A Distributed Simplification System For Large-scale Complex Scene

The real-time rendering of large-scale scene is currently hot in the field of computer graphics research. As the massive data of large-scale scene, the current graphics hardware is too slow to render it at real-time. A commonly used method to accelerate the rendering is using mesh simplification to build a multi-resolution model of the scene, then at rendering time an appropriate simplified model can be selected instead of the refined one. But when the scene is a dynamic, or it can be edited, the pre-simplified model may become invalid, and then it comes up to us the need of real-time simplification. Unfortunately most of current methods do not meet the need of real-time simplification. A natural solution is partitioning the simplification task into several subtasks, and then distributes them to different simplification ends to execute parallel. If we have enough ends, then simplifying at real-time can be realized. This is the idea of Distributed Simplification System.This thesis presents the design and implementation of a distributed simplification system for large-scale scene.The system is composed of a central controller end and several simplification ends. The central controller end is responsible for the task distribution function, which selects an appropriate simplification end for each task. The simplification end receives the task and executes it. It can be multi-threaded, so it can execute several task parallel at one time. Among the first of the thesis it introduces the physical structure and mandate process of the system, in forms and texts. Then it introduces the central controller end and the simplification end in order: first is the distribute algorithm of the central controller end and next is the multi-task process of the simplification end. Finally, the net transmission agreement is also detailed introduced.After the introduction of each part in the simplification system, the thesis subsequently introduces the simplification algorithm and the structure used in the system. First it briefly introduces the Gaps algorithm, which is the core of the simplification end. Then comes the data format and interface of a genetic triangle mesh structure. This structure can be used to facilitate the design of the Gaps algorithm.At the last part of the thesis we illustrate a design of a drawing system, which can utilize the distributed simplification system. We briefly introduce the structure and algorithm of the drawing system, and then use this system to bring forth the practically efficiency of our simplification system.