北京大学学报(自然科学版)

基于MODIS数据的东北亚森林时序变化分析

付安民1,2,孙国清1,3,过志峰1,王殿中1   

  1. 1. 中国科学院遥感应用研究所遥感科学国家重点实验室, 北京 100101; 2. 国家林业局调查规划设计院,北京 100714; 3. Department of Geography, University of Maryland, College Park, MD 20742;
  • 收稿日期:2009-08-11 出版日期:2010-09-20 发布日期:2010-09-20

Forest Changes Detection in the Northeastern Asia Using MODIS Imagery

FU Anmin1,2, SUN Guoqing1,3, GUO Zhifeng1, WANG Dianzhong1   

  1. 1. State Key Laboratory of Remote Sensing Science, Institute of Remote Sensing Applications, Chinese Academy of Sciences, Beijing 100101; 2. Academy of Forest Inventory and Planning, State Forestry Administration, Beijing 100714; 3. Department of Geography, University of Maryland, College Park, MD 20742;
  • Received:2009-08-11 Online:2010-09-20 Published:2010-09-20

摘要: 基于MODIS传感器在2000?2006年观测的MOD13A1影像产品, 对东北亚森林的覆盖变化进行了专题分析和制图。一方面, 通过生长季节的累加归一化植被指数 GNDVI 和成对双样本 T 检验方法, 在像素尺度上反映发生变化的强度和显著性特征; 另一方面, 对地物交错区域的边缘条带进行缓冲区判读分析, 去除由于多期影像配准位移所导致的错误判断。然后, 利用TM/ ETM+ 影像得到的森林变化 图对上述的森林变化检测结果在两个核心区域进行了比较验证。在 MOD13A1 像素内面积变化 5 ha 以上, 且3 ×3 窗口的空间配准尺度下, 两区域的森林变化制图精度分别达到了80. 24% 和 88. 73% 。

关键词: 北方森林, 森林扰动变化, MODIS, TM/ETM+

Abstract: MOD13A1 products observed by MODIS sensor during the years from 2000 to 2006 were used to map forest dynamic changes in the Northeastern Asia, The mapping analysis was based on the trajectory of multi-year-series normalized difference vegetation index (NDVI) and its spatial morphologic characteristics. The intensity of the changes for a detected forest pixel was characterized by the degradation of the GNDVI and the statistical significance of the changes was determined by a T-test of paired-samples. The false identification of the changes in the transition areas between different vegetation types, which was mainly caused by the mis-registration of the multi-temporal MOD13A1 data, was detected by spatial morphologic analysis and eliminated from final change map. The results from MOD13A1 data were compared with forest change maps derived from TM/ ETM+ images in two local regions. The accuracies were 80.24% and 88. 73%, at 5 ha and 3×3 window level, respectively.

Key words: boreal forest, forest disturbance, MODIS, TM/ETM+

中图分类号: