Detection of fabric flaws based on machine vision

In order to realize the real-time detection of pattern fabric defects in the process of weaving, a machine vision-based rapid real-time detection method using area subtraction and automatic segmentation was proposed.This method first calculates the minimums of the horizonal and vertical superposition distance matching functions (DFMs) and analyzes the weight of every minimum, obtains the element period of pattern fabric accurately; then it determines the size of the area to be subtracted, calculates the gradient of the subtracted image, segments the gradient image with a marker-based watershed algorithm, it can fastly and accurately segment out the flaws area. Experiments shows that, the method can accurately test the flaws area, and has a high instantaneity that an average time of detection is about 200 ms and a low fallout ratio that the accuracy rate is above 98%, which fully meet the requirements of the industrial field.