项目名称: 基于DEM样本的交互式地形合成方法研究
项目编号: No.61502088
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 计算机科学学科
项目作者: 邹昆
作者单位: 电子科技大学中山学院
项目金额: 21万元
中文摘要: 地形生成是计算机图形学领域的经典问题,有着广泛应用。相对于过程化和基于物理模型的方法,基于数字高程模型(DEM)样本的地形合成能提供更方便直观的用户控制,且所生成地形的真实感更强,但速度问题是阻碍其应用的一大障碍,此外现有方法也存在交互控制方式单一且不够人性化的不足。本项目选定基于DEM样本的交互式地形合成为研究方向,通过利用GPU并行计算、KD树以及基于邻域相关性的传播对各合成阶段加速,并对多细节等级的地形合成进行研究,在实现高质高效合成的同时,避免了块大小的人为调节,还可为用户提供渐进显示的合成结果。此外,通过对多种现有的人机交互接口进行有机整合,并提供人性化的地形特征绘制自动完成功能以及特征复制画刷,实现更加方便、直观而多样化的用户控制接口。
中文关键词: 地形合成;数字高程模型;基于相关性的传播;地形特征提取;基于草图的建模
英文摘要: Terrain generation is a classical problem in Computer Graphics, and has a wide range of applications. Compared with procedural and physics-based terrain generation methods, digital elevation model (DEM) sample-based terrain synthesis can provide more convenient and intuitive user control, and generate more realistic terrains, but speed is a big obstacle for its application, also, the user control interfaces of existing methods are monotonous and not humanized . This project chooses DEM sample-based interactive terrain synthesis as the research direction. By utilizing GPU parallel computing, KD tree and neighborhood coherency-based propagation to accelerate each synthesis stage, and by research on multi-level of detail synthesis scheme, high-quality and efficient synthesis are achieved, the artificial tuning of patch size is avoided, and users are provided with gradual display of synthesis results. Besides, by organic integration of some existing human-computer interaction interfaces, and based on the implementation of humanized autocomplete function for terrain feature drawing and feature copying brush, convenient, intuitive and diverse user control interface is realized.
英文关键词: terrain synthesis;digital elevation model;coherency-based propagation;terrain feature extraction;sketch-based modeling