Research on Real Time Rendering Algorithm for Large Scale Complex Scenes based on Unity
Gang Liu,Ziyi Wang,M. Tan
Abstract:Real time rendering of large-scale complex scenes is widely used in video games, virtual reality, film and other fields. The development of graphics hardware technology. Powerful graphics hardware technology to provide powerful technical support for the gorgeous game screen and the impact of the movie, but the user's requirements for graphics rendering quality far exceeds the scope of graphics hardware rendering capabilities. How to improve the rendering efficiency of large scale complex scenes with software is becoming a hot topic in the field of computer graphics. In recent years, a large number of complex scene rendering algorithm. The commonly used algorithms include occlusion culling algorithm, LOD algorithm, instantiation technology and so on. The author on the basis of previous research, the software realization of LOD occlusion culling algorithm and view dependent parallel optimization, and combine with the same rendering pipeline, the formation of a new large-scale complex scene rendering algorithm efficient. First of all, the author abandons the traditional hardware occlusion query algorithm, and implements the parallel optimization of the occlusion culling algorithm. The key of the hierarchical occlusion map algorithm is the construction of occlusion map. In the preprocessing stage, the occlusion database is derived by removing the specific conditions of the occlusion and removing the object which is not suitable for the occlusion. In the operational phase, covering algorithm from the database to extract the current set of view under the shelter, which rendered to the image space and construct the hierarchical occlusion map, finally use the hierarchical occlusion map in vivo with visual objects by calculating occlusion culling. The view dependent LOD algorithm and occlusion culling algorithm are combined by an array of visible objects. In the preprocessing stage, the algorithm is used to process the scene objects. At runtime, the algorithm extracts objects from the array of visible objects one by one, and calculates the appropriate LOD according to the distance between the object and the current viewpoint and the visibility of the object in the preprocessing phase. Finally, the triangle of the LOD of the calculated object is written into the triangle sequence to be rendered and sent to the rendering pipeline. In this paper, the algorithm is applied to the large-scale factory scene rendering in real-time. Experiments show that the algorithm can guarantee the rendering quality, real-time rendering of large-scale factory scene than the single LOD algorithm or occlusion culling algorithm have different degrees of improvement in the efficiency of rendering, and the scene is more complex, more can reflect the efficiency of algorithm.
Computer Science