Fabric Pilling Computer Analysis And Evaluation,

In this paper, several methods based on image technologies, which make use of newly developed theories in information science, such as wavelet analysis, pattern recognition and etc, are proposed for automatic detection of pilling on surface of plain fabrics. Image processing technique has the advantages of quickness, accuracy and automation. At present, many manual-testing works can be replaced by image processing technique.How to evaluate the grade of textile pilling is an important thing of textile testing fields. The common way we use is to judge the grades according to the standard sample, but it is greatly subjective and will produce different results among different people. It is difficult for us to precisely and quantificationally describe the amount of pilling precisely and quantificationally. The subject will introduce the image processing technique into the textile field, evaluating the amount of pilling by image processing technique and has some impacts on the objective and quick evaluation of the amount of pilling.The content includes:1. On a basis of the current image processing system, the image of the standard sample cards of the carded woven fabrics were analyzed and the pillings were extracted, which based on multiresolution analysis, the weaves of fabrics are wiped off successfully.2. By the means of the wavelet alternation, the noise is wiped off successfully.3. By area filter, knots of yarn and fuzz are discarded from the target and the image only consists of pillings was left.4. To obtain the characteristic parameters from the pilling image and select the most sensitive parameter, the total area of pilling ball, the amount of pilling ball, the standard deviation of area in woven fabrics were selected . The standard characteristic vectors on evaluating the grade of pilling are set up separatively according to the standard samples.5. In standard condition, make friction on the sample of carded knitted fabrics and woven fabrics by YG501 pilling tester.6. According to the principal of matching standard sample, by the pattern recognition to evaluate the amount of pilling, the result is correspondent with the method of manual discriminant. The results of experimental are satisfied and the method using computer visionand digital image processing technology to recognize pilling is promising.