Shape reconstruction from generalized projections

Abstract

In this thesis we develop methods for recovering the three-dimensional shape of an object from generalized projections. We particularly focus on the problems encountered when data are presented as discrete image fields. We demonstrate the usefulness of the Fourier transform in transferring the image data and shape model projections to a domain more suitable for gradient based optimization. To substantiate the general applicability of our methods to observational astronomy, we reconstruct shape models for several asteroids observed with adaptive optics, thermal infrared interferometry, or range-Doppler radar. The reconstructions are carried out with the ADAM software package that we have designed for general use.