An appearance defect online inspection system of mesh fabric based on machine vision has been developed. The mean filter method is adopted to eliminate noise. An adaptive threshold method based on brightness is presented to eliminate the effects of uneven illumination and separate the foreground and background. By analyzing the texture characteristics and defect features of mesh fabric, the mesh fabric defect identification method based on texture features is proposed. The inspection results show that the speed of the online inspection system could reach 60m/min and the defect recognition rate could reach 95% or more, which should meet the automation requirements of enterprises quite well.
展开▼