摘要
利用机器人可以大批量、不间断作业的优势,设计基于机器视觉和卷积神经网络的无人化智能装卸方法。使用机器视觉系统采集无人化智能装卸机器人装卸时的图像数据,通过自适应卷积神经网络的无人化智能装卸图像识别方法获得待装卸物体,将其在图像坐标系下的坐标值运用张氏标定法转化为机器人坐标系下的坐标值,根据转化结果采用快速扩展随机树算法驱动机器人运动到坐标位置,实现待装卸物体的无人化智能装卸。实验结果表明:该方法能识别出全部待装卸物体;标定所得重投影误差值小,无人化智能装卸机器人运动路径短,并能有效避开所有障碍物,能高效、精准地装卸全部待装卸物体。
Using the advantage that robots can work in large quantities and continuously,an unmanned intelligent loading and unloading method based on machine vision and convolutional neural network is studied.Based on the basic theory of machine vision and convolutional neural network,machine vision system is used to collect the image data of unmanned intelligent loading and unloading robot,and the object to be loaded and unloaded is obtained through the unmanned intelligent loading and unloading image recognition method based on adaptive convolutional neural network,coordinate value in the image coordinate system is transformed into the coordinate value in the robot coordinate system by Zhang's calibration method.According to the transformation results,timproved fast extended random tree algorithm is selected to drive robot to the coordinate position,so as to realize the unmanned intelligent loading and unloading of the object to be loaded and unloaded.The experimental results show that this method can identify all the objects to be loaded and unloaded.The re projection error values obtained from calibration is small,unmanned intelligent loading and unloading robot has a short motion path,can effectively avoid all obstacles,and can efficiently and accurately load and unload all objects to be loaded and unloaded.
作者
陶加贵
韩飞
汪伦
赵恒
TAO Jia-gui;HAN Fei;WANG Lun;ZHAO Heng(State Grid Jiangsu Electric Power Company Electric Power Research Institute,Nanjing 210000 China;State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 210000 China)
出处
《自动化技术与应用》
2024年第4期26-30,共5页
Techniques of Automation and Applications
基金
国家电网有限公司总部科技项目资助(1400-202118268A-0-0-00)。
关键词
机器视觉
卷积神经网络
无人化
智能装卸
机器人
张氏标定法
machine vision
Convolutional Neural Network
unmanned
intelligent loading and unloading
Robot
Zhang's calibration method
