Faults Detection with Image Processing Methods in Textile Sector

The quality control process for determining fabric faults causes high cost and time loss when performed with human eyes. Adding a real-time system to the fabric production mechanisms is very difficult to detect mistakes. In this study, 5 MP Raspberry Pi camera module is used to detect the image of the fabric moving through the conveyor band and a Raspberry Pi 3 minicomputer to process the received image and display it on the touch screen. In addition, studies on the inspection of the faults in the woven fabrics, which are the most applied areas of image processing methods, have been investigated and thresholding method, HSV color space transformation and morphological processes have been used in order to detect the fault in any fabric accurately and correctly. The most suitable threshold values for these methods have been determined by testing in the images obtained as a result of the investigations.