Length Measurement And Software Development Of Natural Cellulose Fiber Based On Image Processing Technology

Because of its excellent hygroscopicity and permeability,the natural cellulose fiber has become a very important raw material in the textile industry.The consumption and market share of cotton fiber is the largest.And the akund fiber is a new potential natural fiber found in recent years.Fiber length directly affects the quality of natural cellulose fibers and follow-up of textile production,processing technology and yarn quality and trade links.The existing fiber length measuring instruments have their own defects,such as the sample preparation of Y111 is time-consuming and laborious,and the test results vary from person to person.AFIS and HVI are expensive and their test results are not allowed.To summarise,the main weakness of fiber length measuring instruments is the test results can not be traced back.For the current situation of fiber length testing method and previous studies,this subject studies and expands deeply on fiber length testing methods on the basis of image processing technology.In this paper,the overall experimental scheme for testing the fiber length by image method is established,including fiber sample preparation,image acquisition,fiber image processing,length data processing and experimental comparison.The test method is also tested and applied.Finally,the fiber length testing software based on image processing technology was designed and developed.The main contents and conclusions of this paper are as follows:(1)According to the characteristics of fiber,the methods of fiber dispersion preparation are explored.For cotton fiber,the combed roving was selected for sample preparation on the black velvet board.The cohesion of the roving is weak and it has good straightness,and the fiber length was long and neat.For akund fiber,in view of its low fineness and large electrostatic effect,the polyimide plate is selected to sample the fiber by frictionaladsorption.The amount of single sample preparation of cotton and akund fibers can up to 20-30 roots.(2)After the preprocessing and fiber length calculation of the image,the fiber problems with a few large breakpoints and some cross patterns in the fiber image are analyzed and solved.For the case of large breakpoint,the relevant complementary algorithm and code are compiled.If the distance between the target endpoints is less than 150 pixels,it is judged as executing the complementary program,and the validity of the algorithm is verified experimentally.For the cross-shaped fibers,the basic situation of the intersection of two points and the intersections of two sections are mainly analyzed.The algorithm of cross-separation and the code are written.The key of the algorithm lies in the recognition of cross points and the matching of fiber sections.The actual fiber separation and algorithm separation were compared to verify the accuracy of the cross fiber separation algorithm.(3)For cotton fiber,the average length of the root length obtained from the length data is about 3mm longer than the AFIS test result,and the average length of the upper quartile is relatively close to the AFIS test result.The coefficient of variation of root length test results are 7.37%,and the experimental results of the number of short fiber velvet rate slightly smaller than AFIS.This is mainly because the combed roving fiber has good self-uniformity,less staple fiber content,and the image method has basically no secondary damage to the sample,and the test result is close to the length of the sample itself.(4)Software design and development of testing method of natural cellulose fiber based on image processing technology were carried out by using MATLAB software.Software features include fiber length test and length data processing,the interface is simple and easy to understand.In the software,the collected fiber images are loaded,the results of the image processing can be directly obtained and the fiber length data obtained from the test,the number of fiber counts.When the number of roots reach the requirements,the length distribution data and various length indexes can be obtained.(5)According to the fiber length data measured by the image method,we can see that the length distribution of the akund fiber is similar to that of the cotton fiber,showing the continuous distribution from long to short,with the general characteristics of the natural fiber length distribution.The longest akund fiber is about 50 mm,the shortest fiber is about 15 mm,and the length is concentrated in the range of 25.5 ~ 38.5mm,which is ingood agreement with the existing research.The fiber below 16 mm is less,partly because the image test method overcomes the problem of fiber loss and loss,the test length is closer to the natural straightening state of the fiber,and the measurement process is accurate and traceable.The average length of the fibers is 31.73 mm.Compared with the results of the test by Fischer et al.and the hand-pulled ruler method,the accuracy and reliability of the measurement method can be basically verified.