Analysis of woven fabric structure using image analysis and artificial intelligence

An integrated hardware and software system has been developed that can automate the analysis process of various woven fabric structures. Although the analysis of woven structure is one of the most important steps in the fabric design and quality control process, it has been dependent only on human skills with primitive devices. In this study, a dedicated hardware system has been developed to obtain an ultra-high resolution fabric images. Then a series of image analysis technique was applied to locate the intersecting regions of warps and wefts on those images. Finally, an artificial neural network was formed to determine the woven structure of fabric based on the two shape parameters extracted from those regions.