Image Processing Reference
In-Depth Information
Fig. 2.17
The effect of gray-level resolution. The gray-level resolution is from
left
to
right
: 256,
16,and4graylevels
A similar situation is present for the representation of the amount of incident
light within a cell. The number of photons hitting a cell can be tremendously high
requiring an equally high digital number to represent this information. However,
since the human eye is not even close to being able to distinguish the exact number
of photons, we can quantify the number of photons hitting a cell. Often this quanti-
zation results in a representation of one byte (8 bits), since one byte corresponds to
the way memory is organized inside a computer (see Appendix A for an introduc-
tion to bits and bytes). In the case of 8-bit quantization, a charge of 0 volt will be
quantized to 0 and a high charge quantized to 255. Other gray-level quantizations
are sometimes used. The effect of changing the gray-level quantization (also called
the
gray-level resolution
) can be seen in Fig.
2.17
. Down to 16 gray levels the image
will frequently still look realistic, but with a clearly visible quantization effect. The
gray-level resolution is usually specified in number of bits. While, typical gray-level
resolutions are 8-, 10-, and 12-bit corresponding to 256, 1024, and 4096 gray levels,
8-bit images are the most common and are the topic of this text.
In the case of an overexposed image, a number of cells might have charges above
the maximum measurable charge. These cells are all quantized to 255. There is no
way of knowing just how much incident light entered such a cell and we therefore
say that the cell is
saturated
. This situation should be avoided by setting the shutter
(and/or aperture), and saturated cells should be handled carefully in any video and
image processing system. When a cell is saturated it can affect the neighbor pixels
by increasing their charges. This is known as
blooming
and is yet another argument
for avoiding saturation.
2.4
The Digital Image
To transform the information from the sensor into an image, each cell content is
now converted into a pixel value in the range:
. Such a value is interpreted
as the amount of light hitting a cell during the exposure time. This is denoted the
intensity
of a pixel. It is visualized as a shade of gray denoted a
gray-scale value
or
gray-level value
ranging from black
(
0
)
to white
(
255
)
, see Fig.
2.18
.
[
0
,
255
]
