Abstract. In
practice, it often happens that different objects on the same image appear with
varying degrees of distortion (for example, because of defocusing due to being
location at the different distances from the receiving device). In such cases,
there are difficulties in the real image recovery, since the distorting
hardware function (Point spread function (PSF)) is not the same across the
entire image. Therefore, when working with images on which areas are distorted
in various ways, as a rule, the original improved image is divided into parts
(sprites), which, presumably, were affected by one distorting hardware function
(Point spread function PSF). And after that, autonomously for each part
(sprite), they make a reconstruction, and as a result, they fold the restored
fragments into a general image.
This
work presents a technique for automatically determining fragments on spectrally
distorted images that are promising for deconvolution. The proposed procedure,
using a floating sprite, pixel-by-pixel, calculates the image recovery
coefficient Cri (an abbreviation of Coefficient of recoverability image). Cri
has the physical meaning of estimating the fraction of the amplitude spectrum
of the image under study, located in a given neighborhood of the universal
reference spectrum ((URS) or, same General-purpose reference spectrum (GRS)).
This makes it possible to reduce the influence of the human factor on the
choice of location and position of fragments that are promising to improve
their quality. The technique has been tested both on model and real images.
With its help, regions (fragments) are defined reliably as undistorted, and
with distortions of different degrees.
Key words: spectral
distortion, coefficient of
recoverability image, fragments distorted to varying degrees.
References
1.
Gonzalez
R.C., Woods R.E.. Digital image processing. 2nd ed. published by Pearson
Education, Inc, publishing as Prentice Hall. 2002. ISBN: 978-0-2011-8075-6.
2.
Yagola A.G. , Koshev N.A.
"Restoration of blurred and defocused color images". Vychislitelnye
metody i programmirovanie - Computational
methods and programming. 2008, Vol. 9, pp.207-212. (In Russian)
3.
Zrazhevsky A.Yu., Kokoshkin A.V., Novichihin E.P., Titov
S.V. Improvement Quality of Radio Images.
Nelineinyi mir - Nonlinear world. 2010.
No. 9, pp. 582-590. (In Russian)
4.
Gulyaev Yu.V., Zrazhevsky A.Yu, Kokoshkin A.V., Korotkov V.A.,
Cherepenin V.A. Correction of the spacial spectrum distorted by the optical
system using the reference image method. Part 2. Adaptive reference image
method.
Zhurnal Radioelektroniki - Journal of Radio Electronics, 2013. No.
12. Available at:
http://jre.cplire.ru/jre/dec13/2/text.html (In Russian)
5.
Kokoshkin A.V., Korotkov V.A., Korotkov K.V., Novichihin E.P. “Using
the method of renormalization limited to restore the distorted image in the
presence of interference and noise with unknown parameters.”
Zhurnal Radioelektroniki - Journal of Radio Electronics, 2015. No. 7, Available at:
http://jre.cplire.ru/jre/jul15/4/text.html
(In Russian)
6.
Kokoshkin A.V., Korotkov V.A., Korotkov K.V., Novichihin E.P. Restoration
of the images consisting of fragments with various degrees of defocusing. JZhurnal
Radioelektroniki - Journal of Radio Electronics], 2015, No. 10. Available
at: http://jre.cplire.ru/jre/oct15/6/text.html (In Russian)
7.
Kokoshkin A.V., Korotkov V.A., Korotkov K.V., Novichihin E.P. “The
method of predicting possible improvements in the quality of distorted images.”
Zhurnal Radioelektroniki - Journal of Radio Electronics, 2015, No. 6. Available
at:
http://jre.cplire.ru/jre/jun15/5/text.html (In Russian)
8.
Gulyaev Yu.V.,
Zrazhevsky A.Yu, Kokoshkin A.V., Korotkov V.A., Cherepenin V.A. “Correction of
the spacial spectrum distorted by the optical system using the reference image
method. Part 3. General-purpose reference spectrum.” //
Zhurnal Radioelektroniki - Journal of Radio Electronics, 2013.
No. 12. Available at: http://jre.cplire.ru/jre/dec13/3/text.html (In Russian)