基于深度学习的布匹瑕疵数据可视化系统  

Deep Learning Based Cloth Defect Data Visualization System

作　　者：刘建彬 LIU Jianbin(Guangdong Ocean University,Zhanjiang Guangdong 524088,China)

机构地区：[1]广东海洋大学,广东湛江524088

出　　处：《信息与电脑》2023年第4期198-201,共4页Information & Computer

摘　　要：在布匹瑕疵检测领域中,传统的瑕疵检测手段主要是采用人工检测的方式,其弊端很多,如检出率低等。为解决人工检测的各类问题,设计了基于深度学习的布匹瑕疵数据可视化系统,视觉算法部分采用改进后YOLOv5模型作为瑕疵检测模型对布匹的20种瑕疵进行检测,能够达到准确率95.7%,检测速度为33 ms,并且结合工业机械臂使用高清摄像头来进行图像以及视频的采集,在经过深度学习输出的相关数据经过后端并在前端平台上进行展示,从而实现自动化的检测及智能分析平台。In the field of fabric defect detection,the traditional defect detection means is mainly used in manual detection,which has many drawbacks,such as low detection rate,etc.;in order to solve the various problems of manual detection,a deep learning based fabric defect data visualization system is designed,and the visual algorithm part uses the improved YOLOv5 model as the defect detection model to detect 20 kinds of defects in fabric,which can achieve an accuracy rate of 95.7%accuracy,33ms detection speed,and combined with industrial robotic arm using high-definition camera for image and video acquisition,after deep learning output of relevant data through the back-end and on the front-end platform for display,so as to achieve an automated detection and intelligent analysis platform.

关 键 词：YOLOv5算法 机械臂 数据可视化 深度学习 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]