Unveiling Textile Defects a Novel Fabric Defect Identification Based on Integrated Image Processing with Logical Learning Scheme

When it comes to controlling the quality of the cloth that is produced, defect identification is a crucial step. Considering the inefficiency and length of time wasted by human fabric defect detection methods, it would be more efficient to use computer vision technology to automatically detect fabric problems. This would allow manufacturers to better meet their demands. Recent years have seen a surge in interest in the topic of fabric defect identification as a result of the increasing cost of labour and the rapid development of automation within the textile industry. For the purpose of automated fabric flaw identification, we presented a learning-based methodology. We start by cropping the original image to a specific size and shape using a square slider. All of the cropped images are subsequently enhanced using improved histogram equalization. This work primarily aims to create a system called Logical Learning Scheme for Fabric Defect Detection (LLSFDD) that uses image processing techniques for simple and efficient defect detection and classification. To test how well the model works, it is cross-validated with a Convolutional Neural Network (CNN). Weft and warp threads that are missing, as well as oil stains and holes, are the most common types of flaws found. The four-step process is the basis of the system. It starts with taking the picture, then uses image analysis to remove parasitic details and make the picture sharper. Afterwards, find the rates of straight lines, dark regions, and voids; these three metrics will describe the flaws that were previously indicated. Lastly, utilise a suggested model, the Logical Learning Scheme for Fabric problem Detection, to identify the kind of fabric problem.