Intelligent reconstruction of grain structures in images of 2D cuts of crystalline materials

Some basic mechanical properties of metallic materials with grain structure depend on the grain size. The grain structure can be recognised in two-dimensional cuts of crystalline materials or in surface scans. The morphological analysis of such micrographs is very often complicated due to the signal noise. An intelligent method of grains reconstruction based on neural networks is presented in this contribution. # 2004 Published by Elsevier B.V.