肉牛生产性能智能监测技术研究进展

doi:10.3864/j.issn.0578-1752.2025.23.021

摘要/Abstract

摘要：

随着我国肉牛规模化养殖的快速发展，以物联网、大数据及人工智能为代表的现代智能肉牛养殖技术水平得到不断提高。肉牛个体的身份识别及体重、体尺和采食量等生产性能的实时监测对提高牧场饲养管理水平、降低人员工作量、加快肉牛育种选育进程具有重要意义。个体识别是肉牛个体生产性能监测的基础，当前主要依赖于RFID识别技术和基于图像的深度学习个体识别技术。RFID技术识别精度高，但面临成本高、识别距离短、佩戴工作量大的问题。而基于图像的深度学习识别技术通过分析肉牛的体表花纹、耳标文本、鼻纹、虹膜、视网膜、面部和侧面轮廓等独特生物特征实现个体识别。但其识别效果可能受光照条件和动物个体差异的影响。未来需着力研发能够适应不同环境条件并实现精准、快速、动态的肉牛机器视觉识别技术。肉牛体尺与体重的精准估测主要通过采用2D、3D相机拍摄的二维、三维图像，经过关键特征点的提取分析计算实现。2D相机具有设备获取简单、成本低的优势，但其在测定过程中需要已知尺寸的参照物，且在测量胸围、腹围等曲面特征体尺指标时存在局限性，直接影响了相关体尺测定和体重估测的精准度。相比之下，3D相机能够全方位、立体化获取肉牛体表结构及其与设备的距离信息，从而为多维度体尺体重指标的精准测量提供可能。肉牛采食量自动监测对判断饲料效率至关重要。自动计重料槽借助压力传感器精准计量采食前后食槽重量差值，实现采食量的自动精准测定，但其安装成本高、饲喂不便等问题限制了其应用范围。通过肉牛采食前后饲料的深度图像变化或通过传感器记录采食行为也可有效估测采食量。然而，在实际应用中，饲料组成的复杂性及形态变化可能对监测准确性造成干扰。基于机器视觉的肉牛生产性能测定技术在肉牛生产性能测定上取得了显著的进展，但仍面临诸多挑战，如数据处理量大、环境干扰影响结果准确性和数据深度挖掘利用不足等。未来可采取以下策略：通过边缘计算技术、优化阶段性检测等策略降低设备数据计算负担，提高系统响应速度；探索基于单视角深度相机的三维重建技术，以提高肉牛体尺和体重监测在实际生产应用的可行性；研发适用于不同品种、各生长阶段肉牛的通用预测模型，以增强技术的普适性与实用性；加强多模态数据融合，提高肉牛生产性能监测数据的综合应用价值。智能监测技术是肉牛养殖现代化的关键，通过技术创新与融合，有望实现低成本、高精度、广泛适用的智能化肉牛生产性能监测技术，从而推动肉牛产业智能化升级，提升生产效率与经济效益，满足市场需求。文章系统总结了肉牛养殖过程身份识别及体重、体尺和采食量的智能监测技术，深入分析了我国肉牛生产性能智能监测技术当前面临的挑战和未来发展趋势，旨在为相关智能化监测技术的研发和应用提供参考。

关键词: 肉牛, 生产性能, 智能化, 体尺, 体重, 深度学习, 机器视觉

Abstract:

With the rapid development of large-scale beef breeding in our country, modern smart beef breeding technology, including the Internet of Things, big data and artificial intelligence, has been continuously improved. The identification of individual beef cattle and real-time monitoring of production performance, such as body weight, body size, and feed intake, are crucial for improving feeding management, reducing personnel workload, and accelerating the breeding process of beef cattle. Individual identification is the foundation for monitoring individual production performance. Current methods primarily rely on RFID identification technology and image-based deep learning individual identification technology. While RFID offers high accuracy, it faces challenges such as high cost, short identification distance, and significant workload for tagging. Image-based deep learning identification technology identifies individuals by analyzing unique biometric features like body surface patterns, ear tag text, nose prints, iris, retina, facial features, and side profiles. However, its effectiveness can be affected by lighting conditions and individual differences. In the future, it is necessary to develop precise, rapid and dynamic recognition machine vision recognition technology for beef cattle that can adapt to different environmental conditions. Images captured by 2D and 3D cameras can be used for estimation of body size and weight after key feature extraction and analysis. 2D cameras have the advantages of simple equipment acquisition and low cost. However, its reliance on reference objects of known dimensions during the measurement process, as well as the measurement limitations of curved surface characteristic body size indicators such as chest circumference and abdominal circumference, directly affect the accuracy of related body size measurement and body weight estimation. In contrast, 3D cameras can obtain the external structure of beef cattle and the distance information between them and the equipment in a comprehensive and three-dimensional manner, thus providing the possibility for precise measurement of multi-dimensional body weight indicators. Automated monitoring of beef cattle feed intake is vital for assessing feed efficiency. Automatic weighing feed troughs accurately measure intake by calculating the weight difference before and after feeding using pressure sensors. However, the challenges such as high installation costs and operational inconvenience have largely confined their application scope. Feed intake can also be effectively estimated through depth image changes before and after feeding or by recording feeding behavior by using relevant sensors. Nevertheless, in practical applications, the complexity of feed composition can affect monitoring accuracy. The technology for determining the production performance of beef cattle based on machine vision has made remarkable progress. However, it still faces many challenges, such as large amounts of data processing, environmental interference affecting the accuracy of results, and insufficient data development and utilization. In the future, strategies such as edge computing technology and optimizing phased detection can be adopted to reduce the computing pressure of device data and improve the agility of system response. Exploring 3D reconstruction technology based on single-view depth cameras could improve the feasibility of applying body dimension and weight monitoring in practical production settings. Efforts should be dedicated to developing universal prediction models applicable to different breeds and various growth stages to enhance the versatility and practicality of the technology. Strengthening multimodal data fusion will improve the comprehensive application of beef cattle production performance monitoring data. Intelligent monitoring technology is the key to the modernization of beef cattle breeding. Through technological innovation and integration, it is expected to achieve low-cost, high-precision and widely applicable intelligent monitoring technology for the production performance of beef cattle, promoting the intelligent upgrade of the beef cattle industry, improving production efficiency and economic benefits, and meeting market demands. This review summarized intelligent monitoring technologies for beef cattle identification, as well as for estimating body weight, body size, and feed intake. It also discussed the challenges and future development trends of intelligent monitoring technology for beef cattle production performance in China, aiming to provide references for the research and application of related intelligent monitoring technologies.

