3.4 Input and output considerations
CCDs are not inherently able to detect colour, only brightness. So it is necessary to rely on the fact that any colour of light can be made up from the three primary colours of light: red, blue and green. (Note that the three primary colours of light are different from the three primary colours of pigments.) Each CCD in the array is therefore overlaid with a red, blue or green filter and so detects the brightness of, respectively, the red light, the blue light or the green light falling on it. The filters are arranged in a mosaic pattern, and later processing has to recombine the outputs of groups of CCDs to arrive at the colour of the light in that general area. Figure 9 illustrates this idea for a very small array of CCDs.
You will notice that in Figure 9 there are twice as many green filters as red or blue ones (because the human eye is not equally sensitive to the three primary colours of light). So in fact it is the outputs of groups of four CCDs which must be combined to find the colour of light in that general area, as indicated in Figure 9.
Each CCD in the array produces an analogue electrical output that corresponds to the brightness of the filtered light falling on it. Each CCD's output has to be converted to digital form by an A-D converter, and then groups of outputs are processed to arrive at the original colour of the light. Finally, the overall colour and brightness of each pixel is encoded.
You might imagine that the number of pixels in the image will correspond to the number of groups of four CCDs in the array, but actually this is not necessarily the case. First, by averaging values obtained for different groups of four CCDs it is possible to arrive at a one-to-one correspondence between the number of CCDs and the number of pixels, as Figure 10 illustrates. And just to complicate things further, it is possible to create an image with more pixels by ‘guessing’ intermediate colour and brightness values between adjacent CCDs.
From the foregoing you will be able to see that there is a great deal of processing associated with the production of a bit-map image from the CCD array's output – and I have not even mentioned the compression that will take place. For this reason many digital cameras, including the one described in this course, contain a second processor. This second processor is a digital signal processor or DSP, which is a processor specially designed for repetitive yet demanding tasks like image processing.
So far as output is concerned, the camera provides a screen on which the user can view photos already taken, and also the shot about to be taken. The output process of showing images on this screen works as follows. The screen, which is backlit, consists of an array of tiny liquid crystals that can be made transparent or opaque in response to electrical signals. It is therefore known as a liquid-crystal display or LCD. To make the display coloured, the liquid crystals are grouped in threes, one with a red filter, one with a blue and one with a green. The LCD's associated output subsystem receives colour and brightness data from the processor and sends appropriate electrical signals to control the transparency of the liquid crystals on the screen.
Box 8: CMOS image sensors
The digital camera discussed in this block uses an array of CCDs to capture the colour and brightness of the image. An alternative method of capturing the colour and brightness is to use what is known as a CMOS image sensor. These sensors are made using the same technology as silicon chips and can therefore be integrated onto the same chip as a processor, which can be very convenient in small portable items such as cameras.
CMOS image sensors consist of arrays of light-sensitive cells, and in some cases groups of cells are used to detect red, blue and green light (much as for CCDs). One manufacturer, however, has patented a process which has a big advantage over CCDs: each individual cell can be made to detect the brightnesses of all three of red, blue and green light. This means that no averaging processes for groups of cells are needed for this manufacturer's CMOS arrays. Against this advantage must be set the disadvantage that more complex circuitry is required to detect the outputs of the cells.
As well as being used in digital cameras, CMOS image sensors are used in some mobile phones and web cams.