Hennadii Khudov et al., International Journal of Emerging Trends in Engineering Research, 8(7), July 2020, 3927 - 3934 3927 ABSTRACT The result of decoding the images obtained from on-board optical-electronic surveillance systems depends on the quality of image segmentation, especially taking into account the peculiarities of their production (heterogeneous background, variability of different parts of the image, the presence of noise). The main techniques, criteria and indicators of image segmentation, their advantages and disadvantages are analyzed. It is proposed to evaluate the information indicator (Kulbak-Leibner distance) of thematic segmentation of the optoelectronic image by the Canney method. The analysis of the main stages of the Cannes method is carried out: smoothing, gradient search, suppression of false maxima, double threshold filtering, tracing of the uncertainty region. The result of segmentation of the optoelectronic image by the Canney method is given, the calculation of the Kulbak-Leibner distance on its dependence on the scale factor of the original image is carried out. Key words : the optoelectronic image, on-board surveillance system, segmentation, the Canny edge detection algorithm, the Kullback-Leibler divergence. 1. INTRODUCTION In modern conditions of network-centric and hybrid wars and anti-terrorist operations, about 80% of reconnaissance missions, 60% of security tasks and 50% of fire assault tasks are solved using information obtained from on-board surveillance systems (unmanned aerial vehicles and space surveillance systems) [1]–[6]. The efficiency of image interpretation obtained from on-board surveillance systems can be represented by four categories [7]: "A" – confident interpretation without the use of additional materials; "B" – interpretation is possible in-house using additional materials; "C" – interpretation is possible only using field research; "D" – interpretation is not possible. Table 1 shows the summary results of the possibility of interpretation of objects using optoelectronic images obtained from Pleiades, WorldView spacecraft (SC) and when performing aerial photography using the VisionMap A3 camera [8]. 236 objects of the following types were analyzed: land relief; hydrography; localities; socio-economic objects; road networks and road constructions; vegetation and soil. Table 1: The summary results of the possibility of interpretation of objects Output data type Number of objects A B C D SC Pleiades 103 92 21 20 SC WorldView 106 90 19 21 Camera Vision A3 126 78 16 16 From the analysis of the results shown in table 1, it can be seen that for slightly less than half of the objects they are interpreted with confidence (category "A"). Approximately 10% of objects cannot be interpreted (category "D"). And about 50% of objects can be interpreted using additional materials. In the presence of even additional materials, difficulties arise when solving the problem of interpreting Estimation of the Kullback-Leibler Divergence for Canny Edge Detector of Optoelectronic Images Segmentation Hennadii Khudov 1 , Rostyslav Khudov 2 , Irina Khizhnyak 3 , Volodymyr Loza 4 , Taras Kravets 5 , Sergii Kibitkin 6 ,1 Department of Radar Troops Tactic, Ivan Kozhedub Kharkiv National Air Force University, Kharkiv, Ukraine, 2345kh_hg@ukr.net 2 Department of Theoretical and Applied Informatics, Kharkiv National University named after V. N. Karazin, Kharkiv, Ukraine, rhudov@gmail.com 3 Department of Mathematical and Software Automated Control Systems, Ivan Kozhedub Kharkiv National Air Force University, Kharkiv, Ukraine, khizh_ia@ukr.net 4 Department of Communications and Automated Control Systems, Ivan Cherniakhovskyi National Defence University, Kyiv, Ukraine, vladymyrloza@gmail.com 5 Department of Artillery Facility of Rocket Forces and Artillery, Hetman Petro Sahaidachnyi National Army Academy, Lviv, Ukraine, 2345kh_hg@ukr.net 6 Department of Aviation equipment and aerial reconnaissance complexes, Ivan Kozhedub Kharkiv National Air Force University, Kharkiv, Ukraine, sergejkibitkin@ukr.net ISSN 2347 - 3983 Volume 8. No. 7, July 2020 International Journal of Emerging Trends in Engineering Research Available Online at http://www.warse.org/IJETER/static/pdf/file/ijeter162872020.pdf https://doi.org/10.30534/ijeter/2020/162872020