苹果叶片磷含量高光谱估测模型研究

Study on hyperspectral estimation model of phosphorus content on apple leaves

为了监测苹果树生长阶段的营养情况,本研究采用山东省肥城市潮泉镇下寨村两个果园基地获取的2012、2013年苹果树整个生长期的叶片磷含量数据和相应的叶片光谱反射率数据,建立苹果叶片磷含量估测模型.通过进行叶片磷含量和光谱反射率及其一阶导数之间的相关性分析,筛选出叶片磷含量的敏感波段分别为530 nm、553 nm、558 nm、704 nm、722 nm、732 nm.分别采用多元线性回归和逐步回归,构建苹果叶片磷含量估测模型.结果表明:基于逐步回归构建的模型,具备较高的拟合度,其建模的R2和RMSE分别为0.88、0.046 g/100 g,其验证的R2和RMSE分别为0.43、0.04 g/100 g,该模型可以高精度地估测苹果叶片磷含量.研究结果可以高效率监测苹果生长期的营养状况,从而为进行精准施肥提供了理论依据.

In order to monitor the nutritional status of apple trees at the growth stage,this study adopted the 2012 and 2013 apple tree leaf phosphorus content data and corresponding leaf spectral reflectance data obtained from two orchard bases in Xiɑzhɑi Village,Chɑoquɑn Town,Feichenɡ City,Shɑndonɡ Province,to establish the apple leaf phosphorus content estimation model.Through correlation analysis between leaf phosphorus content,spectral reflectance an-d its first derivative,sensitive bands of leaf phosphorus content were selected as 530 nm,553 nm,558 nm,704 nm,722 nm and 732 nm,respectively.Multiple linear regression and stepwise regression were used to establish phosphorus content estimation models of apple leaves.The results showed that the model constructed based on stepwise regression had a high degree of fitting.The R2 and RMSE of the model were 0.88 and 0.046 g/100 g,respectively,and the R2 and RMSE verified were 0.43 and 0.04 g/100 g,respectively.The model could accurately estimate the phosphorus content in apple leaves.The results of this study can effectively monitor the nutritional status of apple during its growing period,thus providing a theor-etical basis for precise fertilization.