Information Technology Reference
In-Depth Information
Fig. 5.12 Recovery of images and removal of non-Gaussian noise by the Belousov-Zhabotinsky
medium functioning in the excitable mode: (a) removal of non-Gaussian noise, (b) recovery of the
image
recovery of an image in which random defects have been formed, and the removal
of non-Gaussian noise clogging the image.
Note that in the excitable mode of the functioning of the medium, determining
the skeleton of the image was possible for only a limited class of extended images
formed by sufficiently thin fragments. In general, this operation is executed by a
medium that operates in the trigger mode. Numerical simulation of this operation is
shown in Fig.
5.6
. Unfortunately, a medium of the Belousov-Zhabotinsky type in
the trigger mode is not sensitive to light radiation, which does not allow to
determine experimentally the skeleton of a figure of an arbitrary shape.
A more complete comparison of the results of image processing by the method
of mathematical morphology and the evolution of the images in a reaction-diffu-
sion medium of Belousov-Zhabotinsky are given in Table
5.1
. The primary con-
clusion to be drawn from this comparison is that the results of the image evolution
in the medium actually coincide with the operations of image processing. The basis
of this is apparently constituted by nonlinear mechanisms inherent
to both
approaches.
Nevertheless, differences between these approaches should be emphasized,
although they do not affect the main conclusions.
The method of mathematical morphology allows working with a variety of
structural elements of different shape. This makes it possible to fix the more subtle
features of the image being processed, which is its distinct advantage.
At the same time, the complexity of the behavior in the Belousov-Zhabotinsky
media is very high. On the one hand, as will be shown below, this may result in
image processing artifacts that do not reflect the actual structural characteristics of
the image. To avoid this,
the mode of image processing by the Belousov-
