Reconstruction of the Lung Geometry for the Simulation of Aerosol Transport

Abstract

ABSTRACT

Most computer simulations of gas and particle transport in the human lungs use idealized lung geometries. In this study we present an approach to make real lung geometries available for computer simulations. Two-dimensional images of a cast of canine lungs were obtained by high-resolution computer tomography (HRCT). These images were then combined to a three-dimensional grayscale volume data set that was processed by means of image processing. From the resulting three-dimensional volume data set a three-dimensional polygonal surface representation consisting of triangles and quadrilaterals was obtained. The surface reconstruction method used, surface primitive algorithm (SPA),⁽¹⁵⁾ does not falsify the original CT data and thus the surface representation shows an excellent conformity with the original cast. This polygonal surface can be used to generate a finite element mesh necessary to perform computational fluid dynamics calculations. Future aspects of the present work will be the production of hollow casts from the polygonal surface reconstruction, so that gas and particle flow fields can also be studied experimentally. Thus, for the first time a realistic lung geometry can be used for studying particle transport theoretically and experimentally through intrapulmonary airways.