Optical image reconstruction in highly scattering media.

摘要

Light propagation in turbid media is treated by the theory of radiative transfer. We show that the cumulant approximation to radiative transfer is a major improvement over the diffusion approximation and it describes correctly both the ballistic photons at early times and diffusive photons at later times. Comparisons between the second order cumulant approximation, the diffusion approximation and Monte Carlo simulations are presented. A transport forward model for optical imaging based on the second order cumulant approximation is then proposed. The weight function for image reconstruction under this new model is shown to address the deficiency of the diffusion approximation and reveals the effect of the initial ballistic motion of incident photons.; The framework for image reconstruction for weak inhomogeneities using the Born approximation and the Green's function method is presented, following a brief review of the inverse problem and regularization methods. We present several three-dimensional near real time image reconstruction algorithms which exploit the inherent symmetry of planar geometries and transform the image reconstruction problem into reconstructions of decoupled spatial Fourier components of the optical property in the Fourier space. In particular, based on the theory of propagation of a spatial Fourier component of the scattered wave field inside a turbid medium, a new optical diffuse imaging methodology is presented which uses the two-dimensional Fourier transform of photon intensity on a plane to detect inhomogeneities in a highly scattering turbid medium when illuminated by a picosecond (near) plane wave pulse.; Finally, we provide a summary and end with a discussion about future directions of the optical tomographic imaging research.