Guarnido-Lopez Pablo, Ramirez-Agudelo John-Fredy, Denimal Emmanuel, Benaouda Mohammed
Institut Agro Dijon, 26 bd Docteur Petitjean, 21079 Dijon, France.
Grupo de Investigación en Ciencias Agrarias-GRICA, Escuela de Producción Animal, Facultad de Ciencias Agrarias, Universidad de Antioquia, Medellin 050010, Colombia.
Animals (Basel). 2024 Sep 30;14(19):2821. doi: 10.3390/ani14192821.
This study highlights the importance of monitoring cattle feeding behavior using the YOLO algorithm for object detection. Videos of six Charolais bulls were recorded on a French farm, and three feeding behaviors (biting, chewing, visiting) were identified and labeled using Roboflow. YOLOv8 and YOLOv10 were compared for their performance in detecting these behaviors. YOLOv10 outperformed YOLOv8 with slightly higher precision, recall, mAP50, and mAP50-95 scores. Although both algorithms demonstrated similar overall accuracy (around 90%), YOLOv8 reached optimal training faster and exhibited less overfitting. Confusion matrices indicated similar patterns of prediction errors for both versions, but YOLOv10 showed better consistency. This study concludes that while both YOLOv8 and YOLOv10 are effective in detecting cattle feeding behaviors, YOLOv10 exhibited superior average performance, learning rate, and speed, making it more suitable for practical field applications.
本研究强调了使用YOLO目标检测算法监测牛进食行为的重要性。在法国一个农场录制了六头夏洛来公牛的视频，并使用Roboflow识别并标记了三种进食行为（咬、嚼、靠近）。比较了YOLOv8和YOLOv10在检测这些行为方面的性能。YOLOv10的表现优于YOLOv8，其精度、召回率、mAP50和mAP50 - 95得分略高。尽管两种算法的总体准确率相似（约90%），但YOLOv8更快达到最优训练，且过拟合现象较少。混淆矩阵表明两个版本的预测误差模式相似，但YOLOv10表现出更好的一致性。本研究得出结论，虽然YOLOv8和YOLOv10在检测牛进食行为方面都有效，但YOLOv10在平均性能表现更好、学习率和速度更快，使其更适合实际现场应用。
