智能化农业装备学报（中英文）

基于YOLO模型的动物体尺关键点检测研究

预览PDF | 下载PDF

刘地华 ¹ , 司秀丽 ¹^,² , 张树敏 ³ , 马丽 ⁴^,^*^,^*

智能化农业装备学报（中英文） | 2026,7(2): 182-192

收起

智能化农业装备学报（中英文） | 2026 , 7 (2) : 182 -192

基于YOLO模型的动物体尺关键点检测研究

全屏

刘地华¹, 司秀丽¹^,², 张树敏³, 马丽⁴^,^*^,^*

作者信息

^1.吉林农业大学信息技术学院，吉林长春，130118

^2.吉林省农业人工智能研究基地，吉林长春，130118

^3.吉林农业大学动物科学技术学院，吉林长春，130118

^4.无锡学院，江苏无锡，214105

通讯作者:

^* 马丽，女，1980年生，吉林长春人，博士，教授；研究方向为智慧农业。E-mail：mali@cwxu.edu.cn

作者简介:

刘地华，男，1998年生，江西吉安人，硕士研究生；研究方向为智慧农业。E-mail：liudihuaemail@163.com

Animal body measurement keypoint detection based on YOLOv8-Pose

Dihua Liu¹, Xiuli Si¹^,², shumin Zhang³, Li Ma⁴

Affiliations

^1.School of Information Technology, Jilin Agricultural University, Changchun 130118, China

^2.Jilin Agricultural Artificial Intelligence Research Base, Changchun 130118, China

^3.School of Animal Science and Technology, Jilin Agricultural University, Changchun 130118, China

^4.Wuxi University, Wuxi 214105, China

Corresponding Author:

^* Ma Li, E-mail: mali@cwxu.edu.cn

Liu Dihua, E-mail: liudihuaemail@163.com

出版时间: 2026-05-15 doi: 10.12398/j.issn.2096-7217.2026.02.017

文章导航

摘要

收起

对动物体尺测量关键点进行精确定位，实现智能畜牧健康监控和自动化测量具有重要意义。现有关键点检测模型往往存在：参数量较大，不利于在边缘设备上的部署；部分轻量化模型的检测精度难以满足定位精度的要求。为此，研究设计一种基于改进YOLOv8-Pose架构的轻量化模型，在保证实时性的同时有效提升了关键点检测精度。该模型在主干网络的浅层以及3个检测头之前集成坐标注意力机制，将空间坐标信息有效编码到通道注意力中，增强模型对关键点空间分布感知能力，提升关键点定位的精确度和稳定性。浅层主干网络引入强制锐化模块，强化边缘和轮廓等细节信息，平衡全局语义信息与局部细节信息的表达能力。此外，模型在浅层主干网络引入强制锐化模块，强化边缘和轮廓等细节信息，平衡全局语义信息与局部细节信息的表达能力。另外，模型通过分支方式引入更高分辨率的特征图，向3个检测头提供细粒度的空间信息，弥补深层特征图在空间分辨率上的不足。试验结果表明，所提出的改进模型在复杂马匹关键点检测任务中精度达到82.6%，召回率为81.3%，mAP50-95提升至58.1%。相较于YOLOv8-Pose，模型精度、召回率和mAP50-95分别提高了21.0%、39.2%和45.3%，在复杂姿态和多变场景下表现出更强的关键点定位能力。模型参数量由3.361 M增加至4.611 M，增幅约为37.2%；浮点计算量由9.6×10⁹增加至10.4×10⁹，增幅仅为8.3%。在保持推理速度基本不受影响的前提下，模型性能实现了显著提升，表明该方法以较小的模型代价换取较大的性能收益，在精度与效率之间取得良好的平衡，对智能畜牧养殖和自动测量等应用具有重要推动作用。

关键词

YOLOv8-Pose / 动物关键点检测 / 局部特征增强 / 坐标注意力 / 智能畜牧

Abstract

收起

Accurate localization of keypoints for animal body measurement is of great importance for intelligent livestock health monitoring and automated measurement systems. However, existing keypoint detection models often suffer from two major limitations: a large number of parameters which are unfavorable for deployment on edge devices, and insufficient localization accuracy in some lightweight models that fails to meet the precision requirements of body measurement tasks. To address these challenges, this paper proposes a lightweight keypoint detection model based on an improved YOLOv8-Pose architecture, which effectively enhances detection accuracy while maintaining real-time performance. In the proposed model, coordinate attention mechanisms are integrated into the shallow layers of the backbone network and before the three heads, enabling spatial coordinate information to be effectively embedded into channel attention. This design enhances the model ability to perceive the spatial distribution of keypoints, thereby improving the accuracy and stability of keypoint localization. In addition, a forced sharpening module is introduced in the shallow backbone to strengthen edge and contour features, facilitating a better balance between global semantic representation and local detail preservation. Furthermore, a high-resolution feature branch is incorporated to provide fine-grained spatial information to the three detection heads, compensating for the loss of spatial resolution in deeper feature maps. Experimental results demonstrate that the proposed model achieves a precision of 82.6%, a recall of 81.3%, and an mAP50-95 of 58.1% on complex horse keypoint detection tasks. Compared with YOLOv8-Pose, the proposed method achieves improvements of 21.0%, 39.2%, and 45.3% in precision, recall, and mAP50-95, respectively, exhibiting superior robustness and localization capability under complex poses and diverse environmental conditions. In terms of model complexity, the number of parameters increases from 3.361 M to 4.611 M, corresponding to an increase of approximately 37.2%, while the floating-point operations rises from 9.6×10⁹ to 10.4×10⁹, an increase of only 8.3%. Despite this limited computational overhead, the proposed model delivers substantial performance gains without significantly affecting inference speed. These results indicate that the proposed approach achieves a favorable trade-off between accuracy and efficiency by incurring a relatively small model cost for considerable performance improvement, demonstrating its strong potential for intelligent livestock farming and automated body measurement applications.

