摘要
随着全球气候变暖,我国洪涝灾害发生的频率及影响范围都不断增加。通过野外模拟试验,研究植被(玉米、甜菜)在水浸胁迫下的光谱变化特征,以构建高光谱遥感模型对水涝灾害范围进行监测。试验于2008年5月—8月在英国诺丁汉大学Sutton Bonington校区(52.8°N,1.2°W)进行,每周采集一次样本并在室内测量其光谱数据。试验结果表明植被光谱在550,800~1 300nm区域反射率都稍有降低,而在680nm区域反射率则略微增大。选取NDVI,SIPI,PRI,SRPI,GNDVI及R800*R550/R680共六个植被指数识别水浸胁迫下的植被,研究表明,指数SIPI与R800*R550/R680对水浸胁迫玉米比较敏感,而指数SIPI,PRI及R800*R550/R680对水浸胁迫甜菜比较敏感。为寻找最优的识别模型,计算对照与水浸胁迫植被指数之间的归一化均值距离并进行对比分析,发现植被指数R800*R550/R680的归一化均值距离在胁迫早期即大于其他指数的距离,说明该指数识别水浸胁迫植被的能力优于其他指数,且具有较强的敏感性与稳健性。因此,可以利用该指数快速地提取水浸面积,为救灾减灾决策提供信息支持。
With the global climate warming, flooding disasters frequently occurred and its influence scope constantly increased in China. The objective of the present paper was to study the leaf spectral features of vegetation （maize and beetroot） under water logging stress and design a hyperspectral remote sensing model to monitor the flooding disasters through a field simulated experi ment. The experiment was carried out in the Sutton Bonington Campus of University of Nottingham（52. 8°N, 1.2°W） from May to August in 2008, and samples were collected one time every week and spectra were measured in the laboratory. The result showed that the reflectance of the maize and beetroot decreased in the 550 and 800～1 300 nm region, and the reflectance slightly increased in the 680 nm region. This paper chose NDVI, SIPI, PRI, SRPI, GNDVI and Rs00 ＊ R550/R680 to identify the vegeta tion under waterlogging stress, respectively. The result suggested that the SIPI and Rs00 ＊ R550/R680 was sensitive for maize un der waterlogging stress, and then SIPI and PRI and R800 ＊ Rsso/R68o was sensitive for beetroot under waterlogging stress. In or der to seek the best identifiable model, the normalized distances between means of control and stressed vegetation indices were calculated and analyzed, the result indicated that the distance of R800 ~ R^so/R^8~ is more than that of indices‘ in the early stress stage, illustrated that the index identifiable ability for waterlogging stress is better than other indices, then the index has the strong sensitivity and stability. Therefore, the index R800 ＊ R550/R680 could be used to quickly extract flooding disaster area by using hyperspectral remote sensing, and would provide information support for disaster relief decisions.
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2013年第11期3106-3110,共5页
Spectroscopy and Spectral Analysis
基金
国家自然科学基金项目(41101397)
教育部博士点基金项目(20100023120007)
国家留学基金委联合资助
关键词
光谱特征
水浸胁迫
植被
识别模型对比分析
归一化均值距离
Spectral features
Waterlogging stress
Vegetatiom Identification model
Normalized distance between means
