Comparison of different platforms for ceramic tiles visual inspection and classification algorithm

Automation of the inspection and classification stage in ceramic tiles production includes the development of machine vision systems. It is performed by digital cameras and different image processing algorithms. In this work we presents the development, implementation and performance comparison of the Moving Average with Local Difference (MALD) method for ceramic tile surface defects detection. The execution times of the image processing algorithms need to be shorter due to needs for faster ceramic tile production. The execution time of the MALD algorithm has been implemented to speed up on three different platforms: CPU, GPU and FPGA. For every platform minimum of two algorithms were developed and implemented in MATLAB's MEX/C++, C++, CUDA/C++, VHDL and Assembly programming languages. Execution times are measured and compared for different algorithms and their implementations.