Fabric quality analysis [7,8,9] is important in textile engineering. Inspecting the non-woven basis weight on-line is a significant target toward improving the quality of the nonwoven material [3]. The discrepancy among various manufacturing processes affects the distribution of fiber, and also causes changes to non-woven basis weight. The short-distance deviation represents the quality of fiber carding, the long-distance deviation reflects the non-uniform quantity of product material, and the periodic deviation indicates flaws in manufacturing processes. Therefore, inspecting the non-woven basis weight on-line to monitor the quality during the production process, to avoid production flaws and to reduce the production costs are the major issues of the textile industry. 1The methods used for inspecting non-woven, web basis weight or uniformity can be either contact or non-contact methods. Usually the contact methods cause damage to the material. Owing to the direct contact of non-woven fabrics in the measurement process, the method of using floating pressurized rollers [17] to measure the non-woven thickness for converting into non-woven weight can easily damage or change the physical nature of the original nonAbstract This article explores a method of inspecting the non-woven basis weight on-line, by combining the exponential law of absorption and image processing technique. In the experiment, a theoretic equation for the charge-coupled device image-capture transmittance and non-woven basis weight was acquired through the exponential law of absorption. By using light source compensation and the least-squares approximation, the best approximate equations of transmittance and basis weight were acquired. According to the experimental results, the method of inspecting non-woven basis weight by image processing as proposed in this paper can reduce the uniformity of image gray scale from a coefficient of variation (C.V.) of 5.5 to 0.7%, raise the R 2 of basis weight and transmittance to 0.998, and decrease the basis weight inspection error to 0.33%. The proposed method effectively enhances the accuracy and stability of non-contact inspection of non-woven basis weight. The result acquired herein can also be applied to the on-line uniformity inspection system with web as the basic material.

By using the effective distance between clusters (EDC) as the basis for feature selection, this paper achieves a significant and effective feature for textile yarn grading, and further upgrades the operational efficiency of such grading. The results, such as feature selection processing to principal axis vectors (PAVS) by EDC, show that the feature's average number and average total distance of mistaken ranking by EDC are only 33.3% and 16.7% of those by Karhunen-Loeve (K-L) expansion, respectively. Furthermore, EDC can be applied directly to the feature selection of property vectors (PVS) and can reduce the measured items of pvs without lowering identification precision. Compared with our previous method of textile yarn grading, EDC provides 16.7% greater efficiency both in measuring pvs and calculating PAV1 time.

No abstract

No abstract

The grade of textile yarns is an important index in evaluating the yarn's market value. This paper uses the backpropagation neural network (BNN) and Karhunen-Loeve (K-L) expansion method to construct a new and highly accurate grading system. Outcomes show that a highly accurate and neutral grading system can be obtained if the BNN learning sample is comprehensive or by adopting the BNN with a relearning technique (self-healing). Considering the possibility of reducing the dimension of BNN input vectors without losing the accuracy, this paper preprocesses the BNN grading system using the K-L expansion. Experiments demonstrate that the K-L expansion provides a way to reduce the input dimensions, and that a single principle axis value of the BNN with the K-L expansion grading system is able to grade textile yarns. In addition, the experiment demonstrates that as the input dimensions are reduced to four in a self-healing neural network with the K-L expansion, the grading system provides the high accuracy and robustness.