Key words: beef cattle, production performance, intelligent, body size, weight, deep learning, machine vision

张帆, 唐湘方, 杨亮, 王辉, 陈睿鹏, 熊本海. 肉牛生产性能智能监测技术研究进展[J]. 中国农业科学, 2025, 58(23): 5081-5096.

ZHANG Fan, TANG XiangFang, YANG Liang, WANG Hui, CHEN RuiPeng, XIONG BenHai. Research Progress of Intelligent Monitoring Technology for Beef Cattle Production Performance[J]. Scientia Agricultura Sinica, 2025, 58(23): 5081-5096.

0
/ / 推荐

导出引用管理器 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

链接本文: https://www.chinaagrisci.com/CN/10.3864/j.issn.0578-1752.2025.23.021

https://www.chinaagrisci.com/CN/Y2025/V58/I23/5081

图/表 3

表1

肉牛个体识别技术对比"

监测部位 Monitoring site	监测设备 Monitoring equipment	主要识别算法 Main identification method	准确率 Accuracy rate	特点 Characteristic	参考文献 Reference
耳标文本 Ear tag text	三千万像素照相机 30-megapixel camera	CRNN	92.30%	利用耳标进行文本识别，但对图像采集角度及摄像机的像素要求较高 Text recognition using ear tags is possible but needs specific image acquisition angles and high camera resolution	[22]
耳标文本 Ear tag text	SP007 Sricam IP防水摄像机与红外夜视仪 SP007 Sricam IP waterproof camera with infrared night vision device	YOLOv3	> 95%	引入纠错算法，避免误读 An ad hoc error correction algorithm is presented to avoid misreading	[12]
鼻纹 Muzzle	X-T4相机 X-T4 camera	VGG	98.7%	通过加权交叉熵损失函数和数据增强处理提高结果精度 The identification accuracy is improved through the weighted cross-entropy loss function and data augmentation processing	[23]
虹膜 Iris	二维图像 2D image	2DLP-LDA	94.07%	优秀的泛化能力，可识别存在局部遮挡和形变等质量缺陷的图像 Good generalization ability and can identify images with quality defects such as local occlusion and deformation	[26]
视网膜 Retina	二维图像 2D image	SURF	92.25%	利用视网膜血管识别，准确率高 The recognition by retinal blood vessels has a high accuracy rate	[27]
面部 Face	索尼A6000和华为P30 SONY A6000 and Huawei P30	MobileNetV1为主干网络，K-means++聚类 MobileNetV1 as the backbone network, and K-means++ clustering algorithm	无遮挡时99.86%；遮挡小于30%时90%以上 99.86% with no occlusion, and over 90% when occlusion is less than 30%	在遮挡情况下仍可识别 It can still be recognized even in occluded conditions	[30]
面部 Face	二维图像 2D image	YOLO-Unet组合网络模型 YOLO-Unet combined network	90.48%	引入背景消去模块，提高准确率11.99% The model with background removal improve recognition accuracy of 11.99%	[28]
面部 Face	二维图像 2D image	SOLOv2	98.06%	添加SOLOv2实例分割模型获取牛脸轮廓信息，提高牛面部识别准确率 Add the SOLOv2 instance segmentation model to obtain the contour information of cow faces and improve the accuracy of cow face recognition	[29]
侧面 Body side	华为Mate30摄像机 Huawei Mate30 camera	TLAINS-InceptionV3	99.74%	解决肉牛无明显自身特征带来的识别准确率问题 Solve the problem of recognition accuracy caused by the lack of obvious self-characteristics of beef cattle	[32]

表1

表2

肉牛体尺测定技术研究对比"

