Locating defects on shirt collars using image processing

Recent developments in the hardware and software mean that the automation of visual fabric inspection tasks is becoming feasible at low cost. This paper investigates the techniques that can be used to solve the problem of repetitive, tedious and physically demanding human inspection for defects in shirt collars. The faults studied in this work are those found in nine types of defects that can be present on shirt collar panels. Two statistical methods: moving group average, and moving divided group average are proposed. In addition, highlighting and variance techniques are applied to an image with moving group average and signature counting. These techniques gave an indication of fast computation time to detect the defects on the image, which is needed in manufacturing, and could be applied to most automated inspection systems.