Real-time atmospheric cloud visualization framework for urban simulation system

Abstract

Clouds are important aspects of any natural outdoor scene especially in urban simulation system and atmospheric analysis. However, the existing methods within computer graphics only offer the simplest of cloud representations. The problem is how to provide a means of simulating heterogeneous types of clouds and other atmospheric features that are suitable for virtual environments. Methods for cloud visualization are available within the area of urban planning, but numerical calculation of the systems are computationally intensive which give more numerical accuracy than what are required for graphics and restricted to the law of physics. Furthermore, pure physics-driven methods are not always producing the visual plausible results. Consequently, three-dimensional clouds are very difficult to model and render when considering the efficiency and visual accuracy aspects. The purpose of this research is to propose a cloud visualization framework for urban simulation system that would be able to produce high real-time rate and higher degree of visual realism. In order to solve the data representation complexity and the computation overhead problem, the generalized multiple data representations and the cloud rendering method are proposed respectively. Combinations of ground, atmospheric, cloud and shadow models in the proposed framework is the key for realizing the real-time cloud visualization in urban simulation system.