监测部位 Monitoring site	测定指标 Monitoring parameters	监测设备 Monitoring equipment	主要识别算法 Main identification method	准确率 Accuracy rate	特点 Characteristic	参考文献 Reference
侧面 Body side	体高、体斜长、胸深、蹄径 Body height, body oblique length, chest depth, hoof diameter	单个RGB相机 Single RGB camera	YOLO v5s对肉牛目标进行定位，Lite-HRNet提取关键点 YOLO v5s detects the beef cattle target, and Lite-HRNet detects the key points	相对误差分别6.75%、7.55%、8.00%、8.97% With average relative error of 6.75 %, 7.55 %, 8.00 % and 8.97%	能准确测量不同距离和光照条件下的体尺，但结果易受背景和拍摄角度影响 Accurately measure the body sizes under different distances and lighting conditions, but the accuracy is affected by the background and shooting angle	[39]
侧面 Body side	体高、体长 Body height, body length	二维图像 2D image	Image 2	误差小于0.04 m Average prediction deviation < 0.04 m	成本低，但胸围预测准确率低 Low cost, but low prediction accuracy of chest circumference	[40]
侧面 Body side	体高、体斜长 Body height, body oblique length	二维图像 2D image	YOLOv5	相对误差分别为4.49%、4.70% With relative error of 4.49% and 4.70%	以已知尺寸的耳标为参照计算体尺 Using the ear tag of known size as reference to calculate body sizes	[41]
侧面 Body side	体高，胸深、背高、体斜长、腰高 Body height, chest depth, back height, body oblique length, waist height	单个IFM O3D303 3D激光雷达 IFM O3D303 3D LiDAR camera	GPT三维重建 Greedy Projection Triangulation reconstruction	精度2 mm，相对误差< 2% With accuracy of 2 mm and relative error close to 2%	精度高，但操作复杂，且动物需要保持静止 High precision, but the operation is complex and the animal needs to remain stationary	[42]
5个角度 Five angles	体高、臀高、胸围、腹围、体斜长 Body height, hip height, chest circumference, abdominal circumference, body oblique length	5个Kinect DK Five Kinect DK sensors	BTSS进行关键区域分割，基于姿态的调整模型进行结果校正 Segmentation of key regions by bidirectional tomographic slice segmentation method, along with Posture-based Measurement Adjustment model to correct result	相对误差分别为1.84%、3.47%、1.56%、2.36%、1.14% With relative errors 1.84%, 3.47%, 1.56% 2.36% and 1.14%	多摄像头实时三维重建，精确度高 Multi-camera real-time 3D reconstruction with high accuracy	[43]
侧面 Body side	体高、胸围、背高、腰高、体斜长 Body height, chest circumference, back height, waist height, body oblique length	单个IFM O3D303 Single IFM O3D303	按照动物左右对称原理合并后，经孔洞修复后计算体尺 After merging according to the principle of left-right symmetry of animals, the body sizes were calculated after hole repair	平均相对误差5.42% With average relative error 5.42%	基于对称原理构建真实肉牛轮廓用于体尺测量 The real beef cattle contour was constructed based on the principle of symmetry for body size measurement	[44]

