Cluster analysis in ceramic tiles images orientation

The image processing and analysis applied in manufacturing industry is increasing. The applications are mostly in quality control, such as visual inspection. If the processing and analysis are used in production lines for ceramic tiles, it could be very demanding and it has to be very fast. So the use of powerful computer architectures such as computer clusters is applied. In this paper we present the image processing method of ceramic tiles that determine image object dimension and orientation. Processing was done on Beowulf cluster using different number of nodes and processors and starting MPICH and LAM/MPI jobs

[1]  N. Otsu A threshold selection method from gray level histograms , 1979 .

[2]  Axel Keller,et al.  Anatomy of a Resource Management System for HPC-Clusters , 2000 .

[3]  Tarek A. El-Ghazawi,et al.  Performance evaluation of selected job management systems , 2002, Proceedings 16th International Parallel and Distributed Processing Symposium.

[4]  J. F. Reid,et al.  RGB calibration for color image analysis in machine vision , 1996, IEEE Trans. Image Process..