Image Processing Reference
In-Depth Information
1.2 Image Sensor Output and Structure of Image Signal
In this section, the structure of an image signal captured by image sensors is discussed.
1.2.1 Monochrome Still Images
As discussed in Section 1.1, the only image information required for monochrome still
images is light intensity and space (position). The basic device configuration of image sen-
sors is shown in Figure 1.4. Image sensors have an
image area
on which optical images
are focused and are converted to image signals for output. In the image area, unit cells
called pixels are arranged in a matrix in a plane. Each pixel has a sensor part typified by a
photodiode, which absorbs incident light to generate a certain quantity of signal charges
according to the light intensity. Thus, the light intensity information for a pixel is obtained
at each sensor part. Figure 1.5b shows a partially enlarged image of Figure 1.5a, and the
image signal of the same area captured by image sensors is shown in Figure 1.5c. The
density of each rectangular block is the output of each pixel and the light intensity infor-
mation
S
(
x
i
,
y
j
) at the coordinate point (
x
i
,
y
j
) of each pixel. Expressing a two-dimensional
coordinate point (
x
i
,
y
j
) using
r
k
, the signal can be written as
S
(
r
k
). A set of
S
(
r
k
) at all
r
k
in the
image area makes up the information for one monochrome still image.
Dividing the image area into pixels that have a finite area, as shown in Figure 1.6, means
fixing the area size and coordinate point at which light intensity information is picked
up. The figure also shows that the continuous analog quantity, representing the position
information, is replaced by discrete coordinate points. That is,
x
i
and
y
j
cannot take arbi-
trary values, but are built-in coordinate points that are fixed in imaging systems, which
are the digitization of space coordinates. Using digitization in this system, since two-
dimensional space information can be treated as determinate coordinate points, only light
intensity information should be obtained. Therefore, the three-dimensional information
of a monochrome still image is compressed into one dimension. Since the overall total
Sensor part
(photodiode)
Image area
Pixel
Output part
Scanning function part
Image sensor chip
FIGURE 1.4
Basic device configuration of image sensors.
