An identification method of cheese yarn ends based on machine vision
To efficiently extract cheese yarn ends in the warping process of textile weaving and solve the problems of low work efficiency of manual and machine thread hunting,and complex devices,a method for identifying cheese yarn ends based on machine vision was proposed.Firstly,distortion correction was performed on the collected cheese yarn images,and a cheese yarn correction model was established based on perspective projection and local perspective transformation to obtain an ideal cheese side expansion image.Secondly,the 4-neighbor connection rule was used to mark the connected domain of the corrected image,and the connected domain containing the most curve parts was taken out.Finally,the tortuosity of the connected domain was calculated,and the one reaching the threshold was the thread end to be identified.The results of identifying thread end-containing images and thread end-free images show that the local perspective transformation correction method is nearly 30％faster than heavy mapping,showing high correction accuracy;the tortuosity threshold discrimination method can effectively filter out non-thread end images and improve the recognition accuracy of thread end images.Using machine vision to identify the ends of the cheese yarn can effectively detect the thread end part of each cheese yarn,reduce manual intervention on the production line,and provide reference for building an efficient and energy-saving intelligent production model.
machine visionidentification of the ends of the cheese yarncorrection modellocal perspective transformationconnected domains
NETL
NSTL
万方数据
维普
