论文部分内容阅读
高光谱技术是一种快速无损监测植被生物量的有效方法,但土壤背景的干扰一直是生物量监测的主要限制因素之一。本研究试图利用盲源分离(blind source separation,BSS)法分离出净植被光谱,达到消除土壤背景影响,提高小麦生物量估算精度的目的。本研究对110组小麦冠层光谱数据进行快速独立分量分析(fast independent component analysis,Fast ICA)处理,提取净植被光谱,并对比了Fast ICA处理前后所建的偏最小二乘回归(partial least squares regression,PLSR)模型估算精度。结果表明:Fast ICA算法可有效分离土壤光谱和植被光谱;且基于净植被光谱建立的小麦生物量估算模型精度得到明显提升,建模集RPDc(ratio of performance to deviation of the calibration)和交叉验证集RPDcv(ratio of performance to deviation of the cross calibration)分别由原始光谱的1.83和1.64提高至2.77和2.09;可见,Fast ICA可以作为有效的光谱数据预处理方法,显著提高小麦生物量的估算精度,为利用遥感技术进行大尺度、精准监测生物量提供了方法支持和理论依据。
Hyperspectral technology is an efficient and rapid method for non-destructive monitoring of vegetation biomass, but interference from soil backgrounds has been one of the major limiting factors in biomass monitoring. In this study, we attempted to separate the net vegetation spectrum by blind source separation (BSS) method to eliminate the influence of soil background and improve the estimation accuracy of wheat biomass. In this study, fast independent component analysis (Fast ICA) of 110 canopy wheat canopy spectra was used to extract the net vegetation spectra, and the partial least squares regression was compared with that of Fast ICA regression, PLSR) model estimation accuracy. The results show that the Fast ICA algorithm can effectively separate the soil spectrum from the vegetation spectrum, and the accuracy of wheat biomass estimation model based on net vegetation spectrum has been significantly improved. The model set RPDc (ratio of performance to deviation of the calibration) and cross validation set RPDcv (ratio of performance to deviation of the cross calibration) increased from 1.83 and 1.64 of the original spectrum to 2.77 and 2.09, respectively. It can be seen that Fast ICA can be used as an effective pretreatment method for spectral data to significantly improve the estimation accuracy of wheat biomass, Using remote sensing technology for large-scale, accurate monitoring of biomass provides a method to support and theoretical basis.