E.A. Samoylin - Dr.Sci. (Eng.), Professor of Chair of Military Education-Research Center of Military Air Forces «Military Aviation Academy named for prof. N.E. Zhukovsky and Yu.A. Gagarin» (Voronezh)
V.V. Shipko - Post-Graduated, Military Academy of Military Education-Research Center of Military Air Forces «Military Aviation Academy named for prof. N.E. Zhukovsky and Yu.A. Gagarin» (Voronezh)
M.G. Ivanov - Operator of Scientific Company of Military Education-Research Center of Military Air Forces «Military Aviation Academy named for prof. N.E. Zhukovsky and Yu.A. Gagarin» (Voronezh)

It is known that in many countries development of optical-electronic (laser) suppression complexes is intensively conducted recently. Influence of laser radiation on a photodetector of optical-electronic system leads to a flare or failure of photosensitive elements of a photodetector that in turn leads to impulse noise on digital images. In accordance with this, the aim of this paper is to improve the detection accuracy of impulse noise on three-channel digital images received by optical-electronic systems in conditions of laser suppression. In this paper we proposed the algorithm of detection of impulse noise on the digital images registered by a three-channel photo¬detector in conditions of influence of laser radiation. The algorithm is based on the property of interchannel redundancy of multichannel digital images and assumes calculation of kernel gradients in each of color channels and their comparison. Results of experimental studies of the impulse interferences mechanism are given in color channels of the digital image at hit of laser radiation of various power and wave length on a three-channel photodetector. The dependence of the calculated values of probability of impulse interferences emergence in three channels of the image from density of power and length of a laser radiation wave is shown. Results of calculation of error detection of the first and second sort for offered algorithm, as well as known rank procedures are presented. Calculation of errors is carried out in all range of probability of noise emergence on the image, defined by the density of power of the laser radiation applied to the the photodetector. The error values of detection of the first and second sort of the developed algorithm received thus are lower in comparison with known such algorithms. The results indicate that the proposed algorithm allows to solve the problem of detection of impulse noise by increasing the number of spectral channels, and therefore, increases the color components of digital images in modern jam-resistant optoelectronic systems.