A New Pattern Recognition Method for Parametric Modelling of Random Fields
S. Klostermann, D. Vogt, S. Lippert, O. von Estorff · Civil-comp proceedings · 2011
To predict more reliably the real system's behaviour by means of simulation, probabilistic models can be created that account for aleatory uncertainties. Local effects of uncertainty can be modelled by random fields in the n-dimensional feature space. In this paper general constraints for the parameterisation of random fields are identified. A new multi-scale method for a pattern recognition based parameter-isation of random fields is developed: The approximated differential scale-space (ADSS). The ADSS yields parameter sets that describe the relevant characteristics of the random field and allows synthesising of simulation samples. The capabilities of the new method are demonstrated for one- and two-dimensional application examples. © Civil-Comp Press, 2011.