The invention relates to a gray fabric broken 
yarn online detection method based on 
computer vision, which is characterized by comprising the following steps: Step 01, providing an industrial camera to 
shoot to-be-detected gray fabric so as to obtain a gray fabric picture; Step 02, providing a control computer to read the gray fabric picture and performing pretreatment on the picture; Step 03, utilizing broken 
yarn recognition program to compare the gray fabric picture with a preset fabric template by the control computer, and when the matching ratio can not meet the preset matching threshold value, sending a halt command to the communication port by the control computer; Step 04: providing a control module to receive the halt command so as to control a 
warp knitting machine to run a motor to 
cut off power supply and stop, so that the 
system is suspended and waits for settlement. According to the gray fabric broken 
yarn online detection method based on 
computer vision, disclosed by the invention, the defects of the traditional broken yarn detection method are eliminated, the detection accuracy, efficiency and capacity of resisting disturbance are improved by 
combing the method of one-dimensional and two-dimensional 
signal of the picture, the variety of to-be-detected gray fabric is increased, and better universality is realized.