基于并丝机与深度学习的丝线运动研究及检测  

Research on Thread Motion and Detection Based on Parallel Thread Machine and Deep Learning

作　　者：张瑞琪 吴震宇[1,2] 张银龙 

机构地区：[1]浙江理工大学机械工程学院,浙江 杭州 [2]浙江理工大学新昌技术创新研究院,浙江 绍兴

出　　处：《建模与仿真》2023年第6期5236-5250,共15页Modeling and Simulation

摘　　要：并丝机是纺织机械常见的一种绕线设备,其绕线过程中伴随主轴的转动和排线架的往返将丝线有序的排列在线轴上,由于排线架与丝线运动的特点,导致并丝机丝线排列在线轴上时不同的区域丝线在线轴上的运动量存在差异。通过高速相机对并丝机丝线运动过程进行拍摄,后利用TEMA分析软件对丝线进行运动学分析。由于丝线在运动过程中存在差异,所以会造成堆叠以及交叉等瑕疵,基于Yolov8模型,深度学习和检测平台实现对丝线瑕疵的检测识别,目的是检测并丝机并丝的实际效果。The winding machine is common winding equipment in textile machinery. The winding process is accompanied by the rotation of the spindle and the round trip of the wire rack. Due to the charac-teristics of the movement of the wire rack and the thread, the amount of exercise of the thread in different areas of the wire is different when the thread of the wire drawing machine is arranged on the spool. A high-speed camera was used to capture the motion process of the silk thread of the drawing machine, and then the kinematics of the silk thread was analyzed by TEMA analysis soft-ware. Due to the differences in the movement of silk threads, defects such as stacking and crossing will be caused, based on the Yolov8 model, deep learning and detection platform to achieve the de-tection and identification of silk defects, the purpose is to detect the actual effect of silk combining machine.

关 键 词：并丝机丝线 运动学分析 高速相机 深度学习 Yolov8 

分 类 号：TM2[一般工业技术—材料科学与工程]