The Identification Of Cotton And Lyocell Fibers In The Longitudinal View Based Of Automatic Fibers Image-Acquisition And Identification System

At present,the artificial method is widely used to identify component in blended products of cotton and Lyocell fibers.And it depends on subjective judgments and experience,which is bound to affect the objectivity and accuracy of results.Furthermore,the method is time-consuming and energy-consuming with low efficiency.So the department of commodity inspection needs a high-speed,accurate and objective method to identify cotton and Lyocell fibers.On the base of the successful establishment of"the automatic fibers image acquisition and identification system",cotton and ramie fibers can be identified easily.This research focuses on the functional expansion of the system.Based on the factor analysis of six characteristic parameters used to identify cotton and ramie fibers,two indirect characteristic parameters of cotton and Lyocell fibers are extracted.And these two indirect parameters efficiently quantify differences between cotton and Lyocell fiber.For the new identification program based on two indirect characteristic parameters,the results of all corresponding experiments verify the recognition rate of cotton and Lyocell fibers can reach more than 95%.The research is composed of two parts.The first part is to analyze characteristics of the cotton and Lyocell fibers on the basis of six characteristic parameters used to identify cotton and ramie fibers.Compared with cotton and ramie fibers,cotton and Lyocell fibers have more obvious differences in longitudinal view,so six characteristic parameters of cotton and ramie fibers are used to identify cotton and Lyocell fibers.The corresponding recognition rate can only reach about 65%when 2,000 cotton and 2,000 Lyocell fibers are tested.Obviously,it can not meet the requirements in the practical application.In order to further improve the recognition rate,through the observation of the values of the six characteristic parameters,some of characteristic parameters of Lyocell fibers are zero,which means they will not make any contribution to the corresponding recognition result.That will lead to the system error,so the six characteristic parameters need be further analyzed.If the characteristic parameters are re-extracted to meet the requirements of identifying cotton and Lyocell fibers,the whole system will be rebuilt and the original stable and effective part will also be given up.So the indirect characteristic parameters with linear combination of the six original ones are expected to be obtained,which can meet good independence and separability.Some of characteristic parameters are ineffective,while all the indirect characteristic parameters are effective. The second part is the new identification program is put forward and verified.On the base of the factor analysis,the two public factors are obtained with the linear combination of six characteristic parameters.The further analysis shows they have good independence and obvious separability,so they are defined as the indirect characteristic parameters.A single indirect characteristic of cotton and Lyocell fibers has not very good independence.But in the 2-dimentional space which coordinates respectively present two indirect characteristic parameters,the displacement areas of cotton and Lyocell fibers have obvious separability. Wang Rongwu,the member of the research group,employs two indirect characteristic parameters to establish the new identification program.In the experiments of identifying 2919 cotton fibers and 2919 Lyocell fibers, the result verifies the recognition rate can reach more than 95%.And for the small sample size,the identification becomes more stable.If effective indirect characteristic parameters can be extracted from three different kinds of fibers and represent good independence and obvious separability,the three different kinds of fibers can be identified.