For Precision Animal Husbandry: Precise Detection of Specific Body Parts of Sika Deer Based on Improved YOLO11 DOI Creative Commons

Jinfan Wei,

Haotian Gong, Lan Luo

и другие.

Agriculture, Год журнала: 2025, Номер 15(11), С. 1218 - 1218

Опубликована: Июнь 3, 2025

The breeding of sika deer has significant economic value in China. However, the traditional management methods have problems such as low efficiency, easy triggering strong stress responses, and damage to animal welfare. Therefore, development non-contact, automated, precise monitoring technologies become an urgent need for sustainable this industry. In response demand, study designed a model MFW-YOLO based on YOLO11, aiming achieve detection specific body parts real environment. Improvements include: designing lightweight efficient hybrid backbone network, MobileNetV4HybridSmall; multi-scale fast pyramid pooling module (SPPFMscale) is proposed. WIoU v3 loss function used replace default function. To verify effectiveness method, we constructed dataset containing 1025 images, covering five categories. experimental results show that improved performs well. Its mAP50 MAP50-95 reached 91.9% 64.5%, respectively. This also demonstrates outstanding efficiency. number parameters only 62% (5.9 million) original model, computational load 60% (12.8 GFLOPs) average inference time 3.8 ms. work provides algorithmic support achieving non-contact intelligent deer, assisting automated (deer antler collection preparation), improving welfare, demonstrating application potential deep learning technology modern precision husbandry.

Язык: Английский

For Precision Animal Husbandry: Precise Detection of Specific Body Parts of Sika Deer Based on Improved YOLO11 DOI Creative Commons

Jinfan Wei,

Haotian Gong, Lan Luo

и другие.

Agriculture, Год журнала: 2025, Номер 15(11), С. 1218 - 1218

Опубликована: Июнь 3, 2025

The breeding of sika deer has significant economic value in China. However, the traditional management methods have problems such as low efficiency, easy triggering strong stress responses, and damage to animal welfare. Therefore, development non-contact, automated, precise monitoring technologies become an urgent need for sustainable this industry. In response demand, study designed a model MFW-YOLO based on YOLO11, aiming achieve detection specific body parts real environment. Improvements include: designing lightweight efficient hybrid backbone network, MobileNetV4HybridSmall; multi-scale fast pyramid pooling module (SPPFMscale) is proposed. WIoU v3 loss function used replace default function. To verify effectiveness method, we constructed dataset containing 1025 images, covering five categories. experimental results show that improved performs well. Its mAP50 MAP50-95 reached 91.9% 64.5%, respectively. This also demonstrates outstanding efficiency. number parameters only 62% (5.9 million) original model, computational load 60% (12.8 GFLOPs) average inference time 3.8 ms. work provides algorithmic support achieving non-contact intelligent deer, assisting automated (deer antler collection preparation), improving welfare, demonstrating application potential deep learning technology modern precision husbandry.

Язык: Английский

Процитировано

0