Entropy Based Defect Detection for Patterned Fabrics

This research work focuses on identifying defects in patterned textiles, which are essential for improving fabric quality and reducing financial losses. The proposed hybrid denoising algorithm, ksvd dwt, effectively removes various types of noise from fabric photographs. To enhance fault detection in patterned fabrics, a feature extraction and selection strategy are employed, using robust feature descriptors like SIFT and SURF. Particle Swarm Optimization (PSO) is applied for feature selection, which, combined with the entropy-based feature selection method, significantly improves the accuracy of defect identification in patterned fabrics. The study emphasizes the importance of two-step processes involving denoising and feature extraction to distinguish between defective and defect-free fabric images.