Real-time rendering method based on energy consumption-error precomputation

Abstract

The invention discloses a real-time rendering method based on energy consumption-error precomputation, comprising: determining the spatial level structure of the scene to be rendered through adaptive subdivision of the space positions and look space of the camera browsable to the user in the 3D scene to be rendered; during the process of adaptive subdivision of the space, for each position subspace obtained at the completion of each subdivision, obtaining the error and energy consumption of the camera for rendering the 3D scene using a plurality of sets of preset rendering parameters in each look subspace at each vertex of the bounding volume that bounds the position subspace, and Pareto curve of the corresponding vertex and look subspace is built based on the error and energy consumption; based on the current camera viewpoint information, searching and obtaining the target Pareto curve in the spatial level structure to determine a set of rendering parameters satisfying the precomputation condition as optimum rendering parameters to perform rendering. The present invention not only saves a great deal of energy, but also ensures the quality of rendering result and extends the battery life.