Robust multi-weather pothole detection: An enhanced YOLOv9 trained on the MWPD dataset

Real-time pothole detection is crucial for advancing road safety and infrastructure management, particularly in challenging multi-weather conditions. Deep learning-based techniques, especially object detection models, have demonstrated higher accuracy than other approaches. This research proposes an improved YOLOv9 model, specifically designed for detecting road potholes in multi-weather conditions. To optimize performance, ADown layers were replaced with standard convolutional (Conv) layers at specific positions, enhancing feature extraction efficiency while reducing computational load. A custom dataset, the Multi-Weather Pothole Detection (MWPD) dataset, was developed, comprising roadway pothole images captured under varied environmental conditions. Data augmentation techniques, including color perturbation, contrast adjustment, Gaussian noise addition, flipping, and rotation, were applied to enhance training robustness. To ensure a reliable evaluation, a 5-fold cross-validation strategy was employed, partitioning the MWPD dataset into five equal subsets to minimize bias and variance. Using the evaluation benchmarks, the improved YOLOv9 achieved an average mAP@50 of 95% and an F1-score of 91%, outperforming the baseline YOLOv9 model on the MWPD dataset. • Optimized YOLOv9 with DWConv and layer swaps improves efficiency and accuracy. • Introduced MWPD dataset with multi-weather pothole images and robust preprocessing. • Achieved 85% mAP@50 and 83% F1-score, surpassing baseline YOLOv9 by 2%. • Cross-dataset validation shows +9.5% precision and +5.3% recall on Kaggle/Roboflow. • Advances scalable deep learning models for road safety under diverse conditions.