Traditionally two methods have been applied to terrain rendering: 1. Cell-based rendering of pre-computed, decimated levels of detail (LOD), and 2. Real-time creation of a terrain skin with continuous LOD whereby only required polygons are drawn based on distance from eye-point. The objective of this research is to investigate a terrain rendering technique that combines benefits of both prior methods, without incurring the high compute costs of a full, real-time meshing solution, given the current cost/performance trends in graphic versus cpu subsystems. Factors other than distance, such as angle of incidence and complexity will be evaluated to select the structure of the image to be rendered.
