Anti-drone systems: An attention based improved YOLOv7 model for a real-time detection and identification of multi-airborne target

• Published in: Intelligent systems with applications Vol. 20; p. 200296
• Main Authors: Yasmine, Ghazlane, Maha, Gmira, Hicham, Medromi
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.11.2023; Elsevier
• Subjects: Airspace safety; Deep learning; Drone identification; Real-time detection; Deep learning; Airspace safety; Real-time detection; Drone identification
• DOI: 10.1016/j.iswa.2023.200296
• ISSN: 2667-3053

•The paper has been revised according the reviewers ‘comments. Yellow and blue highlights indications show the new edited corrections.•For security issues, it becomes highly urgent to use anti-drone system able to identify drones from non-drone objects.•The main objective is to detect and locate the airborne target properly and in real-time.•Developing a model with the best compromise between high performance and fast speed.•Incorporating a series of improvement with CSPResNeXt module in the backbone, transformer block with C3TR and decoupled heads to enhance the performance of the model.•Using a proper methodology to retain the maximum information from the learned features. Recently, with the significant rise of drones, reinforcing and securing aerial security and privacy has become an urgent task. Their malicious use takes benefit from the malevolent deployment which leverages some existing gaps in Artificial Intelligence (AI) and cybersecurity. Anti-drone systems are the spotlighted security solution developed to ensure aerial safety and security against rogue drones. However, the anti-drone systems are constraints to accurate airborne target identification and real-time detection to neutralize the target properly without causing damages. In this paper, we have developed a real-time multi-target detection model based on Yolov7 aiming to detect, identify and locate the airborne target properly and rapidly using a varied dataset which is biased and imbalanced due to the differences between the targets. In order to develop a model with the best compromise between a high performance and fast speed, we have used a series of improvements by incorporating the CSPResNeXt module in the backbone, a transformer block with the C3TR attention mechanism and decoupled head structure to enhance the performance of the model. The comparative and ablation experiments confirm the effectiveness of the proposed ensemble learning-based model. The experiments have shown that the improved model has reached high performance, with 0.97 precision, 0.961 recall, 0.979 map@0.50 and 0.732 and 0.979 map@0.50–0.95. Additionally, the real-time detection condition is satisfied with 92 FPS and an inference speed equal to 0.02 ms per image. The results show that the model succeeds in achieving an optimal balance between inference speed and detection performance. The proposed model achieves competitive results compared with the existing state-of-the-art models.