表2

表3

参考文献 80

[1]	曹兵海, 曹建民, 李俊雅, 郭爱珍, 王之盛, 孙宝忠, 陈昭辉. 2023年度肉牛牦牛产业与技术发展报告. 中国畜牧杂志, 2024, 60(3): 335-338.
	CAO B H, CAO J M, LI J Y, GUO A Z, WANG Z S, SUN B Z, CHEN Z H. Report on the development of beef cattle and yak industry and technology in 2023. Chinese Journal of Animal Science, 2024, 60(3): 335-338. (in Chinese)
[2]	ARCHER J A, BARWICK S A, GRASER H U. Economic evaluation of beef cattle breeding schemes incorporating performance testing of young bulls for feed intake. Australian Journal of Experimental Agriculture, 2004, 44(5): 393. doi: 10.1071/EA02054
[3]	张帆, 周梦婷, 唐湘方, 刘民泽, 杨振刚, 熊本海. 规模化肉牛场数字化管控平台的开发与应用: 以阳信亿利源5G数字化牧场为例. 农业大数据学报, 2024, 6(1): 68-81. doi: 10.19788/j.issn.2096-6369.000009
	ZHANG F, ZHOU M T, TANG X F, LIU M Z, YANG Z G, XIONG B H. Development and application of digital control platform in largescale beef cattle farm: Take the 5G digital ranch of Yangxin Yi Liyuan Halal Meat Co., Ltd as an example. Journal of Agricultural Big Data, 2024, 6(1): 68-81. (in Chinese) doi: 10.19788/j.issn.2096-6369.000009
[4]	罗西尔, 卢小龙, 刘庆友, 崔奎青. 肉牛智慧养殖技术研究进展. 华南农业大学学报, 2024, 45(5): 661-671.
	LUO X E, LU X L, LIU Q Y, CUI K Q. Research progress on intelligent farming techniques of beef cattle. Journal of South China Agricultural University, 2024, 45(5): 661-671. (in Chinese)
[5]	杨亮, 张帆, 王辉, 熊本海. 北京地区家畜养殖数字化现状. 中国乳业, 2023(1): 35-40.
	YANG L, ZHANG F, WANG H, XIONG B H. Digital status of livestock breeding in Beijing. China Dairy, 2023(1): 35-40. (in Chinese)
[6]	万彬彬, 贾建刚, 邓锐强, 何江, 高江璐, 张燕茹, 李帅, 韩建欣. 肉牛生产性能测定问题分析. 中国畜禽种业, 2023, 19(6): 38-42.
	WAN B B, JIA J G, DENG R Q, HE J, GAO J L, ZHANG Y R, LI S, HAN J X. Analysis on the problems of beef cattle production performance measurement. The Chinese Livestock and Poultry Breeding, 2023, 19(6): 38-42. (in Chinese)
[7]	TASDEMIR S, URKMEZ A, INAL S. Determination of body measurements on the Holstein cows using digital image analysis and estimation of live weight with regression analysis. Computers and Electronics in Agriculture, 2011, 76(2): 189-197. doi: 10.1016/j.compag.2011.02.001
[8]	RUCHAY A, KOBER V, DOROFEEV K, KOLPAKOV V, MIROSHNIKOV S. Accurate body measurement of live cattle using three depth cameras and non-rigid 3-D shape recovery. Computers and Electronics in Agriculture, 2020, 179: 105821. doi: 10.1016/j.compag.2020.105821
[9]	刘波, 朱伟兴, 纪滨, 马长华. 基于射线轮廓点匹配的生猪红外与可见光图像自动配准. 农业工程学报, 2013, 29(2): 153-160, 297.
	LIU B, ZHU W X, JI B, MA C H. Automatic registration of IR and optical pig images based on contour match of radial line feature points. Transactions of the Chinese Society of Agricultural Engineering, 2013, 29(2): 153-160, 297. (in Chinese)
[10]	王少华. 基于视频分析和深度学习的奶牛爬跨行为检测方法研究[D]. 杨凌: 西北农林科技大学, 2021.
	WANG S H. Detection methods of cow mounting behavior based on video analysis and deep learning[D]. Yangling: Northwest A & F University, 2021. (in Chinese)
[11]	杨胜楠, 赵建敏, 杨梅, 赵宇飞. 基于去偏置项SoftMax和紧致度量损失函数的牛脸识别方法. 黑龙江畜牧兽医, 2024(4): 36-42.
	YANG S N, ZHAO J M, YANG M, ZHAO Y F. Cattle face recognition method based on debiased term SoftMax and compact metric loss function. Heilongjiang Animal Science and Veterinary Medicine, 2024(4): 36-42. (in Chinese)
[12]	PRETTO A, SAVIO G, GOTTARDO F, UCCHEDDU F, CONCHERI G. A novel low-cost visual ear tag based identification system for precision beef cattle livestock farming. Information Processing in Agriculture, 2024, 11(1): 117-126. doi: 10.1016/j.inpa.2022.10.003
[13]	张永芳, 黎亮, 董丽虹, 王海斗, 王朋, 谢向宇. RFID传感标签制备工艺研究进展. 材料导报, 2023, 37(22): 47-56.
	ZHANG Y F, LI L, DONG L H, WANG H D, WANG P, XIE X Y. Research progress of RFID sensor tag preparation technology. Materials Reports, 2023, 37(22): 47-56. (in Chinese)
[14]	黄彬, 朱立学. RFID技术在农业生产领域的应用现状与展望. 现代农业装备, 2024, 45(6): 56-62.
	HUANG B, ZHU L X. Main application of RFID technology in agricultural production. Modern Agricultural Equipment, 2024, 45(6): 56-62. (in Chinese)
[15]	熊本海, 钱平, 罗清尧, 吕健强. 基于奶牛个体体况的精细饲养方案的设计与实现. 农业工程学报, 2005, 21(10): 118-123.
	XIONG B H, QIAN P, LUO Q Y, LÜ J Q. Design and realization of solution to precision feeding of dairy cattle based on single body status. Transactions of the Chinese Society of Agricultural Engineering, 2005, 21(10): 118-123. (in Chinese)
[16]	AWAD A I. From classical methods to animal biometrics: a review on cattle identification and tracking. Computers and Electronics in Agriculture, 2016, 123: 423-435. doi: 10.1016/j.compag.2016.03.014
[17]	许兴时, 王云飞, 邓红兴, 宋怀波. 基于跨模态共享特征学习的夜间牛脸识别方法. 华南农业大学学报, 2024, 45(5): 793-801.
	XU X S, WANG Y F, DENG H X, SONG H B. Nighttime cattle face recognition based on cross-modal shared feature learning. Journal of South China Agricultural University, 2024, 45(5): 793-801. (in Chinese)
[18]	ANDREW W, GAO J, MULLAN S, CAMPBELL N, DOWSEY A W, BURGHARDT T. Visual identification of individual Holstein- Friesian cattle via deep metric learning. Computers and Electronics in Agriculture, 2021, 185: 106133. doi: 10.1016/j.compag.2021.106133
[19]	OKURA F, IKUMA S, MAKIHARA Y, MURAMATSU D, NAKADA K, YAGI Y. RGB-D video-based individual identification of dairy cows using gait and texture analyses. Computers and Electronics in Agriculture, 2019, 165: 104944. doi: 10.1016/j.compag.2019.104944
