VEHICLES AND FORTIFICATIONS LOCALIZATION ON THE IMAGE BASED ON ARTIFICIAL INTELLIGENCE APROACHES

Abstract

Warfare, especially defensive warfare, is based not only on strategy and tactics, but also on the ability to quickly analyze the situation and adapt to constantly changing conditions. The effective use of intelligence data, including the accurate identification of enemy positions, is crucial for the successful planning and execution of defensive or offensive operations. Therefore, the development of advanced image analysis methods capable of automatically localizing military objects is becoming one of the priority areas of modern military engineering.

Developing a model that identifies objects through a UAV camera lens, especially for military purposes, poses some challenges and nuances that need to be overcome. One possible problem is the lack of data for quality model training, as previously described, although significant, it is not the only one. Initial task for the research will be to collect dataset of images from UAV of vehicles and fortification structures. This paper presents the results of research on the localization of military equipment, as well as fortifications and objects in UAV images, using artificial intelligence tools, namely the YOLO method. The main goal of the study is to achieve high accuracy and efficiency of object segmentation in combat conditions. To ensure optimal speed and accuracy of detection, the YOLOv8 model is used, which allows to effectively classify equipment and fortifications in images. The object of research is the processes of localization of military equipment and fortifications using artificial intelligence. The subject of the study is the methods and algorithms used to localize objects in the image from UAV.

The study introduces a novel application of artificial intelligence, specifically leveraging YOLO technology, for the detection of military equipment and fortifications in UAV imagery. This approach integrates advanced deep learning techniques with innovative algorithms to enhance the precision and dependability of identifying and categorizing military objects.