计算机视觉技术在现代畜牧业中的应用与研究进展  

Advance and Application of Computer Vision Technology in Modern Animal Husbandry

作　　者：闫鹏飞 肖莎 张志勇 赵铖 刘聪 米凯臣 刘静峰 YAN Pengfei;XIAO Sha;ZHANG Zhiyong;ZHAO Cheng;LIU Cong;MI Kaichen;LIU Jingfeng(New Hope Liuhe Co.,Ltd./Key Laboratory of Feed and Livestock and Poultry Products Quality&Safety Control,Ministry of Agriculture,Chengdu 610023,China)

机构地区：[1]新希望六和股份有限公司/农业农村部饲料及畜禽产品质量安全控制重点实验室,四川成都610023

出　　处：《中国猪业》2024年第5期83-89,共7页China Swine Industry

基　　金：科技创新2030--“新一代人工智能”重大项目(2021ZD0113800)。

摘　　要：当前,实现对动物健康的连续和自动化监测,以及在发现动物异常时及时采取精准干预措施,已成为研究的热点。计算机视觉技术可以获取目标动物图像,并通过对所获取原始数据的分析,提取与目标行为相关的特征进行机器学习模型训练,从而实现动物异常行为的监测。本文综述了目标分类、目标检测、目标追踪、目标分割等计算机视觉技术在畜牧业中的应用,并分析了其应用原理,最后展望了计算机视觉技术在畜牧业中的应用前景及优化方向。Currently,the realization of continuous and automated monitoring of animal health,as well as timely and precise interventions when animal abnormalities were detected,had become a research hotspot.Computer vision technology could acquire target animal im-ages,and through the analysis of the acquired raw data,extract the features related to the target behavior for machine learning model training,so as to realize the monitoring of abnormal animal behavior.This paper reviewed the application of computer vision technology in animal husbandry such as target classification,target detection,target tracking,target segmentation,etc.,and analyzed the principle of its application,and finally looked forward to the prospect of the application of computer vision technology in animal husbandry and the direction of optimization.

关 键 词：养猪业 人工智能 计算机 视觉 目标 监测 畜牧业 研究进展 

分 类 号：S818.9[农业科学—畜牧学] TP391.4[农业科学—畜牧兽医]