
7 Binarization and Color Separation
103
All the parameters of this dialog are set using the sliders or by typing appropriate
values in corresponding fields.
1. Set the
Radius
value
–
the radius within which the program will be
analysing the level of noise to be cleaned. The greater this value, the more
pixels surrounding the color transition boundary will be analyzed.
2. Select the
Smooth
checkbox. Using smoothing with the
Binarize
checkbox
set to off produces a cleaner image with the background evened and line
objects revealed.
3. Set the
Binarize
checkbox to create a monochrome image.
4. Click
Ok
after achieving satisfactory results in the preview window.
Color Separation
A real map or color diagram is usually made up of a small number of colors.
However as a result of scanning a paper original we can get a color raster image
having tenths or even hundreds of thousand of colors.
Two similar procedures are described in this section that make possible to separate
color image dots in non-overlapping sets, i.e. categories. These procedures are
used to extract the colors the original image was created with. The objects of one
sort are usually marked with the same color; therefore we are able to separate
necessary image objects.
Each category is based on a set of basic colors. A set of basic colors that belong to
all categories defines the separation of original image dots in non-overlapping basic
subsets.
To distribute colors into subsets WiseImage performs the following procedure. The
color difference between each dot of a color image and all specified basic colors is
calculated. This dot is put in a subset of basic color that has the minimum color
difference from the color of this dot in the RGB area. Thus, all the original image
dots are split to basic subsets, related to the specified basic colors.