spots on the track as they pass below it. The data is stored in digital form, a magnet­ized spot on the rotating surface representing a 1, a demag­netized spot a zero. As with an audio-cassette, information can be erased and re-recorded any number of times.

Magnetic tape can also be used in certain computer applications, but it is less versatile than disk, since it is not possible to move instantly to a particular spot on the tape to read or write data at that spot. Reading from the tape is rather like reading a novel: you must start at the beginning and read each word in sequence until you reach the end. This is called sequential access.

Reading from a disk is more like reading from a reference book: you look up the location of the section you require in an index and turn straight to the appropriate page. This is called random access. Each program or file of data on the disk must be assigned a name, and an index of these names is stored at the beginning of the disk. The location of the file on the disk's surface is also stored in an index, and when the computer receives a command to load a file into RAM it is able to read this location from the index and move the read/ write head directly to the required track on the disk's surface.