Key words

YOLOv8-Pose / animal keypoint detection / local feature enhancement / coordinate attention / intelligent livestock farming

引用本文

刘地华, 司秀丽, 张树敏, 马丽. 基于YOLO模型的动物体尺关键点检测研究. 智能化农业装备学报（中英文）, 2026 , 7 (2) : 182 -192 . DOI: 10.12398/j.issn.2096-7217.2026.02.017

Dihua Liu, Xiuli Si, shumin Zhang, Li Ma. Animal body measurement keypoint detection based on YOLOv8-Pose[J]. Journal of Intelligent Agricultural Mechanization, 2026 , 7 (2) : 182 -192 . DOI: 10.12398/j.issn.2096-7217.2026.02.017

基金

收起

国家重点研发计划子课题(20200136)
吉林省科技发展计划项目(20240303066)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

[1]

Norton T, Berckmans D, Vranken E, et al. Precision livestock farming: Building ‘digital representations’ to bring the animals closer to the farmer [J]. Animal, 2019, 13(12): 3009-3017.

[2]

李熙莹，刘恒，苏淑妍，等．无接触式猪体体尺测量技术研究进展[J]．华中农业大学学报，2024，43（5）：167-177.

Li Xiying, Liu Heng, Su Shuyan, et al. Progress on non-contact measurement of pig body size [J]. Journal of Huazhong Agricultural University, 2024, 43(5): 167-177.

[3]

王克林，刘龙申，陈金鑫，等．人工智能在畜禽养殖机器人中的应用与展望[J]．智能化农业装备学报（中英文），2025，6（2）：44-57.

Wang Kelin, Liu Longshen, Chen Jinxin, et al. Application and prospect of artificial intelligence in livestock and poultry farming robots [J]. Journal of Intelligent Agricultural Mechanization, 2025, 6(2): 44-57.

[4]

Russello H, Pezzuolo A, Guarino M, et al. Video-based automatic lameness detection of dairy cows using pose estimation and multiple locomotion traits [J]. Computers and Electronics in Agriculture, 2024, 223: 109040.

[5]

Zhao S, Li J, Wang Y, et al. Deep learning-based approach for behavior recognition and pose estimation in cage-reared ducks: leveraging cage-net elimination to propel precision improvement [J]. Computers and Electronics in Agriculture, 2025, 236: 110469.

[6]

董力中，孟祥宝，潘明，等．基于姿态与时序特征的猪只行为识别方法[J]．农业工程学报，2022，38（5）：148-157.

Dong Lizhong, Meng Xiangbao, Pan Ming, et al. Recognizing pig behavior on posture and temporal featuresusing computer vision [J]. Transactions of the Chinese Society of Agricultural Engineering, 2022, 38(5): 148-157.

[7]

刘莫尘，褚镇源，崔明诗，等．基于改进YOLO v8-Pose的红熟期草莓识别和果柄关键点检测[J/OL]．农业机械学报，1-12[2025-12-15]．http://kns.cnki.net/kcms/detail/11.1964.S.20230911.2034.046.html.

[8]

Pan Y, Chang J, Dong Z, et al. PFLO: a high-throughput pose estimation model for field maize based on YOLO architecture [J]. Plant methods, 2025, 21(1): 51.

[9]

Xu Z, Liu Y, Wang H, et al. A geodesic distance regression-based semantic keypoints detection method for pig point clouds and body size measurement [J]. Computers and Electronics in Agriculture, 2025, 234: 110285.

[10]

李明煌，苏力德，张永，等．基于改进YOLOv8n-pose和三维点云分析的蒙古马体尺自动测量方法[J]．智慧农业（中英文），2024，6（4）：91-102.

Li Minghuang, Su Lide, Zhang Yong, et al. Automatic measurement of mongolian horse body based on improved YOLOv8n-pose and 3D point cloud analysis [J]. Smart Agriculture, 2024, 6(4): 91-102.

[11]

李琦，刘伟，赵建敏．基于双目视觉及Mask RCNN的牛体尺无接触测量[J]．黑龙江畜牧兽医，2020（12）：46-50，159-160.

Li Qi, Liu Wei, Zhao Jianmin. Non-contact measurement of bovine body size based on binocular vision and Mask RCNN [J]. Heilongjiang Animal Science and Veterinary Medicine, 2020(12): 46-50, 159-160.

