Computer Graphics Techniques for Modeling and Rendering Weathered Materials

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

Image synthesis is an important problem for such diverse fields as architecture, lighting and industrial design, scientific visualization, simulation systems, entertainment, and advertising. An important, largely unexplored area of computer image generation is the simulation of weathering and its effects on appearance. Weathering results from the interaction of the environment with the materials in the world. Current computer graphics models of materials are based on the physics of idealized surfaces that exhibit ideal diffuse, glossy, or specular reflection. However, such idealized materials rarely exist and are hard to produce. In order to make such idealized materials look realistic, animators and modelers often apply many textures. Unfortunately, this use of texture maps is very ad hoc and labor intensive, and it is in conflict with the use of idealized models. What is needed are physically-based models of real materials that incorporate their possible life histories, and hence, possible appearances. The goal of this project is to devise new computer models of materials that can sustain a range of appearances over time. Through an NSF-CAREER award, we have made significant headway on this problem and built critical infrastructure for future work. In this Accomplishment-Based Renewal (ABR) proposal, we provide a brief summary of several new directions that build on our initial work, including the study and modeling of several additional materials, simulations of environmental factors that affect appearance, surface representations and modes of interaction, a measurement device, and techniques for rendering light transport in materials with complex internal structures. This work has the potential to greatly expand the visual realism of synthetic images and reduce the reliance on tedious ad hoc techniques by providing fundamental, sophisticated models for materials and the processes that affect their time-varying appearance. In addition to the work's importance in computer graphics, this research should benefit such applications as scientific visualization, architecture and preservation, advertising, entertainment, art restoration, automotive design, geology, civil engineering, and environmental sciences. Because the project draws on material from several other disciplines such as applied math, geology, materials science, optics, and physics, it promises to add new dimensions to the field of computer graphics.