Machine-vision based method for edge and surface crack detection in ceramic tile

Inhalt:
Ceramic tile is an important building and decorative material in the construction industry. With the rapid development of the real estate industry in China, people's demand for ceramic tile is increasing. The need for automatic detection of ceramic production from pulverizing to edge grinding has shown important significance to promote the development of ceramic tile industry. This paper proposes an on-line ceramic tile quality detection method with the combination of border boundary and surface feature information. The overall structure with machine vision detection is designed according to the specific quality index of ceramic tile. Aiming at detecting the dimensional specifications on the edge and cracks on the flat surface, the imaging-based sensing theory is studied. Problems of formulating the multi-thresholds and combined corner features cross correlation matching method is resolved. A machine vision platform for ceramic tile detection is built and the proposed method is validated with experimental results.