基于红边位置的木荷叶片叶绿素含量估测模型研究

Estimation model of schima superba leaf chlorophyll content based on red edge position

利用红边参数反演作物参数是定量遥感研究的一个热点,红边参数中红边位置与作物生化组分强相关,为监测作物胁迫提供了一个非常敏感的指标。准确估测植被叶绿素含量,对于研究森林健康和胁迫、森林生产力的估计,碳循环的研究有着重要的意义。介绍几种红边位置算法,并对这些算法及其应用进行了比较,通过选取红边位置的不同敏感波段来估测植被叶片叶绿素含量。经室内光谱获取叶片的光谱数据,采用一阶光谱导数法、平滑处理后一阶光谱导数法、线性四点内插法、五次多项式拟合法四种算法处理光谱数据,获得红边位置变量,并与叶绿素含量进行拟合,构建估测木荷叶片叶绿素含量的回归模型。结果表明:各种算法获取的红边位置变量所构建的回归模型估测叶绿素含量是可行的;五次多项式拟合法估算精度是最高的,其获取红边位置计算相对复杂;线性四点内插法估算精度次之,但计算较简便。

Red edge parameters are widely used to inversely deduce crop parameters in quantitative remote sensing studies. Among them, the red edge position, as a very sensitive indicator for monitoring crop stress, is strongly correlated with crop biochemical components. Accurate estimation of the chlorophyll content of vegetation is of importance for studies on forest health, stress, and productivity estimation, as well as carbon cycle. In this article, several algorithms of red edge position were introduced firstly, their applications were compared, and the leaf chlorophyll content of vegetation was estimated through selecting its different sensitive bands. Then leaf spectral data from indoor spectra were extracted, four algorithms were used （first-order derivative spectrometry, first-order derivative spectrometry after smoothing, four point interpolation, and quintic polynomial fitting） to process spectral data and obtain red edge position variables. Finally the obtained variables were used to fit the chlorophyll content, and various regression models of these algorithms for estimating leaf chlorophyll content were established. The results show that all these established models are feasible to estimate chlorophyll content. Among them, the quintic polynomial fitting method is most accurate, but highly complex in obtaining the red edge position while the four point linear interpolation is next to it in accuracy, but simpler.