ABSTRACT
An appearance defect online inspection system of
mesh fabric has been developed based on machine
vision. 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 results have shown when the online
inspection speed reaches 60m/min, the defect
recognition rate can reach 95% or more and the
online inspection system can meet the automation
requirements of enterprises quite well.