Classification board for real time image segmentation

We present in this paper the realization of a classification board, for real-time image segmentation. The classification of each pixel is completed using a real time extraction of attributs and a geometric classification method by stress polytope training, which ensures a high decision speed (100 ns per pixels) and good performances. The decision operator has been integrated in the form of a full custom circuit, and the extraction of parameters is performed using a single high density FPGA. 1INTRODUCTION To increase productivity and quality on automated manufacturing lines, it is often necessary to use vision-based on-line inspection systems to automatically detect defects in the manufactured parts. In view of image segmentation in real-time (20 ms/image), we have developed a board which realises image segmentation using a FPGA and a classification circuit. The FPGA extracts significant parameters and transmits them to the classification stage. To work at the video rate, it is necessary to take a decision in 100 ns per pixel, in accordance with the parameter-vector (figure 1). AAAA AAAA AAAA AAA AAA AAA AAAA AAAA AAAA AAAA AAAA AAAA AAAA AAA AAA AA AA AAA AAA AAAA AAAA AAAA AAAA AAAA AAA AAAA AAAA AAAA AAA Vect = ( p0, p1, p2, p3)