INTEMA   05428
INSTITUTO DE INVESTIGACIONES EN CIENCIA Y TECNOLOGIA DE MATERIALES
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
PSD Retrieval by Bayesian Data Fusion via Metropolis-Hastings and the Jackknife Procedure
Autor/es:
GLORIA L. FRONTINI; FERNANDO A. OTERO; GUILLERMO E. ELIÇABE
Lugar:
Buenos Aires
Reunión:
Workshop; 2017 XVII Workshop on Information Processing and Control; 2017
Institución organizadora:
RPIC
Resumen:
This article analyzes the performance of combining information from Scanning Electron Microscopy (SEM) micrographs with Static Light Scattering (SLS) measurements for retrieving the so-called Particle Size Distribution (PSD). The corresponding data fusion, which is formulated as a Bayesian inverse problem, is implemented using an emblematic Monte Carlo Markov Chain (MCMC) technique, the Metropolis-Hastings (MH) algorithm. Furthermore, the actual PSD is assumed to be exactly represented by a log-normal distribution, in order to reduce additional processing errors. The prior statistics corresponding to the SEM micrographs have been achieved by means of the Jackknife procedure used as a resampling technique. Monte Carlo-based statistical tools are also employed to assess the quality of these priors. The likelihood term for the Bayesian approach is computed considering independent normal measurements generated from a simplified SLS model, the Local Monodisperse Approximation (LMA), also used as the forward linear model. Finally, an experimental example is analyzed using the priors generated with the proposed procedure and parametrization and resulting estimations are compared to the achieved in a previous article and discussed.