Page 180 - Intelligent Communication Systems
P. 180
7
" ~ CHAPTER 12 COMPUTER VISION 151
/
12.1.3 Quantization of Gray-Level Value
To represent an analog image by discrete gray-level values is called quantization
of the image. For example, when the brightest gray-level value is "000" and the
darkest is "111," this is called a quantization of eight levels. In other words, it is
eight gray-scale representations. Conventionally, 64-256 gray-level representa-
tions are used for quantization.
12.1.4 Gray-Scale Binary Image Display
If the number of pixels in a region of a gray-scale image is fixed, then the display
with binary data, 0 or 1, will be represented by the least amount of data. The fea-
tures of an image can be preserved by proper conversion to its binary image. When
the image is binary data, it can be processed easily.
12.1.5 Threshold Processing
There are a variety of methods for obtaining binary data. One of the easiest uses a
gray-level histogram. Consider an image composed of an object and a background.
In the density histogram, the gray-level value where there is a large change in
the number of pixels is called the threshold. (An example is shown in upcoming
(Figure 12.4.)
In the second method for obtaining binary data, when the number of pixels of
an object is known and the gray level of the object is quite low, the threshold is
obtained from the ratio P of the number of pixels of the object to the total number
of pixels in the image.
12.2 IMAGE DISPLAY
12.2.1 Density Distribution
The distribution of gray-level values in an image with MxN pixels is called the
density histogram. It is not good for image processing if the density distribution is
biased. Therefore it is necessary to make the distribution uniform by changing
either the light source or the direction of the lens.
12.2.2 Density Transformation
In order to get a good-quality image, a full-scale gray-level display is desirable.
To achieve this, the density transformation method is used. As shown in Figure 12.2,
density transformation is achieved via line transformation.
