Watershed Functions Applied to a 3D Image Segmentation Problem for the Analysis of Packed Particle Beds

In an effort to advance the area of particle science and technology the University of Utah has been using cone beam X-ray microtomography as a major tool for the quantitative three dimensional (3D) analysis of packed beds of multiphase particles. In this regard, efficient imaging algorithms have been developed and applied in order to extend two dimensional (2D) image analysis techniques for the solution of three 3D problems. Based on this research, image analyses of different particle samples were carried out for identification, segmentation, and quantitative volume characterization of irregularly shaped particles. The whole analysis involves acquisition of a 3D image by X-ray microtomography and a sequence of image processing tools for thresholding, distance transformation, maxima detection using reconstruction by dilation, and a watershed 3D algorithm for segmentation and identification of irregularly shaped particles in three dimensions. Several examples of the 3D analysis for irregularly shaped particles are presented and the results discussed. Application of these algorithms is demonstrated by the determination of coal washability curves.