复合高低空间分辨率遥感影像的云南省近十年森林扰动提取  

Extraction of Forest Disturbance in the Last Decade in Yunnan Province Based on Composition of Remotely Sensed Image with High and Low Spatial Resolution

作　　者：曹云[1] 祝善友[1] 高玚 张桂欣[1] 张海龙[1] 

机构地区：[1]南京信息工程大学地理与遥感学院,江苏南京210044 [2]南京市浦口区气象局,江苏南京211800

出　　处：《西北林学院学报》2015年第6期203-209,共7页Journal of Northwest Forestry University

基　　金：国家重点基础研究发展计划(2010CB950701);国家自然科学基金(41401471;41001289;41201369)

摘　　要：选择云南省作为研究区域,利用2002、2013年的Landsat数据和2002-2013年的MODIS产品数据,分别构建扰动指数提取森林扰动信息。在此基础上,根据空间位置的对应关系,抽样部分区域,建立MODIS数据与Landsat数据扰动提取结果间的统计关系,将该关系外推到整个研究区域,利用FORMA数据对扰动估算结果进行精度评价。结果表明,复合高低分辨率图像校正后的MODIS扰动提取具有较高精度,与FORMA相比,平均相对误差为10.43%,近10a云南省森林扰动主要发生在大理、丽江、玉溪等林区地带,其中2010年扰动最为剧烈。Selecting Yunnan Province as the study area, this paper adopted the disturbance index （DI） to extract forest disturbance from Landsat data imaged on 2002 and 2013, as well as MODIS products imaged from 2002 to 2013. On this basis, according to the corresponding spatial location of the sampled blocks, the perturbation statistical relationship between the deforestation results of the MODIS data and Landsat data was built, which was then extrapolated to map the forest disturbance within the entire study area, and the accuracy of the disturbance estimation results were evaluated by the FORMA data. The results showed that the calibrated MODIS-detected disturbance by compounding images with high and low resolu- tions had a higher precision. Compared with FORMA, the mean relative error （MRE） was 10.43%. Forest disturbances mainly occurred in Dali, Lijiang, Yuxi in the last decade. Especially, the disturbance of 2010 was the most dramatic.

关 键 词：森林扰动 LANDSAT MODIS 抽样方法 

分 类 号：S758.4[农业科学—森林经理学]