The Evaluation Of Yarn Appearance Quality Based On Machine Vision

Under the background of new normal,Chinese textile must improve yarn quality if we want to improve the competitiveness of the international market.On the basis of promoting the intellectualization of production equipment,we also need to strengthen the detection and evaluation of yarn appearance quality.At present,the common test facility of textile industry at home and abroad are photoelectric method and capacitance method,and there are shortcomings in detection accuracy,detection efficiency and price.With the development of science and technology,the inspection of yarn appearance quality based on machine vision takes full advantage of image processing technology to overcome the shortcomings of photoelectric method,so as to achieve objective evaluation of yarn purposes.However,the existing yarn appearance quality testing system based on machine vision has not graded the yarn appearance quality because of the lack of grading evaluation standards.The purpose of this research is mainly focused on the yarn detection algorithm based on machine vision,hairiness detection algorithm,and the evaluation standard of yarn appearance quality,and completed the quality evaluation.The yarn detection algorithm is mainly divided into four parts:image preprocessing,yarn sliver segmentation and evenness evaluation,yarn hairiness extraction and statistics,algorithm effectiveness contrast analysis.The main contents of this topic are as follows:Build the yarn image acquisition platform based on machine vision.A fast FCM segmentation algorithm based on spatial neighborhood information constraints is proposed in view of the poor anti noise performance of the existing yarn segmentation algorithm.Extract the yarn lines by subpixel subdivision then detect yarn defects by measuring yarn diameter accurately.The hairiness length was calculated by the real length tracking method and the coefficient of variation coefficient of hairiness was calculated.6 specifications(41.65 tex,24.30 tex,19.43 tex,18.22 tex,14.58 tex,11.68tex)cotton ring spun yarns were tested,then the results are compared with the results of USTER CLASSIMAT 5,the contrast experiment shows that the results of the two methods are highly correlated,verified that the yarn appearance quality evaluation system based on machine vision is effective.According to the 2013 edition USTER bulletin and GBT 9996.1-2008 "cotton and chemical fiber pure spinning,blended yarn appearance quality blackboard inspection method" standard,the standard of yarn appearance quality evaluation based on machine vision is established,and the appearance quality evaluation of yarn is completed based on this standard.In order to make it more convenient for users to use this system,an easy to use graphical user interface is designed.Finally,summarize the current research work and the future development trend of the project are prospected.