节点文献

基于光谱特征信息的芦苇生物量反演研究

Retrieval of Reed Biomass Based on Spectral Information

【作者】 陈爱莲

【导师】 李新通;

【作者基本信息】 福建师范大学 , 地图学与地理信息系统, 2009, 硕士

【副题名】以辽宁双台河口自然保护区为试验区

【摘要】 芦苇作为辽宁双台河口湿地的第一生产者,其生长状况直接决定了该湿地生态系统的生态功能,芦苇生物量的遥感研究对湿地生态系统的质量评价和生态系统的保护具有重要的意义。本项研究以辽宁双台河口自然保护区为试验区,根据实地测量的芦苇反射光谱数据,建立该区域芦苇的光谱数据库,提取芦苇光谱维特征参数;并以光谱维特征为依据选取卫星数据,分析卫星数据与实测芦苇光谱特征的相关性,进而应用光谱角角度匹配、光谱特征拟合、二进制编码等三种光谱匹配技术,研究卫星数据光谱与实测芦苇光谱的匹配度,提取影像的芦苇像元,作为大面积自动估算芦苇生物量的基础。分析卫星数据反演所得的DⅥ、NDⅥ、RDⅥ、ARⅥ与实测芦苇生物量的变化趋势和相关关系,选择相关性良好的植被指数拟合其与芦苇生物量的回归关系,建立芦苇生物量估算的回归模型,并应用该模型对二进制编码分类的芦苇像元进行生物量估算和生物量分布特征分析;复合人工分类的芦苇长势图和MODIS等数据探讨TM数据反映芦苇生物量的能力及其不确定性;最后提出了进一步研究的关键和重点。

【Abstract】 Reed is the primary producer of the Shuangtai Estuary Wetland. How well reed grows has a great controlling impact on the ecologic function of this wetland ecosystem. Thus, study on reed biomass using remote sensing technology makes good sense because it will help us in estimating the ecologic function as well as protecting the ecosystem.This research took Shuangtai Estuary Nature Reserve as a case study, during which, reed’s spectrum was collected in field firstly with Fieldspec FSR VNIR., then the Vegetation Spectral Features (VSFs) of reed were extracted and reed spectral library was built up. TM images were selected and spectrum of reed on them was then analysed to see how well it related to field-collected spectrum. Spectrum matching methiods such as Spectrum Angle Matching (SAM), Spectral Feature Fitting (SFF), Binary Encoding (BE) were tried to classify reed pixels from other pixels, then into different kinds with respect to their different field spectrums.Vegetation Indexes (VIs) were derived from TM images, and their relationship with field-collected biomass were studied. Among the VIs, the ones that were well related to biomass were chosen to conduct regression analysis with biomass, and models to estimate reed biomass from VI were finally created. In the end, artificial classification was done to see how well the Matching methods do in classification of reed pixels, as well as how well the estimation models work, and uncertainties and work for future were also discussed.

节点文献中: