基于计算机视觉技术的多品种急救包折叠手势库设计  

Design of Multi-Species First Aid Kit Folding Gesture Library Based on Computer Vision Technology

作　　者：赵继祥 杜宝江 ZHAO Jixiang;DU Baojiang(College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730050,China;Virtual Manufacturing Technology Research Institute,Shanghai University of Technology,Shanghai 200093,China)

机构地区：[1]兰州理工大学电气工程与信息工程学院,甘肃兰州730050 [2]上海理工大学虚拟制造技术研究院,上海200093

出　　处：《自动化仪表》2024年第8期117-121,共5页Process Automation Instrumentation

摘　　要：急救包作为医疗器械生产领域的通用产品之一,具有品种多、用途广、需求量激增的特点。针对急救包生产过程中手工折叠劳动负荷大、品质波动明显、生产效率低、操作标准缺乏等技术问题,分析各类急救器械和辅料码放工艺特征,并映射为折叠手势的技术工序组合。创新点包括:引入Leap Motion双目计算机视觉技术,构建双手空间模型和手势模型;采用卷积神经网络(CNN)建立手势关键点识别模型,形成急救包折叠手势的有序集合。采用图形数据库技术,设计存储各型号急救包的折叠手势库。通过某型号急救包生产案例,验证了设计的可行性。所设计的急救包折叠手势可转化为指挥机器人折叠操作的指令集合,实现急救包生产领域机器换人的智能制造模式。As one of the general products in the field of medical equipment production,first aid kit is characterized by many varieties,wide range of uses and surge in demand.Aiming at the technical problems such as large labor load of manual folding,obvious quality fluctuation,low production efficiency,and lack of operation standards,etc.,in the production process of first aid kits,the characteristics of various types of first aid instruments and auxiliary materials yardage process are analyzed,which is mapped to the combination of technical processes of folding gestures.The innovations include:the introduction of Leap Motion binocular computer vision technology,the construction of two-handed spatial model and gesture model;convolutional neural network(CNN)is used to build a gesture key point recognition model to form an ordered set of folding gestures for first aid kits.Adopt graphical database technology to design and store the folding gesture library of each model of first aid kit.The feasibility of the design is verified through the production case of a certain model of first aid kit.The designed folding gestures of the first aid kit can be converted into a collection of folding operation commands for the robot,which realizes the intelligent manufacturing mode of machine-for-man in the field of first aid kit production.

关 键 词：急救包 手势识别 Leap Motion 计算机视觉 手势库 卷积神经网络 

分 类 号：TH166[机械工程—机械制造及自动化]