Computer vision techniques applied to the quality control of ceramic plates

Abstract

This paper presents a system, based on computer vision techniques, that detects and quantifies different types of defects in ceramic plates. It was developed in collaboration with the industrial ceramic sector and consequently it was focused on the defects that are considered more quality depreciating by the Portuguese industry. They are of three main types: cracks; granules and relief surface. For each type the development was specific as far as image processing techniques and illumination are concerned. The system was tested in pre industrial conditions showing the efficiency of the several developed algorithms and also revealing the perspective of its evolution to an industrial automatic inspection system.