We indicate how granulometries may be useful in the analysis of random sets. We define a suitable size distribution function as a tool in exploratory data analysis and give a new Hanisch-style estimator for it. New Markov random sets are constructed which favour certain sizes above others. Applications and examples on real and simulated data sets are included.

Point processes (msc 60G55), Spatial processes (msc 62M30), Image processing (msc 68U10)
CWI. Probability, Networks and Algorithms [PNA]
Signals and Images

van Lieshout, M.N.M. (1997). Size distributions in stochastic geometry. CWI. Probability, Networks and Algorithms [PNA]. CWI.