学海网

硕士论文-SAR影像在湿地植被地上生物量估测中的应用，正文共62页。
【摘要】 湿地是主要的陆地生态系统之一,研究湿地的生态系统特征和规律有着重要的意义。运用合成孔径雷达(SAR)影像进行生物量估算,是生态学研究中较为新颖的课题。作为一种新型传感器,SAR具有全天时、全天候的特点,不受气候影响,对水反应敏感,对地物具有一定的穿透能力,是极其重要的湿地生态水文信息源。特别是湿地水陆交界复杂的生态环境特征,雷达影像可作为与光学影像估测生物量的补充。本文以三江平原内陆沼泽湿地——洪河国家级自然保护区为例,通过2009年生长季实测采集的17个沼泽植被样方的干湿重数据,结合同一时间段内HH/VV极化下的ASAR影像、PALSAR影像和TM影像,以SAR影像不同极化的后向散射系数、TM影像的归一化植被指数(NDVI)和实测生物量干湿重数据为相关变量,建立回归模型和水云模型的拟合方程,以其余8个样方的实测值作为检验样本,对拟合方程进行精度评价,研究SAR与TM影像估测湿地生物量的差异性。研究的主要结论有：以模型拟合相关度来看,ASAR影像VV极化下线性拟合方程相关度最高,R2达0.622,以精度评价结果及生物量分布图看,则ASAR影像HH极化下的相关结果较好。从水云模型的拟合结果及精度评价结果看,各影像的拟合相关度及残差系数相差不大,但生物量分布图与实际差异较大。ASAR影像HH极化下生物量线性估测结果优于TM影像估测结果,表明雷达遥感影像在湿地生物量估测方面与光学遥感影像具有较强的互补作用。
【Abstract】 Wetland is one of the major land ecosystems with significance to discover the characteristics and ecological laws. The estimation of wetland biomass by Synthesize Aperture Radar (SAR) is a novel subject. As a new type sensor, SAR has advantages for wetland applications such as all time, all weather, sensitive on water and definite capability through the ground. Therefore, it can be apply for generating some important information about wetland ecology and hydrology, especially with the complex ecological habitat of the transitional place between land and water, radar images can be a supplement of optical images for wetland biomass estimation.In this paper, as a case study of SAR application on Honghe National Natural Reserve (HNNR), the wet weight and dry weight field data of 17 quadrats of diverse vegetation in a growing season from the study area in 2009 were gathered and ASAR/PALSAR/TM images were acquired at the same time with field work. The backscattering coefficient of polarimetric radar, NDVI index of TM image and wet and dry weight data was taken as correlation variable, the fitting equations of regression models and water cloud model were built to estimate biomass, a validation was made from 8 measured quadrats as the test samples to evaluate the accuracy assessment and the difference between SAR and TM in wetland biomass estimation. The conclusions of the research are got as following.As the value of fitted models, the relations in linear equations of ASAR image in VV polarization are higher than other conditions, and the value of R2 is 0.622, from the accuracy assessment and the biomass estimation diagrams, the results in HH polarization of ASAR image condition approach to better.As the results of fitted equations and accuracy assessment of Water Cloud Model using SAR images, the difference between correlations and average residual coefficients is closely, but the biomass estimation diagrams varies obviously compare to original circumstances.The results of linear regression biomass estimation in HH polarization of ASAR image situation are better than TM image. It announces the efficient complementary between the SAR image and optical remote sensing image in wetland biomass estimation.
【关键词】 湿地生物量； SAR； 水云模型； 估测；
【Key words】 wetland biomass； SAR； Water Cloud Model； estimation；