Study On Modeling And Rendering Of Point Set Surface

Point sample, as a fundamental primitive for geometric modeling and rendering, has gained increasing attentions in computer graphics community. The primary mission in study on point model is to solve the problem that how to compress and render the mass point cloud. Firstly, we investigated the characteristic of point cloud data and introduced a compression algorithm of point sampled geometry by using KD-tree in this thesis. The compression algorithm found on the theory and method of geometry compression. The algorithm compresses the point sample by dividing it into smaller size on condition that keeps the consistent in visual. The algorithm can compress the space properties of point sample into 1 5. The compressed point sample can be transferred progressively after serialization. Secondly, the thesis purposes another algorithm to render the compressed point data which can render the model without reconstructing the full structure of KD-Tree. The algorithm uses the View-Frustum Culling, Back-Face Culling and Level of Detail method, and boosts the speed and precision on rendering.