Defect Detection and Classification in Ceramic Plates Using Machine Vision and Naïve Bayes Classifier for Computer Aided Manufacturing

  Harpreet Singh,  Kulwinderpal Singh

Detection of defect is an inevitable task in ceramic manufacturing. In spite of the importance of the pictorial inspection of products, many of the inspection processes are performed manually. Here, the development of automatic processes for the examination of ceramic products is important. The difficulty is that manual inspection has some obvious drawbacks, such as being long, the high price involved, and the absence of standardization. Packaging glass or china clay product uses the packaging containers which of exact size that of the product to fit in. In this work, we propose a smart system for the automatic detection of the types of defects in ceramic products: the previous one is detection of a defect in ceramic plates for domestic use, called ceramic deformation, and the latter one is classification of a defect called deformation in plates for domestic use. To evaluate these applications, we used the dataset made available by the previous researchers for defect detection purpose only. The result in the form of graphs and GUI are provided showing the effectiveness of the method. In the result it is found that the overall accuracy for defect detection in ANN is nearly 73% while this accuracy for Naïve Bayes is nearly 85%. Also time taken for total defects analysis by Naïve Bayes is almost 0.02s while for ANN it as almost 0.25s.

Keywords- Machine Vision, Inspection, Defects, Classification, Gray Level, Image Series.

Unique Identification Number - IJEDR1703172
Page Number(s) - 1195-1200
Pubished in - Volume 5 | Issue 3 | September 2017
DOI (Digital Object Identifier) -   
Publisher - IJEDR (ISSN - 2321-9939)

  Harpreet Singh,  Kulwinderpal Singh,   "Defect Detection and Classification in Ceramic Plates Using Machine Vision and Naïve Bayes Classifier for Computer Aided Manufacturing", International Journal of Engineering Development and Research (IJEDR), ISSN:2321-9939, Volume.5, Issue 3, pp.1195-1200, September 2017, Available at :http://www.ijedr.org/papers/IJEDR1703172.pdf