浒苔水体光谱特征参量时间序列数据拟合研究

开展时间序列浒苔水体光谱特征参量的拟合研究可为寻求最佳拟合方法,从而对进一步开展大尺度的高光谱高空间高时间分辨率遥感动态实时监测具有重要的参考价值。在实验室培养3种不同浓度的浒苔水体,用ASD光谱仪测定并获取浒苔水体的时间序列反射光谱,基于反射光谱定量提取光谱特征参量,构建时间序列数据;应用曲线拟合和时间序列ARMA模型开展浒苔水体光谱特征参量的时间序列拟合研究。结果表明:应用二阶或三阶多项式和ARMA模型可较高精度地拟合浒苔水体的光谱特征参量;但时间序列ARMA模型比多项式拟合模型更能反映浒苔水体光谱特征参量的时间序列数据的分布和变动特性。该研究可为浒苔灾害动态监测业务化运行部门提供浒苔光谱数据,同时也为光谱数据的处理和分析提供一种新的方法。

Fitting study of time series spectral variables of Enteromorpha prolifera water body

It is of significance to seek for the optimum fitting method and further dynamically monitor Enteromorpha prolifera using hyperspectral, high spatial resolution and high temporal resolution remotely sensed data. Three different kinds of densities for E. prolifera water body in the lab were cultivated. Reflectance spectra of E. prolifera a water body were measured by ASD spectrometer, spectral variables based on reflectance spectra were extracted and the time series data were constructed. Curve estimation and ARMA (Autoregressive and Moving Average) model were applied to simulate reflectance spectra. Research results demonstrated that spectral variables of E. prolifera were highly simulated by second-order and third-order polynomial function and ARMA model. Moreover, ARMA model can affect the distribution and fluctuation features of reflectance spectra of E. prolifera more significantly than polynomial function. The present study will provide dynamical monitoring business operation department of E. prolifera hazard with reflectance spectra data of E. prolifera and a new technique for data processing and analysis.