自适应多特征融合的真实感地形快速绘制

Fast rendering of realistic terrain based on adaptive multiple features fusion

针对真实地形可视化中数字高程模型(DEM)数据结构复杂且绘制速度不佳的问题,提出一种基于自适应多特征融合的真实感地形快速绘制方法。引入地形高程熵,对真实的DEM高程数据进行特征提取以生成地形总体框架;利用随机中点位移分形算法并根据地形特征优化分形参数来增加地形高频细节;计算视点与地形之间的距离阈值,并对应于层次细节(LOD)等级,以实现地形自适应的调度,再根据不确定性判定因子对地形特征进行更新。最后对本文算法进行并行处理,充分利用图形处理单元(GPU)技术对地形进行加速绘制。实验结果表明,该方法生成的地形具有较高逼真度和较好实时性。

Focusing on the complex data structure of digital elevation model （DEM） and the poor rendering speed in real terrain visualization, we propose a fast rendering method based on adaptive multi-feature fusion for realistic terrains. To introduce the entropy of terrain elevation, real DEM elevation data can be extracted for generating an overall framework. According to the random midpoint displacement fractal algorithm and optimized fractal parameters, the high frequency detail can be increased. To calculate the distance threshold between the viewpoint and the terrain, this corresponds to the level of details （LOD）, so that it can achieve adaptive scheduling. Additionally, by using the uncertainty determinant factor, the characteristics of the terrain profile can be updated. Finally, the algorithm is carried out using parallel processing, taking full advantage of the graphic processing unit （GPU） for accelerating the terrain rendering. The experimental results show that the generated terrain has higher fidelity and better real-time capabilily.