[20]	LU Y Z, WENG Z, ZHENG Z Q, ZHANG Y, GONG C L. Algorithm for cattle identification based on locating key area. Expert Systems with Applications, 2023, 228: 120365. doi: 10.1016/j.eswa.2023.120365
[21]	WENG Z, MENG F S, LIU S Q, ZHANG Y, ZHENG Z Q, GONG C L. Cattle face recognition based on a Two-Branch convolutional neural network. Computers and Electronics in Agriculture, 2022, 196: 106871. doi: 10.1016/j.compag.2022.106871
[22]	童慧琳, 李琦. 基于牛耳标OCR的牛身份识别研究. 计算机时代, 2023(10): 12-16.
	TONG H L, LI Q. Research on cattle identification based on cattle ear tag OCR. Computer Era, 2023(10): 12-16. (in Chinese)
[23]	LI G M, ERICKSON G E, XIONG Y J. Individual beef cattle identification using muzzle images and deep learning techniques. Animals, 2022, 12(11): 1453. doi: 10.3390/ani12111453
[24]	CHEN W K, LEE J C, HAN W Y, SHIH C K, CHANG K C. Iris recognition based on bidimensional empirical mode decomposition and fractal dimension. Information Sciences, 2013, 221: 439-451. doi: 10.1016/j.ins.2012.09.021
[25]	盛大玮, 何孝富, 吕岳. 基于最小二乘原理的牛眼虹膜分割方法. 中国图象图形学报, 2009, 14(10): 2132-2136.
	SHENG D W, HE X F, LÜ Y. A cattle iris segmentation method based on least square principle. Journal of Image and Graphics, 2009, 14(10): 2132-2136. (in Chinese)
[26]	魏征. 基于全局和局部特征相结合的不完美牛眼虹膜识别技术研究[D]. 南京: 东南大学, 2017.
	WEI Z. Research on imperfect bivine iris recognition based on combination of local features and global features[D]. Nanjing: Southeast University, 2017. (in Chinese)
[27]	SAYGILI A, CIHAN P, ERMUTLU C Ş, AYDIN U, AKSOY Ö. CattNIS: Novel identification system of cattle with retinal images based on feature matching method. Computers and Electronics in Agriculture, 2024, 221: 108963. doi: 10.1016/j.compag.2024.108963
[28]	周意, 毛宽民. 基于YOLO-Unet组合网络的牛只个体识别方法研究. 计算机科学, 2025, 52(4): 194-201.
	ZHOU Y, MAO K M. Research on individual identification of cattle based on YOLO-unet combined network. Computer Science, 2025, 52(4): 194-201. (in Chinese)
[29]	叶孟珂, 李宝山, 杨梅, 李琦. 基于深度学习的牛脸识别系统设计. 黑龙江畜牧兽医, 2024(4): 43-48.
	YE M K, LI B S, YANG M, LI Q. Design of cattle face recognition system based on deep learning. Heilongjiang Animal Science and Veterinary Medicine, 2024(4): 43-48. (in Chinese)
[30]	李征. 局部遮挡条件下的西门塔尔肉牛面部识别算法研究[D]. 呼和浩特: 内蒙古大学, 2022.
	LI Z. Research on face recognition algorithm of Simmental beef cattle under the condition of partial occlusion[D]. Hohhot: Inner Mongolia University, 2022. (in Chinese)
[31]	胡立俊, 李旭, 李国亮. 基于改进YOLOv8n的安格斯牛面部识别. 华中农业大学学报, 2025, 44(2): 39-48.
	HU L J, LI X, LI G L. Facial recognition of Angus cattle based on the improved YOLOv8n. Journal of Huazhong Agricultural University, 2025, 44(2): 39-48. (in Chinese)
[32]	张宇. 基于深度学习的肉牛体侧识别方法研究[D]. 包头: 内蒙古科技大学, 2023.
	ZHANG Y. Research on beef cattle body side recognition method based on deep learning[D]. Baotou: Inner Mongolia University of Science & Technology, 2023. (in Chinese)
[33]	CASEY T M, PLAUT K. Circadian clocks and their integration with metabolic and reproductive systems: our current understanding and its application to the management of dairy cows. Journal of Animal Science, 2022, 100(10): skac233.
[34]	王苏皖, 冯小芳, 佟丽佳, 李德生, 王瑜, 封元, 蒋秋斐, 徐军, 陈亚飞, 顾亚玲, 张娟. 安格斯牛断奶阶段生长发育性状的遗传参数估计. 中国畜牧兽医, 2024, 51(6): 2517-2523. doi: 10.16431/j.cnki.1671-7236.2024.06.026
	WANG S W, FENG X F, TONG L J, LI D S, WANG Y, FENG Y, JIANG Q F, XU J, CHEN Y F, GU Y L, ZHANG J. Genetic parameter estimation of growth and developmental traits in Angus cattle during weaning stage. China Animal Husbandry & Veterinary Medicine, 2024, 51(6): 2517-2523. (in Chinese)
[35]	马建飞, 白德日根, 金生云, 李丽杰, 包俊杰, 牛化欣. 草原放牧系统条件下华西牛0-9月龄生长、体尺和肉用指数研究. 中国畜牧杂志, 2024, 60(10): 355-358.
	MA J F, BAI D, JIN S Y, LI L J, BAO J J, NIU H X. Study on growth, body size and meat index of Huaxi cattle aged 0-9 months under grassland grazing system. Chinese Journal of Animal Science, 2024, 60(10): 355-358. (in Chinese)
[36]	KAWASUE K, WIN K D, YOSHIDA K, TOKUNAGA T. Black cattle body shape and temperature measurement using thermography and KINECT sensor. Artificial Life and Robotics, 2017, 22(4): 464-470. doi: 10.1007/s10015-017-0373-2
[37]	LI G X, LIU X L, MA Y F, WANG B B, ZHENG L H, WANG M J. Body size measurement and live body weight estimation for pigs based on back surface point clouds. Biosystems Engineering, 2022, 218: 10-22. doi: 10.1016/j.biosystemseng.2022.03.014
[38]	XIAO J X, LIU G, WANG K J, SI Y S. Cow identification in free-stall barns based on an improved Mask R-CNN and an SVM. Computers and Electronics in Agriculture, 2022, 194: 106738. doi: 10.1016/j.compag.2022.106738
[39]	LI R, WEN Y C, ZHANG S J, XU X S, MA B L, SONG H B. Automated measurement of beef cattle body size via key point detection and monocular depth estimation. Expert Systems with Applications, 2024, 244: 123042. doi: 10.1016/j.eswa.2023.123042
[40]	岳萌萌, 舒涛, 王嘉博, 刘利, 王鹏, 赵晓川, 许珊珊, 王春薇, 柴孟龙, 孙芳, 钟金城. 基于Image 2和岭回归模型估测肉牛体尺、体重. 黑龙江畜牧兽医, 2020(22): 41-43, 49.
	YUE M M, SHU T, WANG J B, LIU L, WANG P, ZHAO X C, XU S S, WANG C W, CHAI M L, SUN F, ZHONG J C. Estimation of beef cattle body length and body weight based on the Image 2 and ridge regress model. Heilongjiang Animal Science and Veterinary Medicine, 2020(22): 41-43, 49. (in Chinese)
[41]	张帅. 基于深度学习的晋南牛体尺及体重预测方法研究[D]. 太谷: 山西农业大学, 2022.
	ZHANG S. Study on prediction method of body size and weight of Jinnan cattle based on deep learning[D]. Taigu: Shanxi Agricultural University, 2022. (in Chinese)
