摘要
土地利用/土地覆被变化(LUCC)是影响生态环境和气候变化的主要驱动力之一,同时又是受其影响的结果。LUCC研究对于开展生态环境变化及气候变化的研究均具有重要的意义。针对目前境外LUCC研究中土地利用/覆被分类效率低的问题,探索一种适用于大数据量而又精度较高的分类方法。以额尔齐斯河国外部分-斋桑湖流域为研究区域,以1990年及2007年的Landsat TM/ETM+夏秋季影像以及DEM作为数据源,综合利用影像光谱、纹理信息参与到决策树构造中,进而利用决策树分类方法分别提取这两个时期的土地利用/覆被空间分布信息,最后分析两个时期的土地利用时空变化状况。实验结果表明:(1)光谱与空间纹理信息参与的决策树分类方法具有较高的分类精度;(2)两个时期的土地利用变化分析发现,近20年来该区域土地利用发生了较大的变化,耕地和灌木林地大面积减少,而低覆盖度草地和未利用地却显著增加。
The land use/land cover change is one of the main driving forces that affect the ecological environment and climate change, and also the result of its impact. LUCC research for the conduct of coo-environmental changes and climate change studies is of great significance. View of the current LUCC study,land use / cover classification accuracy and efficiency is low. On focus this issue, the paper tries to find a new classification method that can be used to the large scale data classification and also has a more classification precision. The paper takes Irtysh River -Zaysan Lake as the study area, and Landsat TM/ETM ＋ image in year 1990 and 2007 as well as Aster DEM as data source, and comprehensive utilization of imaging spectroscopy, texture information to participate in the decision tree structure, and then take decision tree classification method to extract information of the two periods land use/land cover, and finally to analysis the LUCC information of this area. The results show as follows ： （ 1 ） the combination of the image spectrum and texture increased the decision tree＇s classification precision dramatically; （2） The study area has a great change in land use in the past 20 years, especially the farmland and shrub decreased, while the grassland and unused land increased significantly.
出处
《干旱区地理》
CSCD
北大核心
2010年第2期189-195,共7页
Arid Land Geography
基金
中国科学院重要方向性项目(KZCX2-YW-307)
国家自然科学基金(40730633)
国家支撑计划项目(2008BAB42B09-3)
