AI and IoT for Yarn Defect Detection in the Textile Industry: A Systematic Literature Review (TURNITIN CHECK)

The textile industry continues to face yarn defects, which, in turn, reduce product quality and make production
inefficient. Conventional manual inspection methods are unreliable and labour-intensive; therefore, combining AI and IoT
enables intelligent quality control. This work conducts a Systematic Literature Review (SLR) on the applications of AI and
IoT in yarn defect detection, using the PRISMA methodology to select and synthesise articles. A review of publication
records across Scopus, ScienceDirect, IEEE Xplore, and Google Scholar, and thus collected 25 peer-reviewed papers from
2014 to 2024. The research trends were classified by production stage, algorithm type, and global distribution. The results
show that AI methods, especially image processing, neural networks, and deep learning (including CNN-based models),
play a leading role in yarn defect detection, achieving accuracies exceeding 95%. However, the implementation of IoT for
real-time monitoring remains underdeveloped, and few studies have examined in-process defect detection. Post-production
inspection receives the vast majority of contributions, while pre-production and on-production stages receive less attention.
China leads in the number of published papers, followed by Turkey, Egypt, and India. The main challenges lie in combining
AI and IoT with legacy systems, ensuring the reliability of data supply, and handling computational and cost constraints.
This review concludes that when AI harmonises with IoT, it drives transformative shifts in predictive monitoring and smart
manufacturing of textiles. Further studies are expected to focus on real-time IoT-based monitoring, model optimisation, and
low-cost implementation towards fully automated, data-driven yarn defect detection systems.