Research And Realization Of An Algorithm Based On FPGA For Fabric Defect Detection

Fabric defect inspection is a critical step in production quality control of moderntextile. At present, fabric defect detection in textile industry is primarily performedby manual detection, Because of subjective factors, human vision transience andlimited identification ability, there are drawbacks, such as poor efficiency, low accuracy andhigher cost. With the development of machine vision technology, scholars pay highattention to a research of detection technology based on machine vision no matter athome or abroad. Considering detection efficiency and accuracy in a fabric defectinspection, an algorithm based on FPGA has been studied in this paper. The researchsubject has supported by “13115” Major Innovation Project of Shaanxi Province.After reading lots of literature at home and abroad, an algorithm based on FPGA forfabric defect detection was studied and designed in this paper, The main researchworks are presented as follows:(1) a median filter and histogram equalization forimproving image quality;(2) an illumination compensation algorithm foreliminating effection of uneven light;(3) an algorithm of improved BUPCNN(Bidirection Unit-Linking PCNN) for detecting fabric defection;(4) a realization ofthe algorithm based on FPGA. Experiment results show functionality and performanceof the improved algorithm reaches expected requirements for a online detection of fabricdefect, and the algorithm is feasible and valid.