Abstract:
In the textile sector, fabric quality plays a pivotal role in maintaining competitiveness, as defects in fabrics cause detrimental effects on the market. Traditionally, fabric inspection has relied on human intervention for a long time. However, this study aims to address this issue by developing algorithms that can accurately detect defects in single-color knitted fabrics. To effectively identify and analyze defective fabric images, this research has employed multiple methodologies, including Neural Networks, Image Processing, and Morphological operations. These techniques enable the detection and analysis of three common defects (stains, holes, and thread missing) in fabrics. By automating the defect detection process, this system can potentially offer significant benefits to the apparel industry, such as cost and time savings, as well as enhancing the overall efficiency of the quality inspection process.
