A Simplification Method for Radiosity Global-illumination Meshes
一种辐射度全局光照网格模型的简化方法

The solution to the global illumination problem using radiosity method is usually represented by an illumination meshes, which are frequently very finely tessellated in order to capture shadow boundaries and other significant illumination features. So the geometry complexity of a complex scene with global illumination effect is beyond the capability of the current computer graphics workstation. In this paper, an automatic method is presented for simplifying the meshes produced as the solution to global illumination problem using radiosity method. Based on the character of the illumination meshes produced by radiosity and the characteristic of human eyes, it simplifies first by merging faces with maximum relative radiosity difference below the threshold that users define. And then the vertices on the perimeters of the merged face are deleted to further simplify the original meshes model. The model simplification samples based on these methods are given and the results show that the algorithm only significantly reduced the complexity of original meshes model, but also performed well to capture the coherent shadow discontinuity feature of illumination meshes.