Improving Small Drone Detection Through Multi-Scale Processing and Data Augmentation

2025-04-27

AI Summary

To address low detection accuracy in complex scenes where small unmanned aerial vehicles (UAVs) are easily confused with birds, this paper proposes a robust detection method based on YOLOv11. The approach introduces three key innovations: (1) a novel multi-scale prediction aggregation mechanism that jointly leverages patch-wise and full-image features to enhance localization accuracy for small objects; (2) a semantic-aware Copy-Paste data augmentation strategy that improves inter-class discriminability by preserving contextual semantics during instance synthesis; and (3) a trajectory-consistency-driven post-processing module that compensates for missed detections across frames, thereby strengthening temporal robustness. Evaluated on the 8th WOSDETC UAV vs. Bird Detection Challenge at IJCNN 2025, the method ranks among the top three globally, achieving significant improvements in mean Average Precision (mAP) and recall over baseline models—particularly excelling in small-object detection tasks.

Technology Category

• Computer Vision: Object Detection & Categorization
• Machine Learning: Multi-instance/Multi-view Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Detecting small drones, often indistinguishable from birds, is crucial for modern surveillance. This work introduces a drone detection methodology built upon the medium-sized YOLOv11 object detection model. To enhance its performance on small targets, we implemented a multi-scale approach in which the input image is processed both as a whole and in segmented parts, with subsequent prediction aggregation. We also utilized a copy-paste data augmentation technique to enrich the training dataset with diverse drone and bird examples. Finally, we implemented a post-processing technique that leverages frame-to-frame consistency to mitigate missed detections. The proposed approach attained a top-3 ranking in the 8th WOSDETC Drone-vsBird Detection Grand Challenge, held at the 2025 International Joint Conference on Neural Networks (IJCNN), showcasing its capability to detect drones in complex environments effectively.

Problem

• Detecting small drones indistinguishable from birds
• Enhancing drone detection using multi-scale processing
• Improving accuracy with data augmentation and post-processing

Innovation

• Multi-scale processing enhances small drone detection
• Copy-paste data augmentation diversifies training examples
• Frame-to-frame consistency reduces missed detections