基于高光谱的甜菜SPAD值估算研究

Models of Estimating Sugar Beet SPAD Using Hyperspectral

叶绿素作为植物体内参与光合作用的重要色素,其含量对作物生长状况、产量和品质有很大影响。为此,利用野外便携式ASD光谱仪,实测了田间甜菜冠层光谱数据,且用SPAD-502叶绿素仪测定叶片SPAD值。基于原始光谱和一阶导数光谱与SPAD值相关性,选取植被指数和波段深度信息建立SPAD值预测模型,并用对照田试验数据对模型进行验证。通过对比植被指数建立的回归模型及波段深度分析,结合多元逐步回归建立的估算模型可知,波段深度比(BDR)结合SMLR建立的估算模型验证结果最好(RMSE=2.54,RE=4.5%)。研究结果表明:导数处理能提高光谱数据与SPAD值相关系数,波段深度信息结合多元逐步回归相比植被指数能提高SPAD值估算精度。

Chlorophyll plays an important role in photosynthesis processing in the plants,the content of which has an effect on the crop conditions,yield and qualities. In this study,we have used a portable ASD spectrometer measured the canopy spectral data of sugar beet,and we also measured SPAD value through SPAD- 502 chlorophyll meter. We selected vegetation index and band information to establish the depth SPAD value prediction model base on the correlation between original spectrum and the first derivative of the SPAD values,and then we used the field test data to validate this model. By comparing the regression model established by the vegetation index and the estimation model established by the depth of the band combines multiple stepwise regression analysis,the result of band depth ratio( BDR) combined estimation model validation SMLR is much better( RMSE = 2. 54,RE = 4. 5%). The results showed that the derivative of the spectral data processing can improve the value of the correlation coefficient with the SPAD,and the band combines depth information multivariate regression,compared to vegetation index estimation model,can improve the accuracy of SPAD value better.