摘要
为提高中空玻璃上下片机械手工作效率,解决钢化玻璃透光性强导致视觉检测尺寸较难的问题,设计了基于视觉的机械手进给距离测量与玻璃尺寸检测系统。首先根据系统工作需求对系统的关键部件进行选型,设计了系统主体硬件结构;其次提出了玻璃尺寸和机械手进给距离检测方法,即在对玻璃边缘与滚轮颜色特征提取的基础上,采用改进骨架提取算法进行玻璃尺寸测量,并基于卡尔曼滤波器利用滚轮长宽比和像素区域面积实现进给距离实时检测;最后设计了系统整体功能软件。该测量系统实现了玻璃尺寸和机械手进给距离的有效测量,具有较高的准确性与快速性,大大提高了测量效率。
In order to improve the working efficiency of the upper and lower manipulator of hollow glass and solve the problem that it is difficult to detect the size by vision due to the strong light transmittance of tempered glass,a vision based manipulator feed distance measurement and glass size detection system was designed.Firstly,the key components of the system were selected according to the working requirements of the system,and the main hardware structure of the system was designed.Secondly,the detection method of glass size and manipulator feed distance was proposed,that is,based on the color feature extraction of glass edge and roller,the improved skeleton extraction algorithm was used to measure the glass size,and the real-time detection of feed distance was realized by using roller aspect ratio and pixel area based on Kalman filter.Finally,the overall function software of the system was designed.The measurement system realizes the effective measurement of glass size and manipulator feed distance,has high accuracy and rapidity,and greatly improves the measurement efficiency.
作者
杜太行
李亚杰
孙曙光
刘辉
梁杰
安春晖
DU Tai-hang;LI Ya-jie;SUN Shu-guang;LIU Hui;LIANG Jie;AN Chun-hui(School of Artificial Intelligence,Hebei University of Technology,Tianjin 300130,China;Hebei Industrial Manipulator Control and Reliability Technology Innovation Center(Hebei University of Water Resources and Electric Engineering),Cangzhou 061000,China;Hebei University of Water Resources and Electric Engineering,Cangzhou 061000,China)
出处
《仪表技术与传感器》
CSCD
北大核心
2022年第9期77-81,88,共6页
Instrument Technique and Sensor
基金
河北省工业机械手控制与可靠性技术创新中心(河北水利电力学院)开放课题资助项目(JXKF2101)
2020年河北省高等学校基本科研业务费研究项目(SYKY2003)。
关键词
中空玻璃上下片机械手
视觉
改进骨架提取算法
卡尔曼滤波
hollow glass upper and lower manipulator
vision
improved skeleton extraction algorithm
Kalman filtering
