The Fabric Structure Analysis And The Defect Detection Based On Time-Frequency Transforms

In fabric Industry, it is very important to identify the character of the fabric without any damage. The Image Processing is a good method to achieve this. We call this method Fabric Image Processing. In fabric Image Processing, Automatic identification of the fabric structure and the characteristic of the fabric surface are both very important.Automatic identification of the fabric structure plays a vital role in the fabric industry. In this paper we present a rapid and accurate approach for automatic identification of the woven fabric structure. The proposed method is based on the geometric feature of peaks in the power spectrum of Fourier transforms. We perform our method into three kinds of jeans fabric: plain, reclining twill and steep twill. Different fabrics exhibit different included angle of peaks in the power spectrum. In addition, the float of fabrics is calculated accurately, too. The efficiency of this approach is illustrated with 96% accuracy in structure detection.Besides, the power spectrum can also be used to defect detection, for every peak in the power spectrum are meaningful. Some peaks denote the main texture of the fabric. If these peaks are deleted, defects may be detected. In this paper, this method can be used to detect the Bamboo fiber.But Fourier transform has its limitation. It cannot be used as an adaptive method. So Adaptive Wavelets is used in this paper to detect the defect of fabrics. Firstly, it introduces the Adaptive Wavelets Analysis Algorithms, and decomposes the fabric images using Adaptive Wavelets coefficients. The experimentation result shows that it can adapt the main texture of the fabric and reject the defect in the fabric.