Ceramic industry has been one of the most reliable industry sectors in Indonesia for
the last few years through its positive performance sales. But, unfortunately, in last
8 years the number of imported ceramic increasing sharply, while the export always
decrease even it is not significant in order to fullfil the national ceramic demand.
According to Asosiasi Aneka Keramik Indonesia (ASAKI), Huge number of import
mostly originated from China, Thailand, and Vietnam because it is very cheap and
have better quality than Indonesian domestic product. Quality need to be increase
because high level of quality produces high customer satisfaction, which usually
supports for high selling price and also cheaper production costs. Ceramics
inspection in Indonesia still done manually. Therefore, a visual inspection system
with digatal imagery can be an effective solution to the problem. Digital image
processing can be used to extract various features of image. The process runs
automatically to minimize human intervention and expected to replace the
inspection process which is still done manually. In this research, focuses on
designing automatic classification system for ceramics defects inspection based on
image processing using Naive Bayes Classifier. The system proved that can classify
five different classes such as non-defect ceramics, scratch, chip off, dry spot, and
crack. It obtained 65.60% of accuracy.
Keywords: Machine Learning, Image Processing, Naive-Bayes Classifier, Visual
Quality Inspection