| 刊名 | Agricultural Biotechnology |
| 作者 | Jizhen WU1, Jianfei SHI1*, Zhiyuan JING2 |
| 作者单位 | 1.College of Information and Electrical Engineering, Heilongjiang Bayi Agricultural University, Daqing 163319, China; 2.Daqing Oilfield Design Institute Co., Ltd., Daqing 163319 163001, China |
| DOI | DOI:10.19759/j.cnki.2164-4993.2025.06.008 |
| 年份 | 2025 |
| 刊期 | 6 |
| 页码 | 36-39 |
| 关键词 | Image recognition; YOLOv11n; Cow behavior recognition; Deep learning |
| 摘要 | To address the issue of low recognition accuracy for eight types of behaviors including standing, walking, drinking, lying, eating, mounting, fighting and limping in complex multi-cow farm environments, a multi-target cow behavior recognition method based on an improved YOLOv11n algorithm was proposed. The detection capability for small targets in images was enhanced by incorporating a DASI module into the backbone network and a MDCR module into the neck network, based on YOLOv11. The improved YOLOv11 algorithm increased the mean average precision from the original 89.5% to 93%, with particularly notable improvements of 8.7% and 6.3% in the average precision for recognizing drinking and walking behaviors, respectively. These results fully demonstrate that the proposed method enhances the model's ability to recognize cow behaviors. |
