THE OBJECTIVE OF UAV CENS BINDING SPECIFYING USING STRUCTURAL AND CONTRAST FEATURES
Р. Г. Сидоренко, О. Sotnikov, В.А. Таршин, Я. М. Кожушко, Y. Karmannyi · Випробування та сертифікація · 2025
The analysis is conducted on the use of autonomous UAVs with combined navigation systems during the repulsion of large- scale armed aggression of the Russian Federation. The author formulated specifying the binding objective task for determining the coordinates of UAVs using CENS. The forming crucial function principle resulting from the comparison of reference image stored on board with the current image generated by CENS during the flight are shown in the article. Analytical expressions for forming CENS of current and equivalent images and their dependence on information features are given as well. The use of fractal dimension values justification for specifying areas in images that differ significantly from the background and are in a given range of fractal dimensions that are most suitable for CENS binding is provided. The study is conducted to identify the binding objective in the UAV CENS using structural and contrast features in the MATLAB software environment, based on the input image obtained from Google Earth Pro. The results of numerical simulation of contrast-segmented images of the binding area on the basis of corresponding segmented by structure images, their fractal dimension fields and histograms are presented. An example of the binding objective identification in the UAV CENS using structural and contrast features is shown. The structure of the two-stage image segmentation algorithm using structural and contrast features is developed in the article. Practical recommendations for the developed approach application are given as well. The proposed two-stage procedure application for binding objective specifying will allow to avoid ambiguity in the adoption of an accurate navigation decision by the CENS and is determined by the navigation parameters of the UAV. An increase in the speed of the CENS is an important benefit of this approach. Due to the absence of scatter in the FR parameter it is possible to use a small sample of current images that characterizes the structural properties of the contrast-selective image. The obtained results can be used for further development and optimization of subsystems for detecting and processing information as well as for the forming UAV control signals under difficult interference conditions. The article can be useful for creating new approaches to improve the accuracy and probability of locating modern air attack means produced by the state.