A.V. Korennoy 1, A.A. Kozhevnikov 2, E.A. Yashchenko 3

1−3 MTRCAF «AFA» (Voronezh, Russia)

1 korennoj@mail.ru; 2 kozhevnikov_a_a@mail.ru; 3 egorka421.91@mail.ru

Statement of the problem. The joint using of images obtained in different parts of the optical range increases their informativeness. In addition, this approach is most relevant in the presence of interfering factors (low illumination, rain, snow, fog, haze, atmospheric turbulence, defocusing of the lens, etc.). Under such conditions, the additional measures are required to improve the quality and informativeness of images. These measures should be complex, considering the images properties in different ranges.

One of the possible approach to the complex problem solution of the recovering distorted images is the Bayesian method of restoring random fields. This method involves the presence of priori information about the original images in the form of their spatial correlation functions.

Goal. Improving a priori information content in problem solution of image recovery by statistical regularization by developing adequate probabilistic models of source images in the visible and infrared ranges.

Results. To describe the model of the image field, an apparatus of stochastic differential equations is used. For still images, three types of linear equations of mathematical physics are considered. There are hyperbolic, parabolic, and elliptical. Their corresponding spatial correlation functions, which can used for a priori probabilistic description of the original images, considered too.

Because of comparison of models correlation functions and real images statistical correlation functions, it established, for a priori description of different type images, which formed in different frequency ranges, a parabolic model at corresponding values of parameters could be use.

To describe the probabilistic relationship between images in different ranges, an expression for the mutually correlative function of images parabolic model in IR and visible ranges obtained.

Practical value. The developed models can used for Bayesian algorithms synthesis of images in the infrared and visible ranges joint processing. The result of processing will depend of adequacy model data, which will describe properties of real images.

Korennoy A.V., Kozhevnikov A.A., Yashchenko E.A. Optical images modeling in infrared and visible ranges. Radiotekhnika. 2020. V. 84. № 12(23). P. 5−. DOI: 10.18127/j00338486-202012(23)-01 (In Russian).

1. Borisov Ju. Infrakrasnye izluchenija. Massovaja radio biblioteka.1976. № 906. S. 5−11 (In Russian).
2. Belousov Ju.I., Postnikov E.S. Osobennosti registracii i analiza teplovyh polej. Infrakrasnaja fotonika. SPb: Universitet ITMO. 2019.  104 s. (In Russian).
3. Vasilenko G.I., Taratorin A.M. Vosstanovlenie izobrazhenij. M.: Radio i svjaz'. 1986. 304 s. (In Russian).
4. Korennoj A.V. Obnaruzhenie, raspoznavanie i opredelenie parametrov obrazov ob’ektov. Metody i algoritmy. M.: Radiotehnika. 2012. 112 s. (In Russian).
5. Korennoj A.V. Matematicheskie modeli polutonovyh izobrazhenij. Radiotehnika. 2007. № 8. S. 79−81 (In Russian).