Pest Detection using State-of-The-Art YOLO Models: A Comparative Study

Agriculture is an essential component of any nation's economy, yet it faces a constant adversary in the form of pests. These unwanted organisms pose a serious threat to agricultural processes because they devour plant yields with a voracious appetite. The implementation of cutting-edge You Only Look Once (YOLO) models for pest detection and species identification is the subject of this study's thorough examination of pest detection techniques. This study aims to identify the relative advantages of the YOLO models and improve pest detection in agriculture. The YOLO models-v5, v7, v8 were analysed and compared to ensure consistency in parameters across all models, resulting in an objective and thorough assessment. With a training time of only about 0.87 hours, YOLOv5 emerged within this paradigm as the standards of time efficiency. YOLOv8 distinguished as the best possible level of performance in contrast, earning a remarkable mean Average Precision ([email protected]) score of 0.99. This study assists in making decisions that are appropriate for agricultural needs, thereby bolstering this vital industry against the ongoing threat of pests.