[42]	HUANG L W, LI S Q, ZHU A Q, FAN X Y, ZHANG C Y, WANG H Y. Non-contact body measurement for Qinchuan cattle with LiDAR sensor. Sensors, 2018, 18(9): 3014. doi: 10.3390/s18093014
[43]	LI J W, MA W H, BAI Q, TULPAN D, GONG M L, SUN Y, XUE X L, ZHAO C J, LI Q F. A posture-based measurement adjustment method for improving the accuracy of beef cattle body size measurement based on point cloud data. Biosystems Engineering, 2023, 230: 171-190. doi: 10.1016/j.biosystemseng.2023.04.014
[44]	ZHANG Q, HOU Z X, HUANG L W, WANG F Y, MENG H Y. Reparation with moving least squares sampling and extraction of body sizes of beef cattle from unilateral point clouds. Computers and Electronics in Agriculture, 2024, 224: 109208. doi: 10.1016/j.compag.2024.109208
[45]	HOU Z X, HUANG L W, ZHANG Q, MIAO Y S. Body weight estimation of beef cattle with 3D deep learning model: PointNet++. Computers and Electronics in Agriculture, 2023, 213: 108184. doi: 10.1016/j.compag.2023.108184
[46]	LI J W, MA W H, LI Q F, ZHAO C J, TULPAN D, YANG S, DING L Y, GAO R H, YU L G, WANG Z Q. Multi-view real-time acquisition and 3D reconstruction of point clouds for beef cattle. Computers and Electronics in Agriculture, 2022, 197: 106987. doi: 10.1016/j.compag.2022.106987
[47]	HOU Z X, ZHANG Q, ZHANG B, ZHANG H M, HUANG L W, WANG M L. CattlePartNet: an identification approach for key region of body size and its application on body measurement of beef cattle. Computers and Electronics in Agriculture, 2025, 232: 110013. doi: 10.1016/j.compag.2025.110013
[48]	GUO H, LI Z B, MA Q, ZHU D H, SU W, WANG K, MARINELLO F. A bilateral symmetry based pose normalization framework applied to livestock body measurement in point clouds. Computers and Electronics in Agriculture, 2019, 160: 59-70. doi: 10.1016/j.compag.2019.03.010
[49]	QIAO Y L, TRUMAN M, SUKKARIEH S. Cattle segmentation and contour extraction based on Mask R-CNN for precision livestock farming. Computers and Electronics in Agriculture, 2019, 165: 104958. doi: 10.1016/j.compag.2019.104958
[50]	XU B B, WANG W S, FALZON G, KWAN P, GUO L F, CHEN G P, TAIT A, SCHNEIDER D. Automated cattle counting using Mask R-CNN in quadcopter vision system. Computers and Electronics in Agriculture, 2020, 171: 105300. doi: 10.1016/j.compag.2020.105300
[51]	邓由飞, 代俊芳, 郭晓旭, 任丽萍, 孟庆翔. 利用体尺指标估测肉牛自然体重的准确性研究. 中国畜牧杂志, 2015, 51(S1): 141-143.
	DENG Y F, DAI J F, GUO X X, REN L P, MENG Q X. Study on the accuracy of estimating natural weight of beef cattle by body size index. Chinese Journal of Animal Science, 2015, 51(S1): 141-143.
[52]	NISBET H, LAMBE N, MILLER G A, DOESCHL-WILSON A, BARCLAY D, WHEATON A, DUTHIE C A. On-farm 3D images of beef cattle for the prediction of carcass classification traits and cold carcass weight. Animal, 2025, 19(6): 101529. doi: 10.1016/j.animal.2025.101529
[53]	GJERGJI M, DE MORAES WEBER V, OTAVIO CAMPOS SILVA L, DA COSTA GOMES R, DE ARAUJO T L A C, PISTORI H, ALVAREZ M. Deep learning techniques for beef cattle body weight prediction. 2020 International Joint Conference on Neural Networks (IJCNN). July 19-24, 2020. Glasgow, United Kingdom. IEEE, 2020: 1-8.
[54]	MILLER G A, HYSLOP J J, BARCLAY D, EDWARDS A, THOMSON W, DUTHIE C A. Using 3D imaging and machine learning to predict liveweight and carcass characteristics of live finishing beef cattle. Frontiers in Sustainable Food Systems, 2019, 3: 30. doi: 10.3389/fsufs.2019.00030
[55]	WEBER V A M, DE LIMA WEBER F, DA SILVA OLIVEIRA A, ASTOLFI G, MENEZES G V, DE ANDRADE PORTO J V, REZENDE F P C, DE MORAES P H, MATSUBARA E T, MATEUS R G, et al. Cattle weight estimation using active contour models and regression trees Bagging. Computers and Electronics in Agriculture, 2020, 179: 105804. doi: 10.1016/j.compag.2020.105804
[56]	KAMCHEN S G, FERNANDES DOS SANTOS E, LOPES L B, VENDRUSCULO L G, CONDOTTA I C F S. Application of depth sensor to estimate body mass and morphometric assessment in Nellore heifers. Livestock Science, 2021, 245: 104442. doi: 10.1016/j.livsci.2021.104442
[57]	LOS S, MÜCHER C A, KRAMER H, FRANKE G J, KAMPHUIS C. Estimating body dimensions and weight of cattle on pasture with 3D models from UAV imagery. Smart Agricultural Technology, 2023, 4: 100167. doi: 10.1016/j.atech.2022.100167
[58]	COMINOTTE A, FERNANDES A F A, DOREA J R R, ROSA G J M, LADEIRA M M, VAN CLEEF E H C B, PEREIRA G L, BALDASSINI W A, MACHADO NETO O R. Automated computer vision system to predict body weight and average daily gain in beef cattle during growing and finishing phases. Livestock Science, 2020, 232: 103904. doi: 10.1016/j.livsci.2019.103904
[59]	陈长博, 路平乐, 苏涛龙, 狄亚鹏, 高永权, 杨永慧, 杨雅楠, 赵生国. 早胜牛母牛体重与体尺性状相关回归分析. 黑龙江畜牧兽医, 2024(8): 31-36.
	CHEN C B, LU P L, SU T L, DI Y P, GAO Y Q, YANG Y H, YANG Y N, ZHAO S G. Correlation regression analysis between body weight and body size traits in Zaosheng cows. Heilongjiang Animal Science and Veterinary Medicine, 2024(8): 31-36. (in Chinese)
[60]	蓝雷斌. 基于机器视觉的牛只体尺体重估测算法研究[D]. 杭州: 杭州电子科技大学, 2024.
	LAN L B. Estimation algorithm of cattle body size and weight based on machine vision[D]. Hangzhou: Hangzhou Dianzi University, 2024. (in Chinese)
