Image segmentation and analysis for porosity measurement

The lack of specific commercial software to analyse CT slice images of metal foams prompted the development of an automated toolbox for reliable and fast characterisation. The CT slice images are first segmented before functionally important parameters are extracted. The segmentation method uses the marker-based watershed transform, supplemented by edge linking. 2D characterisation derives key parameters of individual pores as well as group statistics such as distribution and wall thickness. This paper presents progress made in the refinement of the segmentation step in the use of edge linking, and highlights the results obtained from 2D characterisation. A preliminary performance comparison of the automated toolbox is also made, which shows that the automated toolbox is generally consistent in processing the different types of foams. It also achieves a significant improvement in speed by a factor of 10 over manual segmentation. (c) 2007 Elsevier B.V. All rights reserved.