PolyU IR
 

PolyU Institutional Repository >
Textiles and Clothing >
ITC Theses >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/3953

Title: Characterization of fabric appearance based on image analysis
Authors: Xin, Binjie
Subjects: Hong Kong Polytechnic University -- Dissertations
Textile fabrics
Textile chemistry
Issue Date: 2009
Publisher: The Hong Kong Polytechnic University
Abstract: In order to improve the reliability and reproducibility of quality evaluation and control system, it is necessary to develop some objective methods to evaluate fabric appearance, such as pilling, wrinkling, fuzzy, and so on, instead of subjective methods. In this thesis, we have done the following investigations: 1) the development of effective 3D imaging system for the characterization of fabric surface properties by using the DFPS technology (Dynamic flexible profile scanning technology) and stereo vision technology; 2) the study of illumination effects on pilling evaluation and neural networking modelling of fabric pilling evaluation; 3) the development of objective evaluation system for the appearance evaluation of knitted fabrics before and after washing for the purpose of optimisation of washing process; 4) the development of objective evaluation method for the evaluation of seam puckering appearance by using X-illumination, morphological fractal and Bayes Classifier.
Through our experimental results and modeling analysis, it was found that: 1) the new 3D imaging system based on progressive silhouette imaging techniques could be used for the visualization and characterization of textile surface profile and roughness, the system R&R is acceptable for the measurement of surface thickness and roughness, this method is insensitive to the color and texture effects of textile materials; 2) The relationship between illumination angle and the brightness profile of pills has been modeled theoretically, decreasing the illumination angle could enhance the contrast of pill image while decreasing the brightness of pills; 3) Log-sigmoid function can simulate the judging behaviour of subjective evaluation in comparison with other functions; 4) 44 features were extracted by using 5 image analysis methods to characterize the knit fabric appearance for the optimization of washing process, Principle Component Analysis (PCA) was used to determine the principle factors to describe the fabric appearance, washing temperature and duration have obvious effects upon the fabric pilling appearance, while softening temperature and duration have significant effects to texture of fabrics; 5) the presented image based method based on X-illumination, multi-morphological fractal and Bayes Classifier is feasible for the evaluation of seam puckering appearance, which is insensitive to color and texture attributes; 6) mirror-based stereo 3D imaging system has been developed for the 3D visualization of cloth/garment. The ultimate aim of this research is to develop some effective imaging systems and objective evaluation methods for the quality evaluation of fabric appearance, so as to liberate human being from the tedious, repeatable, time-consuming subjective judging works. It will contribute to the textile industry by establishing a modern objective evaluation system and building up a quick response circle-link between manufacturer and consumers all over the world.
Degree: Ph.D., Institute of Textiles and Clothing, The Hong Kong Polytechnic University, 2009
Description: 245 leaves : ill. ; 30 cm.
PolyU Library Call No.: [THS] LG51 .H577P ITC 2009 Xin
Rights: All rights reserved.
Type: Thesis
URI: http://hdl.handle.net/10397/3953
Appears in Collections:ITC Theses
PolyU Electronic Theses

Files in This Item:

File Description SizeFormat
b23071874_ir.pdfFor All Users (Non-printable) 6.33 MBAdobe PDFView/Open
b23071874_link.htmFor PolyU Users 162 BHTMLView/Open



Facebook Facebook del.icio.us del.icio.us LinkedIn LinkedIn


All items in the PolyU Institutional Repository are protected by copyright, with all rights reserved, unless otherwise indicated.
No item in the PolyU IR may be reproduced for commercial or resale purposes.

 

© Pao Yue-kong Library, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong
Powered by DSpace (Version 1.5.2)  © MIT and HP
Feedback | Privacy Policy Statement | Copyright & Restrictions - Feedback