Robotics Reference
In-Depth Information
the size of the pupil can be controlled, corresponding to the “F” setting
on a camera, in order to control the amount of light that enters the eye.
The lens of the eye inverts the image of whatever the eye sees, and this
inverted image impinges on the retina, where it is detected by millions of
receptors known as
rods
and
cones
. The rods pick up low levels of light
and are connected to the optic nerve, while the cones are more active at
higher levels of light and enable us to perceive colour.
Research into computer vision, also known as machine vision, began
in the 1940s when television was in its infancy. The idea of connecting a
TV camera to a computer led to the invention of a crude image scanner
during the 1950s, and later, in 1960, to the first recognition program
for hand-written characters, based on an automated learning mechanism
called perceptrons.
1
Computers see by means of cameras that replicate
the functions of the human eye. The type of camera most commonly
employed in computer vision systems nowadays is called a CCD camera,
for Charge-Coupled Device, an electronic device that converts light into
electrons and then counts these electrons to measure the levels of light
in the image.
2
Light enters a CCD camera through the lens and is pro-
jected onto a CCD chip, which takes the role of the retina in the human
eye. The CCD chip typically consists of one million or more individual
light detectors (pixels) that are arranged in a matrix pattern, rather like
a gigantic chessboard in which, instead of eight rows and eight columns,
there might be 1,000 or more of each. These are the numbers that are
referred to when talking about the resolution of a TV screen, a computer
monitor or a similar electronic display.
The CCD chip and its associated electronic circuitry count the num-
ber of electrons within each pixel. Each number is stored, so the whole
image can then be represented by a series of numbers, with one number
usually corresponding to each pixel. In Figure 28, for example, the image
of the fish's eye has a dark area along the left hand edge, corresponding
to the 40, 38, 38, 38, 38 in the leftmost column of digits. Note that at
the lower end of this column the numbers are slightly lower (36, 35, 35),
corresponding to the lighter shade on the image, and in the top right-
1
See the section “Artificial Neural Networks” in Chapter 6.
2
A charge-coupled device is a computer chip containing an array of linked (or coupled) electronic
capacitors. When light impinges on the surface of a CCD, some of the electrons in the surface gain
energy from the light and become free to move around, congregating in small, image-forming units
called
pixels
on the surface. These electrons accumulate in the capacitors on each pixel, where they
can be counted by an electronic circuit. The count of electrons in a pixel indicates how much light
has impinged on that pixel.
