Wavelet transform and principal component analysis in fabric defect detection and classification

Fabric defects are determined by quality control staff in textile industry. this process cannot be performed objectively and it constitutes both time and cost difficulties. in this study the cashmere and denim fabric images which are used often in textile industry are tried in both detection and classification process. quality control machine prototype has been manufactured then defected fabric images were obtained with the help of thermal imaging. the fabric defects were detected and classified by using the thermal images. averagely 95% classification accuracy has been achieved on experiments for two different fabric types. according to the experimental results, the fabric quality control process can be made after the drying and fixing, without any further quality control step.