ABSTRACT
Main topic of the article is application methods for defect detection on transparent materials. Paper describes entire evaluation process from image scan procedure right to object classification. In article are describe techniques for image scanning, techniques for elimination of different optic attributes of transparent materials, furthermore techniques for image processing and finally classification techniques. In part describing image attribute classification procedure is shown standard rule classifier and besides that also classifier based on neural network. In this paper individual techniques are introduced at the same sequence as a real process is executed on the each inspected object.
