Development of a vegetation environment classification method based on degree of urbanization

利用人口密度、植被（自然）面积和农业面积数据，提出了一种基于城市化程度的植被环境分类新方法。利用国家数字信息三阶网格，对每个像元半径5km范围内的人口密度和植被面积的下降阶曲线进行聚类分析。为了管理这些数据之间不同的动态范围，开发了一种新的分级方法，使这种新的分类方法能够应用于日本的任何部分和/或整个地区。虽然全国人口调查每5年进行一次，但自1990年以来，数字土地利用信息没有更新。为了定期应用这种新的分类方法，选择近基区作为试验点，利用Landsat TM数据获取土地利用面积的更新信息。首先，对1989年1个场景TM数据进行分类并聚类成3阶网格，与数字土地利用信息进行对比，检验土地覆盖分类精度；在此基础上，采用1996年TM数据对1995年的土地覆盖进行分类，采用1995年全国人口调查数据对1995年的植被环境进行分类。通过本研究，利用Landsat TM数据和全国人口调查数据，证明了新的植被环境分类方法能够周期性地应用于植被环境的细微变化。

A new method concerning vegetation environment classification based on degree of urbanization was developed using population density, vegetation (natural) area and agricultural area data. Classification was done by cluster analysis of descending-order curves made from population density and vegetation area within 5km radius from each pixel, using 3rd-order mesh of National Digital Information. To manage different dynamic ranges between these data, a new scaling method was developed, which made able to apply this new classification method to any part and/or whole area of Japan.Although National Population Survey has been conducted 5 years interval, digital landuse information was not updated since 1990. To apply this new classification method periodically, Kinki district was selected as a test site, and Landsat TM data was used to get updated information of landuse area. First, one scene of TM data of 1989 were classified and aggregated into 3rd-order mesh to test land cover classification accuracy comparing with digital landuse information. As this result was successful, land cover of 1995 were classified using TM data of 1996, and vegetation environment classification method was applied using 1995 National Population Survey data.By this research, the new vegetation environment classification method was proved to be able to apply periodically to fine changes concerning vegetational environment, using Landsat TM data and National Population Survey data.