Objective Evaluation Method On The Smoothness Of Garment Elbow

With the improvement of the economic level of consumption, the requirement of clothing is not just loincloths, warm, more beautiful, neat. In order to continuously meet the needs of the consumers, we must strengthen the quality of clothing appearance of predictiion. At present about clothing crease of evaluation standard, AATCC is a kind of subjective rating method, which internationally used. Although it is simple, easy to operate, but the outcome is uncertain, often vulnerable to factors such as environment, evaluation personnel psychology and so on. With the development of information technology, image processing technology has been widely used in the field of textile and garment. Using image processing t echnology can be more precisely and objectively evaluate the clothing surface roughness. The fabric wrinkle which is produced by the AATCC wrinkling is different from the wrinkle caused in the process of the costume, so the evaluation results of the fabric smoothness can’t objective describe the effect of garment appearance smoothness. We need to explore objective evaluation methods of costume flatness.Aiming at this phenomenon, this text will base on the commonly used fabric on the market as the research object, establish objective evaluation system of using fabric crease recovery angle to predict a costume crease. The system mainly include: the comprehensive crease recovery angle prediction, the prediction of costume objective characteristic value, clothing in grade prediction based on neural network. The research content and results are as follows:(1) By testing fabric crease recovery Angle, using variance analysis of F and Brown-Forsythe test found that the significance of most samples are less than 0.01, so it can be concluded that the crease recovery angle of different angles exists significant difference;Using grey correlation analysis method to extract the comprehensive index of the crease recovery angle, the standard deviation of which and different angles is less than the standard deviation of the average warp and weft, so the comprehensive index that extracted is more reasonable as the fabric wrinkle resistance evaluation index;Using Stepwise regression analysis method to establish a model of comprehensive crease recovery angle and crease recovery angle of different angles.(2) Making fabric into clothes, wrinkling experiment by wearing on humans, taking the wrinkling images and subjective evaluation.Through the study find: wrinkle grade evaluation of a weared costume, wrinkle degree of the right sleeve is higher than the wrinkle degree of the left sleeve; The wrinkle grade of a weared costume tiled significantly greater than the wrinkle grade in condition of clothing; There is a significant correlation between grade, the integral score and the crease recovery angle. That is to say the smaller the crease recovery angle, the smaller the level, the greater the integral score.(3) Through v=0.7,5*5 convolution filtering wrinkle image, obtaining detailed crease details, at the same time,keeping the overall trend of wrinkle texture. On this basis, crease characteristic parameters which extracted by gray difference statistics,Tamura texture statistics and correlation integral, comprehensive crease recovery angle has the high correlation.(4) Through the analysis of the influence parameters of Gabor filter such as frequency, direction.With the principle of biggest entropy to selecte appropriate Gabor filter: f=16,theta=0;f=16,theta=π/12;f=32,theta=π/6;f=8,theta=π/4;f=32,theta=π/3;f=16,theta=5π/12;f=16,theta=π/2. The characteristic parameter extracted by fusion after seven filter is consistent with degree of wrinkle wrinkle image.(5) Through 1-6 layer after wavelet transform to extract the characteristic parameters, the change of the curve can be found: 1-4 layer wavelet transformation, curve upward trend, in 4 layer reach maximum, 5-6 layer curve is on the decline. Indicates: 4 layer wavelet can get more detail information wrinkle. Crease characteristic parametersha、hsextracted by 4 layer wavelet is consistent with the sample wrinkle degree.(6) Comparing to RBF neural network with PSORBF neural network find that the PSORBF neural network obviously has higher precision than RBF neural network; With the increase of crease characteristic parameters, the neural network training error, testing error is reduced.Through a variety of image processing methods for the extraction of crease characteristic parameters, can more accurately predict costume wrinkle grade.This paper try to explore the actual dress garment wrinkle test and evaluation methods, based on image processing technology, to extract objective evaluation index of characterization of a costume crease, provide reference for the quality of garment appearance evaluation.Seting up a evaluation system of costume flatness, which provide reference for selection of fabrics, to predict the appearance quality of different fabrics made into clothes after wearing.