Abstract
The 'Automatic Fabric Evaluation System', that can analyze the structure of woven fabric automatically and evaluate the quality objectively using image analysis, has been developed. Fabric images were captured by CCD camera and preprocessed by Gaussian filtering and histogram equalization. The fabric constructing parameters, such as fabric count, cloth cover, yarn crimp, fabric thickness, and fabric weight per unit area, were measured automatically from the planar and cross-sectional images of woven fabric using image processing and image analysis. The results obtained with the system were compared with the experimental values and sowed good correspondence. As for the quality evaluation of woven fabric, in the case of defected fabric image, defects such as slub or missing pick were detected successfully and in the case of normal fabric image, the uniformity of yarn spacing and the orthogonality of yarn intersecting angle were proposed as criteria for quality evaluation. Therefore, the coefficient of variation of yarn spacing and yarn intersecting angle were measured quantitatively, such that the quality could be compared relatively through values.
