Deep Learning-Based Generalized Efficient Layer Aggregation Network for the Detection of Coal Trucks

Coal transportation is a critical aspect of mining operations, involving the transfer of coal from extraction sites to loading areas. Previous studies have emphasized the effectiveness of object detection models, such as YOLO (You Only Look Once), in monitoring and identifying vehicles to improve safety and efficiency on transportation routes. However, accurately detecting coal trucks remains challenging due to their unique structural and operational characteristics. This study aims to evaluate the effectiveness of the YOLOv9 algorithm in detecting coal trucks compared to earlier YOLO models. The hypothesis suggests that YOLOv9 will outperform previous YOLO versions in terms of accuracy, recall, and overall performance metrics. Using an experimental design, a secondary dataset of 42 images was utilized, divided into 70% for training, 19% for validation, and 11% for testing. Multiple YOLO models were trained and assessed using standard evaluation metrics, including the confusion matrix. The results demonstrated that YOLOv9 achieved superior performance, with an F1 score of 79% and a 100% recall rate, surpassing its predecessors. These findings indicate that YOLOv9 is a promising tool for accurate coal truck detection, offering potential improvements for safer and more efficient monitoring systems in coal transportation logistics.