Robust algorithm of space objects images in a picture of low quality

Аuthors

Shigaev A. K.^1*, Luchkin R. S.^1**, Mashkin M. N.^2***, Tyurin V. S.^3****

1. Korporatsiya "Vympel", 3, 4th ul.8 March, Moscow, 125319, Russia
2. Moscow Aviation Institute (National Research University), 4, Volokolamskoe shosse, Moscow, А-80, GSP-3, 125993, Russia
3. Research Center of rocket and space technologies of Russian Academy of Cosmonautics named K.E. Tsiolkovsky, Pionerskaya str.,16, Korolev, Moscow region, 141074, Russia

*e-mail: MSnakeDT@gmail.com
**e-mail: romluchik@rambler.r
***e-mail: mnmashkin@ya.ru
****e-mail: vst.17@mail.ru

Abstract

The process of restoring the space object localization area in a low-quality picture implies a choice between two below-mentioned hypotheses.

1. Background emission is fixed

2. Object image is found in the picture

In the context of the second hypothesis the space object localization area is selected.

The transition from background to object or vice versa is followed by brightness jump located in the area of background/object border, provided that the background spectra and the hiss spectra are concentrated in a low-frequency region. Thus, we can find the transition from background to object using this assumption. As a result of the algorithm operation the found area of an expected localization of the object image is cut out from the original picture in the form of a square the size of which is multiple of the power of 2.

Definite quality characteristics of solution to the problem are recommended in order to estimate the algorithm efficiency and an informative value of the obtained results.

In the course of the work we’ve synthesized the robust algorithm of selection of a distant object of unknown shape in a low-contrast picture formed in the atmospheric vision system. The algorithm is stable in the context of a probabilistic description of complicated background conditions. A decision-making criterion concerning the image presence in the picture is recommended and the structure of parallel three-channel background information processing is developed. Estimation of the algorithm functioning quality is obtained by means of statistical modeling.

Keywords:

optical image, space object, picture plane, processing algorithm, median filter, contour image

References

1. Vasilenko G.I., Taratorkin A.M. Vosstanovlenie izobrazhenii. (Image restoration), Moscow, Radio i svyaz’, 1986, 304 p.

2. Levshin V.L. Prostranstvennaya fil’tratsiya v opticheskikh sistemakh navigatsii. (Spatial filtration in optical systems of navigation.), Moscow, Sovetskoe radio, 1971, 200 p.

3. Furman Ya.A., Krevetskii A.V., Peredreev A.K., Rozhentsov A.A., Khafizov R.G., Egoshina I.L., Leukhin A.N.. Vvedenie v konturnyi analiz; prilozheniya k obrabotke izobrazhenii i signalov. (Introduction to contour analysis; annexes to processing of images and signals), Moscow, FIZMATLIT, 2003, 592 p.

4. Vasilenko G.I. Golograficheskoe raspoznavanie obrazov. (Holographic recognition of images), Moscow, Sovetskoe radio, 1977, 328 p.

5. Potapov A.A., Gulyaev Yu. V., Nikitov S. A., Pakhomov A. A., German V. A. Noveishie metody obrabotki izobrazhenii. (The newest methods of processing of images), Moscow, FIZMALIT, 2008, 496 p.

Download