Abstract: Fabric defect detection is one of the major stages in the cloth manufacturing sector. Defects such as broken yarns, stains, and holes reduce the quality and value of the produced fabric. Traditionally, these defects are inspected manually, which is time-consuming and leads to operator fatigue over long periods of work. This work presents a machine learning–based fabric defect detection system implemented in LabVIEW for identifying surface defects in commonly used textile materials such as cotton and polyester with plain surfaces. The proposed approach uses extracted texture and statistical features from fabric images, which are classified using a trained machine learning model to distinguish between normal and defective regions. Compared to manual inspection methods, the proposed system provides improved accuracy, consistency, and inspection speed. While machine learning-based approaches require labelled training data and an initial training phase, they offer better robustness to variations in fabric texture and illumination. The automated inspection process reduces human effort, minimizes subjectivity, and improves consistency, making it a practical, cost-effective, and scalable solution for real-time industrial textile quality assessment.

Keywords: Fabric defect detection, support vector machine, texture analysis, industrial automation, quality control

Downloads: | DOI: 10.17148/IJIREEICE.2026.14210