Development of an automatic yarn bobbin abrage detection system by machine vision technology

In purpose of this study, a machine vision system which can automatically detect existence and area of abrage fault on yarn bobbin were developed. The fibers have different dyeability properties and different light reflection characteristics. As a result of accidental mixing of raw materials used at different stages of yarn production (in the blend or package phase), the fault known as "abrage" occurs. If the defective coils are converted to the fabric form, then there will be mistakes in color differences after dyeing process. The abrage faults are seen as color or shade difference on the dyed fabric. Nowadays abrage faults are controlled by the quality control personnel on the bobbins or on the fabric by eye. This process takes long time and it is very tiring. New developments in machine vision and automation technologies provide more sensitive process control and quality inspection in each stage of the production line. Developments in industry 4.0 technology and image processing techniques have been used in many areas in textile industry in last decade. Image processing techniques have been generally used for automatic detection of fiber, yarn and fabric characteristics with improved accuracy and quicker results. In this thesis study, a prototype machine vision system was designed and constructed for automatic detection of yarn bobbin abrage fault. Image processing software were developed and applied on bobbin samples including different types of abrage fault. Finally, the success of the developed machine vision system was statistically evaluated. It was determined that the bobbin abrage faults were detected with 95.84% accuracy.