Research On The Measurement And Visual Evaluation Of Yarn Evenness Based On Image Technology

The performance of the yarn directly determines the quality of the fabric,and the yarn evenness is an important indicator for evaluating the quality of the yarn.It is a necessary condition for controlling and improving the quality of the yarn by testing and analyzing the yarn evenness.However,the existing yarn evenness testing instrument mainly adopts the measuring principle of the combination of capacitance method and photoelectric method.The detection results are easily affected by environmental temperature,humidity and yarn surface hairiness.Moreover,the detection accuracy is low,the cutting length is not flexible,and it is cannot reflect the diameter change in the fine yarn segment.At the same time,the measurement results of existing instruments are mostly statistical indicators,which can not directly reflect the unevenness of yarn and fabric appearance.In allusion to the problems mentioned above,this paper proposes a method for measuring and evaluating the yarn unevenness based on image detection technology,and designs an image dectection system for yarn evenness(IDS-YE).It can measure and obtain pixel level yarn diameter and the general measurement results of yarn unevenness.And the new indexes that can better characterize the unevenness of the yarn in the fine segment are proposed.According to the measurement indexes and the yarn core image obtained by using the image technology,a method for visually evaluating the the yarn unevenness by constructing the electronic blackboard and the simulation fabric is proposed.The main research contents and conclusions of this paper are as follows:(1)An image detection system for yarn evenness,IDS-YE,is established to evaluate the yarn evenness in high real-time.The system integrates three modules:yarn drive control unit module,dynamic acquisition module for high-resolution image and the output module of real-time detection.By optimizing the design of yarn transmission path,allocating the position height of light source,yarn and camera,and coordinating the coordination between software and hardware,the system can meet the requirements of stable acquisition of image at a high speed.This can lay the prerequisite for obtaining accurate and stable measurement results.In order to ensure the real-time performance of the algorithm,a multi-thread system framework based on hard disk buffer technology is used to realize real-time acquisition and processing of yarn images.By fitting the position of the yarn centerline in the yarn images,it is verified that the yarn in the image captured by the system maintains a good verticality,which proves the reliability of the system image acquisition scheme.In order to make the collected yarn image continuous,according to the fluctuation of the overlap position in the sequence image,a processing scheme of image stitching is proposed.The repeatability and discrimination of the system are proved by repeated measurements on several sets of yarn samples.(2)A dynamic interval threshold acquisition method and a hairiness and isolated area removal algorithm are proposed based on the analysis of yarn core and hairiness features.With the steps of threshold processing and maximum area extraction,the effective segmentation of the yarn core image is completed so as to ensure yarn diameter can be measured with high accuracy.In order to guarantee the continuity of the obtained yarn image data,an image stitching algorithm based on the characteristics of the yarn hairiness information is proposed.The overlap position between the sequence images can be calculated,and the overlapping region in the sequence image can be removed according to the position.Compared with the results of manual processing methods,the above methods are highly consistent with the manual methods,which proves the effectiveness of the proposed method and lays a foundation for accurate acquisition of yarn diameter.(3)The yarn diameter data based on time series is extracted from the obtained yarn core image,and an indicator system for measuring the yarn evenness by the image method is established.Combining the definition of the general measurement indexes of the yarn evenness,the calculation method of coefficient of variation of yarn appearance diameter under different fragments,the calculation method of yarn diameter spectrum based on one-dimensional Fourier transform,the calculation method of DR and IDR,the calculation method of variation-length curve,the counting method for the number of frequent defects are carried out to re-express and theory elucidation in this paper.In order to evaluate the yarn unevenness from the fine structure,the micro-measurement indexes of the yarn unevenness are proposed based on the yarn core image and the micron-level detection accuracy.These include CV_P index that represents the variation coefficient of apparent diameter under the fine yarn fragments,and T_C index represents a percentage of the length of the yarn thin.(4)According to the established indicator system,IDS-YE system is used to test the yarns with different yarn counts,and the test results are compared with the existing instrument.Experiments show that for general measurement indicators,the measurement results of the two instruments maintain a high degree of consistency in the part of average diameter value,CV%under different cut length,diameter deviation frequency histogram,spectrum,DR and IDR curve,variation-length curve,the unevenness between segments and unevenness curve within segments,number of frequent yarn defects.The measurement results are not only line with the actual situation in the yarn production,but also the accuracy of the data can be verified in the above parameters.Therefore,the IDS-YE system can be used to evaluate the uniformity of yarn.In the analysis of the proposed micro-measurement indicators,the variation coefficient of the diameter under micro-segment and the number of thin content index are consistent with the trend of the CV value and number of yarn defects measured by the existing instrument.Thus,the proposed index can better reflect the irregularity of yarn fine structure.It is also found that the results of the two indicators show a stronger ability to characterize the unevenness of the yarn while meeting the actual situation.The validity of the proposed micro-indicators of unevenness is verified,and the comprehensive characterization of the yarn unevenness with low precision and high precision is realized.(5)In order to better evaluate the real appearance of the yarn,a method for visually evaluating the yarn evenness by using electronic blackboard and simulation fabric is proposed to make up for the defects of existing testing instruments.According to the proposed image stitching algorithm,the boundary position of yarn core images and the center line position information of the first line of the images are extracted.A mathematical model of electronic blackboard is established to display the yarn evenness based on the requirements of the actual yarn blackboard.Through the establishment of fabric structure change model,cross-section adjustment model,length direction adjustment model,yarn bending model in the fabric and yarn appearance lighting model in the fabric,the construction of simulation fabric is realized by using complete yarn core image and real image.The electronic blackboards and fabrics are respectively constructed by using the yarn data with differnet yarn counts,thick,thin,and nep,respectively.The feasibility of using electronic blackboard and simulation fabric to visually evaluate yarn evenness is verified from the measurement indexes and display effect.Finally,the main research conclusions of this paper are summarized.The shortcomings of the paper are given,and the development of IDS system is prospected.