Hazy Aware-YOLO: An Enhanced UAV Object Detection Model for Foggy Weather via Wavelet Convolution and Attention-Based Optimization

Foggy weather poses substantial challenges for unmanned aerial vehicle (UAV) object detection by severely degrading image contrast, obscuring object structures, and impairing small target recognition, often leading to significant performance deterioration in existing detection models. To address these issues, this work presents an enhanced YOLO11-based framework, called hazy aware-YOLO (HA-YOLO), which is specifically designed for robust UAV object detection in foggy weather. HA-YOLO incorporates wavelet convolution into its structure to suppress haze-induced noise and strengthen multi-scale feature fusion without introducing additional computational overhead. In addition, a novel context-enhanced hybrid self-attention (CEHSA) module is developed, which sequentially combines channel attention aggregation (CAA) and multi-head self-attention (MHSA) to simultaneously capture local contextual cues and mitigate global noise interference. Experimental results demonstrate that the proposed HA-YOLO and its variants achieve higher detection and precision with robustness compared to the baseline YOLO11, while maintaining model efficacy. In particular, in comparison with several state-of-the-art detectors, HA-YOLO exhibits a better balance between detection accuracy and complexity, offering a practical solution for real-time UAV perception tasks in adverse weather conditions.