Unmanned Aerial Vehicles (UAVs), commonly known as drones, have played a significant and evolving role in the Ukrainian conflict, particularly since the 2022 Russian invasion. These UAVs can also carry significant payloads, leading to potentially unforeseen consequences. Therefore, the detection of UAVs near restricted zones has become a crucial concern. Additionally, small UAVs can record video from long distances while remaining undetected by the naked eye. A corresponding network of sensors might be necessary to predict the approach of such objects from a safe distance before they pose a threat to nearby areas. The paper explores the application of machine learning algorithms and sensor fusion for real-time UAV detection and differentiation from other objects, such as birds. Specifically, the study employs YOLOv3 in conjunction with deep sorting algorithms and sensor fusion to achieve precise UAV detection. Conducted research demonstrated that the proposed YOLOv3-based model achieved a high level of accuracy with a rate exceeding 95%. Moreover, YOLOv3 exhibited promising capabilities in real-time UAV and bird detection, indicating its potential for practical application.
Deep learning-based UAV detection
22.10.2024
263972 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Deep Learning Based Pothole Detection
IEEE | 2023
|Deep Learning-Based Speed Breaker Detection
Springer Verlag | 2024
|Ship Detection Based on Deep Learning
British Library Conference Proceedings | 2019
|Deep Learning-Based Vehicle Direction Detection
Springer Verlag | 2021
|