Research On Sorting Tube Machine Key Technologies Of Color Recognition

With the reduction of labor force,the continuous improvement of employing cost and the continuous development of science and technology,spinning enterprises have put forward higher requirements for the automation and intellectualization of equipment.In the spinning process,the yarn tubes of various colors from the winder are mixed together in disorder.The yarn tubes need to be sorted according to the color category before they are used again in the spinning workshop.Relying on manual sorting of yarn tubes in the spinning workshop,the working environment is bad and the sorting method is inefficient,which can no longer meet the needs of re-use of the workshop.This requires more stable and accurate sorting equipment to replace manual sorting.As an auxiliary machine of spinning process in textile industry,the main function of management machine is to automatically sort yarn tubes,and accurate color feature information is the basis of sorting yarn tubes.Because of the lack of advanced science and technology in the past,the early management machine only aimed at the size end of the yarn.Although color recognition was added to the subsequent research and development of the management machine,there were fewer kinds of color recognition,and there were many kinds of yarn color in the workshop,which could not meet the production needs.This subject suits the background of market demand.In order to sort the yarn tubes better,image processing technology is applied to automatic color recognition of the yarn tubes of the management machine.In this paper,a new yarn image acquisition device is designed,and the acquired yarn image is smoothed and sharpened.According to the characteristics of color space,color feature and color recognition method,a color recognition method based on region of interest is proposed.The experimental results show that the color recognition accuracy of this method is high,and the color recognition of management machine is close.Key technology research has certain reference significance for spinning enterprises.The main research contents of this paper include the following aspects:(1)The main research contents of this subject are as follows: image acquisition,physical characteristics and image acquisition structure of yarn tube are studied,image acquisition scheme of yarn tube of management machine is designed,influence of commonlight source,illumination scheme and camera on image acquisition is analyzed,design of illumination scheme and hardware selection of system are completed,image processing functions of raspberry pie and Python-OpenCV are studied.The compiling environment is built,and the editing program calls the camera to collect the yarn image,which realizes the yarn image acquisition and provides the image data for the follow-up research.(2)In image processing,firstly,image quality evaluation method is studied,and no reference evaluation method is selected to evaluate the quality of yarn image.According to the experimental results,the design of image acquisition scheme is proved to be of research significance.Secondly,the principle of image noise is analyzed,and the yarn image is filtered and sharpened to improve the quality of yarn image.(3)The color features of yarn images are studied,and different color space characteristics and color feature representation methods are analyzed.The H and S double channel color histograms of yarn images in HSV space are selected as the color threshold setting criteria.Aiming at the color recognition method of yarn,a color recognition method of region of interest yarn is proposed by using the idea of weight partition.Canny detection algorithm is used to extract the region of interest yarn,and the color recognition of yarn is carried out according to the designed overall flow chart of color recognition.Finally,through the analysis of the experimental results,we can observe the situation of recognizing different colors of yarns.The recognition accuracy is high and the system is stable.The rationality of the yarn image acquisition scheme,yarn image processing,yarn color features and color recognition method based on region of interest are verified.It can provide reference for the automatic sorting of yarns in spinning enterprises.