Point set registration for reduced geometry mismatch during estimation of mass transfer properties in osmotic dehydration of complex-shaped foods

Abstract

The use of point set registration (PSR) in a novel approach for assisted geometry generation of complex-shaped foods during mass transfer modeling of osmotic dehydration (OD) is presented. The procedure is aimed at reducing the effect of geometry mismatch during estimation of diffusion coefficients. The PSR is based on a modified iterated closest point (ICP) algorithm. The proposed methodology was applied to obtain the optimum geometrical parameters describing the shape of white mushroom pilei (Agaricus bisporus) as a revolution solid in the oblate spheroidal coordinate system. Roundness was proposed as a shape factor to evaluate geometry mismatch. An unsteady state 2D diffusion model with the PSR-generated geometry was proposed to describe experimental data, where product shrinkage was included as the contraction of ellipse focus. Moreover, a semi-theoretical approach is presented allowing the estimation of diffusion coefficients by the well-known straight-line fit approach in the proposed geometry. Water and solute diffusivities in the product were estimated and further compared with those previously reported for hemispherical shell geometry. As opposed to hemispherical shell geometry, no significant difference was found () in roundness between PSR-generated geometry and real product shape. Diffusion coefficients estimated with the PSR-generated geometry were significantly lower than those calculated with the hemispherical shell model, indicating that proposed approach may be helpful to reduce geometry mismatch during the estimation of mass transfer properties.