The visualisation and rendering of digital elevation data
The thesis addresses a longstanding cartographic problem, of how to visualise the Earth's surface relief in an effective and meaningful way. The problem is narrowed to relief defined by digital elevation data and visualised as a static, orthographic representation. It is approached in three steps: firstly research focuses on determining the most useful form of graphical representation to be pursued; secondly the theoretical basis of computer visualisation is investigated through a three-model framework, prompting a number of directions where solutions might be developed; and thirdly the development and engineering of a system is reported which models and renders widely available elevation data, and which provides flexibility in its input variables. The developed system is then applied to specific cases of relief visualisation, and new graphical forms are developed. The investigation of past and current approaches to relief representation, and a review of computer-graphic rendering of simpler geometrically defined objects, have revealed some limitations in commonly used relief visualisation systems, but have established the simulation of light and shade as still the most promising line of development. Analysis of the component variables of surface visualisation and rendering has led to the visualisation paradigm of three parametric models - of elevation, illumination and reflectance. Some attractive qualities, including widespread availability, of the contour elevation model have been identified, and a system has been developed which reconstructs surfaces from this data structure in a more effective way than typical current approaches. The system is also designed to support more complex illumination and surface reflectance models than the somewhat simplistic scenarios commonly available. The thesis reports the application of the system to generate surfaces from contour data, and experimentation with multiple coloured light sources and varying degrees of surface specularity. Evaluation of system implementation, and of the qualities of a representative set of graphical products, is addressed through six design criteria within a context defined by a typical mapping application. This has led to conclusions that the system and the new graphical forms have a number of virtues, including close fidelity with the source data, and significant improvements in visualisation.