摘要
针对利用插管式机器人在巡视络筒机设备时,存在无法检测储管纱圆盘均匀排布孔内是否存在管纱的问题,提出了一种基于图像识别的空管状态检测方法.结合络筒机及插管式机器人工作机制,将集嵌入式控制系统、相机模组于一体的空管检测机构固定在竖直机械立柱上,并随插管式机器人做水平往复直线运动.嵌入式控制系统通过捕获插管式机器人发送的锭位信号,实现该锭位储管纱圆盘图像采集,通过学习、调节、分块定域、矫正等准备环节及分级处理环节进行图像处理,并将处理后各均匀排布孔内管纱的实时情况经由通信接口传输至插管式机器人.最后搭建实验平台,进行空管状态检测稳定性测试.结果表明,本文方法检测效果稳定、准确度高,且成本低、安装简单,便于实际工程应用.
Abstract
Objective Manual tube change in winding requires workers for continuous inspection to observe the quantity of tube yarn used in each spindle,which is time-consuming and laborious.It is now possible to solve this problem by using intubation robot to carry out intubation.In this research,an empty tube state detection method based on image recognition is proposed to monitor whether a yarn tube is in each position evenly arranged in the tube storage discs.Method The empty tube inspection mechanism embedded with a control system and a camera module is fixed on the vertical mechanical column,moving horizontally with the intubation robot.By capturing the spindle position signal sent by the intubation robot,the embedded system can complete the image acquisition of the spindle position,carry out image processing by the learning,adjustment,block and localization,correction processing links,and transmit the real-time situation of the tube yarn in each row of holes after processing to the intubation robot by the communication interface.Results After the test preparation stage is completed,the image of the storage tube yarn disc at the best shooting position is captured according to the external interrupt signal triggered by the intubation robot,and the empty tube detection algorithm is called for real-time operation.In order to increase the visibility of the detection results,the identified tube storage yarn disc,tube yarn hole and tube are marked with red,blue and green outlines respectively.In order to facilitate the test and analysis of the experimental detection results,the center position and radius of the storage tube yarn disc,storage tube yarn hole,and yarn tube is stored in the external flash memory chip in real-time according to the processing results of empty tube detection algorithm.The tubes in the storage tube yarn disc are increased and decreased repeatedly and tested repeatedly.Finally,the detection results are calculated for statistics.It can be seen from the start recognition time and the end recognition time that the whole recognition process is only 160 ms,and the image processing speed is at the millisecond level.The recognition result data is sorted out(Fig.13).According to the statistical data,only 2 storage tube yarn holes in this storage tube yarn disc have tube yarn,which is consistent with the actual situation.In order to verify the stability of the test results,the tube yarns in the 1-9 storage tube yarn holes in the 1-9 spindle storage tube yarn discs are taken out in turn,and tube yarns are placed in the remaining storage tube yarn holes,and the 10 and 11 spindle storage tube yarn discs are placed in the full tube and empty tube states respectively(Tab.1).The detection device performs reciprocating detection with the intubation robot for a long time of 24 h,and the detection results are stored in the external flash memory chip in real-time.After statistical analysis,it is found that the data is periodic,and the test results are in line with the actual situation,meeting the requirements of the intubation robot for the stability and accuracy of the empty pipe detection mechanism(Fig.13).Conclusion At present,the testing agency has conducted a pilot test on the intubation robot of a textile enterprise in Zhejiang.The practical application results show that the empty tube detection system has the advantages of fast image processing speed,high recognition accuracy and stability,easy installation and low cost,and can meet the requirements of the intubation robot for the empty tube detection of the storage yarn disc of the yarn bank type automatic winder.
基金项目
浙江省博士后科研项目择优资助一等资助项目(ZJ2021038)
浙江省"尖兵领雁"研发攻关计划资助项目(2022C01065)
浙江省"尖兵领雁"研发攻关计划资助项目(2022C01202)
浙江理工大学科研启动基金(23242083-Y)
NETL
NSTL
维普
万方数据