5201

4871

كارشناسي ارشد

مخابرات-سيستم

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1388

نه،125ص.:مصور،جدول،نمودار

ص.ع.به فارسي

سعيد صدري،رسول امير فتاحي

2/3/89

نيلوفر قيصري،بهزاد نظري

مهندسي برق و كامپيوتر

ID4871

به فارسي و انگليسي: قابل رويت در نسخه ديجيتالي

Automatic Inspection of Palate Surface for Detection of Grate Bar Defects through Digital Gray Level Image Processing Techniques Aida Fooladi Vanda a fooladivanda@ec iut ac ir Date of Submission 2010 04 26 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language FarsiSupervisors Saeed sadri sadri@cc iut ac irRasoul Amirfattahi fattahi@cc iut ac irAbstractAutomatic inspection of products and plants plays a significant role in industries Steel industry is one of thebig industries that in different stages of its production line needs inspection Mobarakeh steel company is thebiggest steel maker company in Iran Inspection of different units of this company e g pelletizing plant hasbeen done by human such that the quality is enhanced and the probabilistic costs are decreased In thepeletizing unit of Mobarakeh steel company 260 pallets exist Pellets are posed on the pallets with 150 cmwidth and 360 cm length The pellets undergo drying and preheating in the stove There are four rows ofgrate bars on the surface of the pallet Each row consists of 90 grate bars Gradually the grate bars can bedamaged because of the high temperature of stove sudden change of temperature and hitting the pellets tothe grate bars The damages cause spaces between grate bars which lead to losses in pelletizing unit If thedamages periodically are observed and reported then these losses can be prevented through replacing thepallets For this reason pallets need permanently to be inspected Traditionaly this inspection has been doneby human In this thesis a method for automatic detection of these damages is presented In the first step thepallet area is segmented from the image by an example based learning approach where a model of an objectclass is derived implicitly from a training set of samples In the second step the pallet image is enhanced andthe space between every two adjacent grate bars is segmented by binarization morphology and labelingapproaches In the next step 66 features geometric grey level histogram co occurance matrix and wavelettransform features are extracted from these regions In the final step the regions are classified based on theextracted features to the following three classes intact deformed and spaced grate bars Consequently thedamages can be detected and also the location of the damages can be indicated on the pallet surface with highaccuracy Keywords Automatic inspection image segmentation image enhancement feature extraction supportvector machine classifier

سعيد صدري،رسول امير فتاحي

نيلوفر قيصري،بهزاد نظري