Optimal Combination of Multi-resolution Models in Visualization Techniques

Abstract

Complex VR scenes, containing hundreds of multi-resolut ion models with millions polygons, cannot achieve real-time performance and verisimilitude at the same time. How to adaptively choose the appropriate level of resolutions according to the data volume and hardware/software environment is therefore a critical issue. This paper we propose an optimization algorithm to choose the model resolutions that maximize scene quality while meeting timing constraints. The mathematical model is given by means of the relationship between scene representation and model resolution. Lagrange multiplier method is used to solve the problem of extreme value solution under time constraints. With the rapid development of virtual reality technology, scenarios become increasingly complex, the number of models (including three-dimension solid models, special effects, and environment models, etc.) is increasing, while the data volume in each model (i.e., the number of vertices and triangular faces) also increasing sharply. Although the performance of current graphic hardware rendering system is also rising, but still cannot meet the large volume of data in real-time scene rendering needs. Construction of multi-resolution modeling of three-dimensional model simplify the models with different levels of details which can be selected as an alternative for different drawing demand, thus can effectively reduce the amount of data volume and improve rendering speed. Details of expression can provide different resolution models is not the same, the details of the high-resolution model of high performance, but the time required to draw a corresponding increase. Therefore, how to draw the time constraints under suitable conditions for the resolution of different models to get the highest allocation model expression become multi-resolution model of the key issues. The traditional model resolution are based on the experience of application developers to determine, developers and hardware performance depending on the complexity of the scene rendering environment, pre-established rules of the respective resolution of the model chosen. (1) LOD models use a simplified method for the establishment of three-dimensional city model provides a way to reference (2) for the virtual campus programs achieve three levels of detail model, and select the level of detail based on the model from the viewpoint of the distance. This approach not only spend a lot of time and effort, but once the environment changes need to draw re-enact the resolution selection rule for the model. (3) The importance of index model, considering the distance, speed, user concerns and other factors model, we propose a model for complex multi-resolution scene selection algorithm. The algorithm can realize limit drawing complex scenes, but could not get the optimal solution of the model of multi-resolut ion selected. Therefore, a multi-resoluti on model limit an adaptive rendering method, the method by plotting time and quantitative model to predict performance capabilities, the establishment of a multi-resolution models show that the mathematical model painting problems, and using Lagrange multiplication real time solve the optimal solution of the model to determine the optimal multi-resolution rendering of the model.