[61]	WANGCHUK K, WANGDI J, MINDU M. Comparison and reliability of techniques to estimate live cattle body weight. Journal of Applied Animal Research, 2018, 46(1): 349-352. doi: 10.1080/09712119.2017.1302876
[62]	刘建明, 杨光维, 李涛, 潘国龙, 乃比江, 范守民. 应用约翰逊法估测美新杂交褐牛体重. 中国牛业科学, 2019, 45(6): 23-25.
	LIU J M, YANG G W, LI T, PAN G L, NAI B J, FAN S M. Estimation of the body weight of Mei-Xin hybrid brown cattle by Johnson method. China Cattle Science, 2019, 45(6): 23-25. (in Chinese)
[63]	DE OLIVEIRA BEZERRA A, DE MORAES WEBER V A, DE LIMA WEBER F, ALVES DE ARRUDA Y, DA COSTA GOMES R, HIROKAWA HIGA G T, PISTORI H, GONÇALVES MATEUS R. Exploring cluster analysis in Nelore cattle visual score attribution. Smart Agricultural Technology, 2024, 8: 100489. doi: 10.1016/j.atech.2024.100489
[64]	赵黄青, 杜蕾, 赵家豪, 马伟东, 王武生, 张金川, 樊安平, 雷初朝, 黄永震. 秦川牛体尺和体重性状通径分析以及多元回归方程的建立. 中国畜牧杂志, 2024, 60(10): 110-113.
	ZHAO H Q, DU L, ZHAO J H, MA W D, WANG W S, ZHANG J C, FAN A P, LEI C Z, HUANG Y Z. Path analysis of body size and weight traits of Qinchuan cattle and establishment of multiple regression equation. Chinese Journal of Animal Science, 2024, 60(10): 110-113. (in Chinese)
[65]	SOUSA R V, TABILE R A, INAMASU R Y, MARTELLO L S. Evaluating a LiDAR sensor and artificial neural network based-model to estimate cattle live weight. //10th International Livestock Environment Symposium (ILES X). American Society of Agricultural and Biological Engineers, 2018: 28-37. DOI:10.13031/iles.18-004.
[66]	DE ANDRADE D K B, VÉRAS A S C, DE ANDRADE FERREIRA M, DOS SANTOS M V F, DE MELO W S, PEREIRA K P. Body composition and net protein and energy requirements for weight gain of crossbred dairy cattle in grazing. Revista Brasileira de Zootecnia, 2009, 38(4): 746-751. doi: 10.1590/S1516-35982009000400022
[67]	张永根. 奶牛采食行为的智能化监测方法及其应用研究进展. 饲料工业, 2020, 41(9): 1-6.
	ZHANG Y G. Research progress on intelligent monitoring methods for monitoring feeding behavior of dairy cows. Feed Industry, 2020, 41(9): 1-6. (in Chinese)
[68]	CHAPINAL N, VEIRA D M, WEARY D M, VON KEYSERLINGK M A G. Technical note: Validation of a system for monitoring individual feeding and drinking behavior and intake in group-housed cattle. Journal of Dairy Science, 2007, 90(12): 5732-5736. pmid: 18024766
[69]	张宏鸣, 武杰, 李永恒, 李书琴, 王红艳, 宋荣杰. 多目标肉牛进食行为识别方法研究. 农业机械学报, 2020, 51(10): 259-267.
	ZHANG H M, WU J, LI Y H, LI S Q, WANG H Y, SONG R J. Recognition method of feeding behavior of multi-target beef cattle. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(10): 259-267. (in Chinese)
[70]	周雅婷, 许童羽, 陈春玲, 周云成, 姜美羲. 基于神经网络算法的肉牛采食行为检测方法. 沈阳农业大学学报, 2016, 47(6): 752-757.
	ZHOU Y T, XU T Y, CHEN C L, ZHOU Y C, JIANG M X. Detection method of beef cattle feeding behavior based on neural network algorithm. Journal of Shenyang Agricultural University, 2016, 47(6): 752-757. (in Chinese)
[71]	周雅婷. 肉牛采食行为识别与采食量模型研究[D]. 沈阳: 沈阳农业大学, 2018.
	ZHOU Y T. Study on the identification of eating behavior of beef cattle and the model of feed intake[D]. Shenyang: Shenyang Agricultural University, 2018. (in Chinese)
[72]	WANG X J, DAI B S, WEI X L, SHEN W Z, ZHANG Y G, XIONG B H. Vision-based measuring method for individual cow feed intake using depth images and a Siamese network. International Journal of Agricultural and Biological Engineering, 2023, 16(3): 233-239. doi: 10.25165/j.ijabe.20231603.7985
[73]	LASSEN J, THOMASEN J R, BORCHERSEN S. Repeatabilities of individual measures of feed intake and body weight on in-house commercial dairy cattle using a 3-dimensional camera system. Journal of Dairy Science, 2023, 106(12): 9105-9114. doi: 10.3168/jds.2022-23177
[74]	QIAO Y L, KONG H, CLARK C, LOMAX S, SU D, EIFFERT S, SUKKARIEH S. Intelligent perception for cattle monitoring: a review for cattle identification, body condition score evaluation, and weight estimation. Computers and Electronics in Agriculture, 2021, 185: 106143. doi: 10.1016/j.compag.2021.106143
[75]	杨亮, 高华杰, 夏阿林, 熊本海. 智能化猪场数字化管控平台创制及应用. 农业大数据学报, 2022, 4(3): 135-146. doi: 10.19788/j.issn.2096-6369.220321
	YANG L, GAO H J, XIA A L, XIONG B H. Creation and application of an intelligent pig farm digital control platform. Journal of Agricultural Big Data, 2022, 4(3): 135-146. (in Chinese) doi: 10.19788/j.issn.2096-6369.220321
[76]	HE C, QIAO Y L, MAO R, LI M, WANG M L. Enhanced LiteHRNet based sheep weight estimation using RGB-D images. Computers and Electronics in Agriculture, 2023, 206: 107667. doi: 10.1016/j.compag.2023.107667
[77]	NYALALA I, OKINDA C, CHEN K J, KOROHOU T, NYALALA L, QI C. Weight and volume estimation of poultry and products based on computer vision systems: A review. Poultry Science, 2021, 100(5): 101072. doi: 10.1016/j.psj.2021.101072
[78]	KAMILARIS A, PRENAFETA-BOLDÚ F X. Deep learning in agriculture: A survey. Computers and Electronics in Agriculture, 2018, 147: 70-90. doi: 10.1016/j.compag.2018.02.016
[79]	VAN HERTEM T, ROOIJAKKERS L, BERCKMANS D, PEÑA FERNÁNDEZ A, NORTON T, BERCKMANS D, VRANKEN E. Appropriate data visualisation is key to Precision Livestock Farming acceptance. Computers and Electronics in Agriculture, 2017, 138: 1-10. doi: 10.1016/j.compag.2017.04.003
[80]	曹淑泉, 王宇, 尚泽琦, 张鹏. 基于视觉三维重建的肉牛定价系统. 物联网技术, 2025, 15(10): 94-96.
	CAO S Q, WANG Y, SHANG Z Q, ZHANG P. Beef cattle pricing system based on visual 3D reconstruction. Internet of Things Technologies, 2025, 15(10): 94-96. (in Chinese)