[12]

Ao D, Hao G, Jie L, et al. Automatic livestock body measurement based on keypoint detection with multiple depth cameras [J]. Computers and Electronics in Agriculture, 2022, 198: 107059.

[13]

Li R, Zhang Y, Wang X, et al. Automated measurement of beef cattle body size via key point detection and monocular depth estimation [J]. Expert Systems with Applications, 2024, 244: 123042.

[14]

Luo X, Zhang H, Li Y, et al. Automated measurement of livestock body based on pose normalisation using statistical shape model [J]. Biosystems Engineering, 2023, 227: 36-51.

[15]

Sapkota R, Flores-Calero M, Qureshi R, et al. YOLO advances to its genesis: a decadal and comprehensive review of the You Only Look Once (YOLO) series [J]. Artificial Intelligence Review, 2025, 58(9): 274.

[16]

吴擎，韦润轩，周乐，等．基于改进YOLOv5s的轻量化鲜茶叶识别方法[J]．智能化农业装备学报（中英文），2025，6（1）：1-14.

Wu Qing, Wei Runxuan, Zhou Le, et al. Lightweight fresh tea leaf recognition method based on improved YOLOv5s [J]. Journal of Intelligent Agricultural Mechanization, 2025, 6(1): 1-14.

[17]

蔡家一，刘世伟，单龙祥，等．基于机器视觉与YOLO v5的裂纹蛋分拣机器人设计与试验[J]．智能化农业装备学报（中英文），2025，6（1）：41-50.

Cai Jiayi, Liu Shiwei, Shan Longxiang, et al. Design and experiment of cracked egg sorting robot based on machine vision and YOLO v5 [J]. Journal of Intelligent Agricultural Mechanization, 2025, 6(1): 41-50.

[18]

陈家政，付根平，黄伟锋，等．基于改进YOLOv5s的云机器人视觉导航方法[J]．智能化农业装备学报（中英文），2025，6（3）：98-110.

Chen Jiazheng, Fu Genping, Huang Weifeng, et al. Visual navigation method based on improved YOLOv5s and cloud robot [J]. Journal of Intelligent Agricultural Mechanization, 2025, 6(3): 98-110.

[19]

Cao J, Tang H, Fang H S, et al. Cross-domain adaptation for animal pose estimation [C]// Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019: 9498-9507.

[20]

Maji D, Nagori S, Mathew M, et al. YOLO-pose: enhancing YOLO for multi person pose estimation using object keypoint similarity loss [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 2637-2646.

[21]

Hou Q, Zhou D, Feng J. Coordinate attention for efficient mobile network design [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 13713-13722.

[22]

Zeiler M D, Fergus R. Visualizing and understanding convolutional networks [C]// European Conference on Computer Vision, Cham: Springer International Publishing, 2014: 818-833.

[23]

Selvaraju R R, Cogswell M, Das A, et al. Grad-cam: Visual explanations from deep networks via gradient-based localization [C]// Proceedings of the IEEE International Conference on Computer Vision, 2017: 618-626.

[24]

Wang H, Wang Z, Du M, et al. Score-CAM: Score-weighted visual explanations for convolutional neural networks [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020: 24-25.

[25]

Cho Y, Kim D, Saeed S, et al. Keypoint detection using higher order Laplacian of Gaussian [J]. IEEE Access, 2020, 8: 10416-10425.

[26]

Wang C Y, Bochkovskiy A, Liao H Y M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors [C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: 7464-7475.

[27]

He K, Gkioxari G, Dollár P, et al. Mask R-CNN [C]//Proceedings of the IEEE International Conference on Computer Vision, 2017: 2961-2969.

[28]

Duan K, Bai S, Xie L, et al. Centernet: Keypoint triplets for object detection [C]// Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019: 6569-6578.

[29]

Sun K, Xiao B, Liu D, et al. Deep high-resolution representation learning for human pose estimation [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 5693-5703.

[30]

Geng Z, Sun K, Xiao B, et al. Bottom-up human pose estimation via disentangled keypoint regression [C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 14676-14686.

2026年第7卷第2期

预览PDF | 下载PDF

引用本文

BibTeX

文章信息

doi: 10.12398/j.issn.2096-7217.2026.02.017

接收时间：2025-12-25
首发时间：2026-11-05
出版时间：2026-05-15

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2025-12-25
修回日期：2026-02-10

基金

Subproject of the National Key R&D Program of China(20200136)

国家重点研发计划子课题(20200136)

Jilin Provincial Science and Technology Development Program Project(20240303066)

吉林省科技发展计划项目(20240303066)

作者信息

^1.吉林农业大学信息技术学院，吉林长春，130118

^2.吉林省农业人工智能研究基地，吉林长春，130118

^3.吉林农业大学动物科学技术学院，吉林长春，130118

^4.无锡学院，江苏无锡，214105

通讯作者:

^* 马丽，女，1980年生，吉林长春人，博士，教授；研究方向为智慧农业。E-mail：mali@cwxu.edu.cn

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/znhnyzbxb/CN/10.12398/j.issn.2096-7217.2026.02.017

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT