Establish of visualization method for large-scale forest landscapes based on the national vegetation survey of Japan.

我们专注于大规模森林景观的高速可视化，因为它从来没有被处理过。首先，我们提出了树木的参数化和分析技术，这些技术能够表示依赖于物种的特定形状。这项技术旨在代表环境部植被研究覆盖的大约200种树木。其次，为了可视化大规模森林，我们提出了一种具有以下三个特点的方法：(1)从日本国家植被调查数据集程序化地生成森林。(2)利用分形噪声函数生成代表森林中树木自然排列的植被边界。(3)为了实现高速可视化，提出了一种适合在GPU上并行处理的树生成方法。该系列技术可以实现全国植被调查生成的大比例尺森林景观的高速可视化。

We concentrate on the high-speed visualization for large-scale forest landscapes, since it has never been dealt with. First, we proposed parameterization and analysis techniques for trees which enable to represent specific shapes dependent on species. This technique aims to represent about 200 species of trees which are covered in vegetation research of ministry of the environment. Second, to visualize large-scale forest, we proposed a method which has the following three characteristic features : (1) It procedurally generates a forest from data sets of the national vegetation survey of Japan. (2) It generates a vegetation boundary which represents a natural arrangement of trees in the forest by using a fractal noise function. (3) To realize high-speed visualization, we propose a generating method of trees which is suitable for parallel processing on GPUs. The series of techniques can realize the high-speed visualization for large-scale forest landscapes generated from the national vegetation survey.