监测部位 Monitoring site	监测设备 Monitoring equipment	主要识别算法 Main identification method	准确率 Accuracy rate	特点 Characteristic	参考文献 Reference
背部 Back	MD-1004NS MD-DVR41	Bagging	相对误差2.67%，误差值13.44 kg With relative error 2.67%, and average error value 13.44 kg	采用背部二维图像估测体重 Estimate the body weight by two-dimensional images of the back	[55]
侧面 Body side	IFM O3D303	PointNet++确定体尺测定关键点 PointNet++ determines the key points for measuring the body sizes	相对误差3.2%，误差值10.2 kg With relative error 3.2%, and average error value 10.2 kg	依据深度相机测定的体尺估算体重 Estimate the body weight based on the body size measured by the depth camera	[45]
背部 Back	Intel RealSense D435i	MLP	相对误差3.13% With relative error 3.13%	利用体积估算体重 Estimate body weight by volume	[56]
侧面 Body side	LiDAR	Potree Desktop	相对误差6.3% With relative error 6.3%	在无人机上搭载深度相机估测体重 Estimate body weight by a depth camera installed on the unmanned aerial vehicle	[57]
背部 Back	Kinect® model 1473	ANN	体重相对误差< 5.32%，日增重误差均< 0.10 kg·d^-1 With relative error of body weight < 5.32%, and the error value of average daily gain all < 0.10 kg·d^-1	同一模型预测肉牛不同阶段体重，并计算日增重 Estimate the body weight and average daily gain of beef cattle at different stages by the same model	[58]