大规模动态点云数据多尺度交互式实时绘制算法研究

项目名称： 大规模动态点云数据多尺度交互式实时绘制算法研究

项目编号： No.60873130

项目类型： 面上项目

立项/批准年度： 2009

项目学科： 生物科学

项目作者： 万旺根

作者单位： 上海大学

项目金额： 30万元

中文摘要： 科学数据和信息爆炸性增长，给人们带来的巨大挑战是如何有效理解和使用这些科学数据。目前很多数据形式是点云数据，其特点是时变、非结构化、不规则、多领域、多维以及大规模，需要新的理论和方法进行处理。本项目根据科学数据可视化原理，提出多尺度绘制方法对点云数据进行实时绘制，通过实现三维交互式可视化，提高科学数据的可理解性，加强人们对科学数据的感知和认知能力；项目提出具有普遍意义的方法，以适应摩尔定律。主要研究内容包括多尺度模型生成算法、点云数据实时渲染算法、快速变化的几何表面重构算法以及实现内外存数据交换的数据预测调度算法。随着三维扫描设备的广泛应用以及模型顶点数的不断增多，基于点云的三维重建和渲染方法将发挥其在三维建模与绘制方面的绝对优势。在大规模三维数据可视化、计算视觉、图形处理和生物视觉建模方面具有广泛应用价值。

中文关键词： 动态点云数据；混合渲染；多尺度；方向性滤波；三维动态显示

英文摘要： With the explosive growth of scientific data and information, a great challenge brought is how to understand and use these scientific data efficiently. Now most form of data is cloud point data, which is characterized by time-variance, unstructured, irregular, multi-domain, multi-dimensional and large-scale, and need new theories and methods to processe. According to principle of scientific data visualization, the project proposes multi-scale, 3D interactive, real-time rendering method for point cloud data, to improve the intelligibility of scientific data, and strengthen people's ability of perception and cognition of scientific data. The project proposed the method of universal significance to adapt Moore's Law. The research topics include multi-scale model generation algorithm, the point cloud data real-time rendering algorithms, reconstruction algorithm in the rapid challenging geometric surface, and achieve internal and external data stored in the exchange of data to predict scheduling algorithm. With the widely use of three-dimensional scanning device and lardge growth of 3D model vertices, the methods based on the three-dimensional reconstruction and rendering of point clouds will play its absolute advantage in three-dimensional modeling and rendering. This method will has great value of use in large-scale three-dimensional data visualization, computer vision, graphics and biological visual modeling.

英文关键词： dynamic point cloud data；hybrid rendering；Multi-scale；directional filtering；3D dynamic display