M.A. Pantyukhin1, A.V. Bogoslovsky2, I.V. Zhigulina3, E.S. Legostaeva4

1-4 MESC AF «N.E. Zhukovsky and Y.A. Gagarin Air Force Academy» (Voronezh, Russia)

Two-dimensional discrete filtering of images solves the problem of selecting objects on an arbitrary background. The output of a discrete 2D-filter is the correlation of the input image and the impulse response of the 2D-filter. Since all spatial directions are equal, the filtering is isotropic. But the image is finite in spatial coordinates, so the edge effects are prominent. Nearby objects can make a negative contribution to the processing result, which will lead to errors in the interpretation of the results.

The aim of the article is to explore the possibility of eliminating edge effects in two-dimensional discrete image filtering, along with removing the negative impact of nearby objects on the processing results.

For this aim it is proposed to use discrete filters with an anisotropic structure. This excludes the development of the edge effects.

An example of processing a test image with anisotropic 5x5 filters is given. It is shown that it is possible to separate images of two closely spaced objects by performing a phased "observation of the image" with a set of anisotropic filters.

An image processing algorithm is proposed. At the first stage parallel processing with anisotropic filters is performed, while at the second stage an output image is formed, in each sample of which we record the maximum of the corresponding reports obtained as a result of anisotropic filtering.

Examples of real image processing are given, confirming that the use of 2D-filters of various configurations makes it possible to completely compensate for edge effects, and joint processing by similar filters increases the efficiency of two-dimensional discrete filtering.

1. Obrabotka mnogomernyh signalov. V 2-h knigah. Kn. 1. Linejnaja mnogomernaja diskretnaja obrabotka signalov. Metody analiza i sinteza. Pod red. A.V. Bogoslovskogo. M.: Radiotehnika, 2013. 168 s. (In Russian).
2. Gonsales R., Vuds R. Cifrovaja obrabotka izobrazhenij. Izd. 3-e, ispr. i dop. M.: Tehnosfera, 2019. 1104 s (In Russian).
3. Pantjuhin M.A., Zhigulina I.V., Beljaev R.V. Dvumernaja diskretnaja fil'tracija izobrazhenij v sistemah tehnicheskogo zrenija. Materialy III Vseross. nauch.-tehnich. konf. «Sostojanie i perspektivy razvitija sovremennoj nauki po napravleniju «Tehnicheskoe zrenie i raspoznavanie obrazov». Anapa. 2021. S. 260–267 (In Russian).
4. Mart'janova A.V. Ispol'zovanie izotropnyh fil'trov v obrabotke izobrazhenij. Informacija: peredacha, obrabotka, vosprijatie. Ekaterinburg: UrFU. 2016. S. 34-41 (In Russian).
5. H'jubel D. Glaz, mozg, zrenie: Per. s angl. M.: Mir. 1990. 239 s. (In Russian).
6. Changizi M. Revoljucija v zrenii: chto, kak i pochemu my vidim na samom dele. M: AST. 2015. 304 s. (In Russian).
7. Remington L. Klinicheskaja anatomija i fiziologija zritel'noj sistemy: Per. s angl. M.: ID Gorodec. 2020. 336 s. (In Russian).
8. Vhodnoe izobrazhenie «Stat'ja ptic». URL: https://www.static.orszagalbum.hu/nagy/ 1/3/2/1/7/1/5/7/4/1/1321715741.jpg (data obrashhenija: 01.09.2022) (In Russian).
9. Pantjuhin M.A., Bogoslovskij A.V., Zhigulina I.V. Dvumernaja diskretnaja fil'tracija testovyh izobrazhenij. Radiotehnika. 2020. № 3(5). S. 64–72. DOI: 10.18127/j00338486-202003(05)-07 (In Russian).
10. Vhodnoe izobrazhenie «Ajerodrom». URL: https://ic.pics.livejournal.com/sergeye2e2/ 4665613/45549/45549_original.jpg (data obrashhenija: 16.01.2023) (In Russian).
11. Vhodnoe izobrazhenie «Zhivotnye». URL: https://image.fhserv.ru/hunting/2015-05-af0458075f983442f3c9e200d6d9b92e__rsu-1000-800.jpg (data obrashhenija: 16.01.2023) (In Russian).
12. Hrjashhev V.V., Priorov A.L., Pavlov V.A., Ivanovskij L.I. Segmentacija ob’ektov na sputnikovyh izobrazhenijah s ispol'zovaniem svertochnyh nejronnyh setej. Uspehi sovremennoj radiojelektroniki. 2019. № 6. S. 28-35 (In Russian).