An Advanced and Automated Neural Network Based Textile Defect Detector

Abstract

All textile industries aim to produce competitive fabrics. The competition enhancement depends mainly on productivity and quality of the fabrics produced by each industry. In the textile sector, there have been an enlarge amount of losses due to faulty fabrics. In the Least Development Countries (LDC) like Bangladesh, whose 25% revenue earning is achieved from textile export, most defects arising in the production process of a textile material are still detected by human inspection. The work of inspectors is very tedious and time consuming. They have to detect small details that can be located in a wide area that is moving through their visual field. The identification rate is about 70%. In addition, the effectiveness of visual inspection decreases quickly with fatigue. Thus, to produce less defective textile for minimizing production cost and time is a vital requirement. Digital image processing techniques have been increasingly applied to textured samples analysis over the last ten years (Rallo et al., 2003). Wastage reduction through accurate and early stage detection of defects in fabrics is also an important aspect of quality improvement. The article in (Meier, 2005) summarized the comparison between human visual inspection and automated inspection. Also, it has been stated in (Stojanovic et al., 2001) that price of textile fabric is reduced by 45% to 65% due to defects. Thus, to reduce error on identifying fabric defects requires more automotive and accurate inspection process. Considering this lacking, this research implements a Textile Defect Detector which uses computer vision methodology with the combination of multi-layer neural networks to identify four classifications of textile defects. Afterwards, a microcontroller based mechanical system is developed to complete the Textile Defect Detector as a real-time control agent that transforms the captured digital image into adjusted resultant output and operates the automated machine (i.e. combination of two leaser beams and production machine), illustrated in Fig. 1. The main purpose of this chapter is to present an advanced and automatic Textile Defect Detector as a first step for a future complete industrial Quality Information System (QIS) in textile industries of Least Development Countries (LDC). The chapter is organized as follows: