Preview	PDF, English (Dissertation) - main document Download (194MB) | Terms of use
Preview	PDF, English (Errata) - Additional Metadata Download (29kB) | Terms of use

Abstract

Astronomy is as old as mankind. Its progress and evolution in history has been parallel to the very process of human development. The advent of modern physical sciences caused an exponential progress in astronomy and, by extension, astrophysics, which has seen amazing development over the twentieth century, and especially in the first decades of the new millennium. In this work we firmly set one foot in astrophysics and the other in scientific visualization, while we develop and use the latter to respond to scientific research questions posed by the former. This thesis has been interdisciplinary since its very inception. It bases on scientific visualization and computer graphics and aims at providing and developing methods and techniques particularly designed to help analyze and understand astrophysical systems and processes. It is composed of two distinct parts: In the first part of this thesis, we borrow from the field of vector field topology, classically concerned with the non-inertial dynamics of static and time-dependent flows, and develop and extend it to enable the study of force-induced, inertial systems, like those governing most of our universe. We then focus on the various time dimensions ingrained in vector field topology, and introduce a framework for its analysis and exploration based on novel derived aggregation fields that capture various properties of the underlying system, and present them in digestible representations in order to aid in the interpretation and analysis of the time domain. Finally, we address the very underlying structures that define complex dynamical systems, and provide a quantitative approach for their extraction and subsequent analysis. The second part of this thesis concerns itself with the exploration and representation of very-large astrometric and astrophysical datasets. In this part, we base on computer graphics and visualization, and develop a technique to efficiently and interactively navigate through catalogs of billions of objects, introduce a method to effectively use floating-point arithmetic in the representation of the known universe without suffering from precision loss, propose a novel logarithmic function to enable limited-resolution depth buffers to function for astronomically large scenes, and present an integrated visualization of relativistic effects, including relativistic aberration due to the observer’s motion and gravitational waves.