基于计算机视觉的织物疵点自动检测.pdf
西安工程大学学报
ofXi’an
Journal PolytechnicUniversity
第23卷第2期(总96期) 2009年4月
Article
ID:1671-850X(2009)02-0196-06
Automatic offabricdefects
inspection
basedon vision
computer
LIU
Shu—guang,QUPing—ge
0fElectronicand
(School Information,Xi’anPolytechnicUniversity,Xi’锄710048。China)
transformandBPneural used and fab-
Abstract:W|avelet networkwere to the
togetherinspectclassify
ticdefects.Awhitefabricis asthe the defectsareoil
plain adoptedsample,anddistinguishing stains,
areacamerawith256×256resolutionisusedinthe
warp-lacking,andweft—lacking.An scheme,a
istransmittedtoa for da-
wavelet thenthe
grabbedimage computer transform,and image
corresponding
are
ta thenusedinBPneural resultshowsthatthefabric
network够input..11le
ratecanbe to95%withabovemethod.
up
words:wavelet neural defects
Key analysis;BPnetwork;fabric
CLC 101.97A
number:"IS
0 Introduction
automatic to and havebecome
Applying techniquesproductionprocesses,inspectionmanagementprocedures
moreandmore fortextile still onhuman
inspection vision,and
important