Research And Implement Of Point-Based Rendering Technology

Followed by a triangular mesh model for the study of geometry processing technology,a new point-based geometric model for geometry processing is now being more and more concerned to researchers.Point-based rendering(PBR) is one of new research direction in the field of digital geometry processing.It is to render an image based on points instead of using the traditional method of triangular mesh and only records the information on points,which leads to a new way of fast rendering for huge sampled data.This paper does a research and implement on PBR and also have designed a way for point clouds based on the broad research and in-depth analysis to the results which have been in the traditional algorithm.The main research and achievement as follows:1.A adaptive algorithm for Point-Based modeling is designed.The algorithm is based on point clouds data grid,and the establishment of K-Nearest neighbors and calculate the normal,and it can establish adaptive radius.This algorithm is not only suitable for uniform distribution of point clouds models,but also will help to improve the uneven distribution of point cloud models' realistic effects.2.The image reconstruction method based on filtering and inpatinting was introduced.As the own characteristics of modeling algorithms,the rendering results may exist two phenomenons.First,the image pixel gray value abrupt changed.Second,as partial sampled data is abnormal lossed,the generation of images will have some holes.This method is designed to solve these problems,and makes the visual effect more realistic.Besides,a prototypical system based on these algorithms is implemented in this paper,and several typical examples are given as well.The results of these examples indicate the availability of the algorithms above.