Deep learning with the YOLO v4 model for drone detection among birds
|
Full Text |
Pdf
|
|
Author |
S. Sharmila and P. Preethy Rebecca
|
|
e-ISSN |
1819-6608 |
|
On Pages
|
641-647
|
|
Volume No. |
20
|
|
Issue No. |
10
|
|
Issue Date |
August 5, 2025
|
|
DOI |
https://doi.org/10.59018/052579
|
|
Keywords |
unmanned aerial vehicles, drones, birds, YOLO v4, convolutional neural networks, computer vision, object detection, instances, accuracy, FPS.
|
Abstract
Aim: Drones, also called UAV’s are becoming popular in various fields, including military and commercial applications. But the challenge lies in detecting drones among birds, which is complex using traditional methods such as radar systems and machine learning algorithms, whose accuracy is limited to a particular area. The aim is to detect the occurrence of drones in images that constitute birds in the background. When a drone is present, algorithms detect and label a restricted area, and it should not alert birds. Significance: To solve this issue, we propose a deep learning algorithm for drone detection using the YOLO V4 Model. The study involves the use of computer vision technology to identify and track drones in areas where birds are present. Methodology: The YOLO v4 model is a state-of-the-art object detection algorithm that uses convolutional neural networks (CNNs) to identify and track objects in real-time. Deep learning algorithm makes use of a large dataset of images and train the model to acknowledge the distinctive features of drones, namely their shape, size, and movement patterns. It allows the model to accurately detect and track drones even in challenging environments with complex backgrounds and lighting conditions. This model significantly improves accuracy, making it an effective tool for drone detection in various operations. Results: Deep Learning mechanism enables a great impact for Object detection in the case of drones. Performance analysis generated improved results in the accuracy of detection compared to traditional methods.
Back