Computer vision based system for localizing structural defects in textile web material using texture analysis techniques

Abstract

Global trade in textiles has played a vital role in the growth of several countries. In India, next to agriculture, the textile industry has an extensive and renowned history in the Indian subcontinent. At global level, Indian textile industry is one of the biggest industries with an enormous amount of raw materials. India is a number one jute producer and a second chief producer of silk and cotton. Textile manufacturing produces fascinating goods every minute for fulfilling the needs of the consumers. India is able to produce technically much advanced textiles than Europe could and there is very wide scope for textile industry because of abundant availability of raw materials such as silk, cotton, jute, wool and man-made fibers. India has got a huge cultural inheritance with a wide range of textile stuffs, method, and types, thereby satisfying the needs of various sets of people and customs across the country. The textile industry includes an extensive assortment of section, starting from hand-woven to capital intensive technological segment. The foremost requirement for textile industry is a source of fiber which can be made into yarn by spinning process. Either by knitting or by weaving, the yarn is processed to produce fabric. for any fabric design, if there is a variation in yarn size or if anything goes wrong in the knitting or weaving process, then it will lead to damage in the fabric quality. As there are various classes of fabric defects, the examination of real fabric defects is essential. Fabric defect detection is of huge importance for managing quality in textile industries. In Indian textile industries, manual inspection is carried out currently for the examination of real fabric defects by subjecting the fabric rolls into the inspection frame. With manual inspection, a well-trained person would inspect all the types of fabrics, identify all the defects correctly and divide them into corresponding classes. In this inspection process, utmost level of attentiveness is maintained only up to 30 minutes only. Further, as the inspection work is very boring and monotonous in nature, there is a chance of human errors with high inspection time. In addition, the speed of Human Visual Inspection System (HVIS) is very slow compared to modern manufacturing industry production rate. Owing to all these drawbacks, Automated Visual Inspection System (AVIS) has been emerging more important than ever. In this scenario, a wide spread application of technology is required not only to upgrade the quality of products and to determine consumer choices but also to overcome locational disadvantages and to reduce overhead cost on unsold inventories. newline