Multilevel data simplification methods for flood process visualizations based on visual perceptions

Abstract

ABSTRACT Visualizations of flood processes require a balance between scientific authenticity and realistic expressions. As the flood area increases, the amounts of time-series data that are produced by flood models often exceed the capacity of computer rendering, leading to poor user experiences and low visualization effectiveness. Therefore, this study carried out visual perception experiments and proposed a multilevel simplification and scheduling method for flood process data based on visual perceptions. This method first simplifies flood data by flood feature point selection according to conclusions related to visual perception, then triangulated irregular networks are built, and finally, flood data scheduling and rendering are achieved through multiple levels of detail combined with visual ranges. An experimental analysis using a 3D visualization prototype system shows that the proposed method can improve the efficiency of visual expressions while preserving flood morphology features.