基于冠层光谱的不同株型冬小麦籽粒蛋白质预测模型

冬小麦冠层光谱因不同株型而异，依不同株型建立模型是提高冬小麦蛋白质预测精度的重要途径之一。该研究利用ASD2500高光谱仪对不同株型冬小麦冠层光谱进行了测定，分析了冬小麦叶片叶绿素含量在冠层垂直方向上的变化及其与籽粒品质指标和冠层光谱特征参量间的相关性。结果表明，冠层叶绿素含量垂直梯度变化因不同生育时期和不同株型而异。同等条件下，其梯度以平展型品种大于直立型品种。并且，当将两种株型品种分别考虑时，第一二叶组之间叶绿素含量的差值(DCC)与小麦籽粒部分品质参数和冠层光谱特征参量具有显著的相关性。通过DCC可以间接地建立籽粒蛋白品质和冠层光谱特征之间的相关模型。通过研究筛选出预测籽粒蛋白质含量(GPC)的最佳时期为灌浆期，最佳光谱特征参量为560 nm的反射峰深度P_Depth560。并且，建立了不同株型品种GPC的预测模型并初步通过验证。

Prediction of grain protein based on canopy spectra in wheat with different plant types

Canopy spectra varied with plant type in winter wheat. So it is an important approach to improve the prediction of wheat protein according to plant type. The canopy spectral reflectance was measured using a ground-based spectroradiometer(ASD2500). It was analyzed that the vertical distribution of leaf chlorophyll content and the correlations between Difference Chlorophyll Content(DCC) and grain quality indices, spectral parameters. The results showed that gradients of leaf chlorophyll contents at different levels varied with plant type and growth stages. Varieties with planophile leaves presented higher DCC than varieties with erectophile leaves under the same conditions. The correlation DCC and grain quality indices, spectral parameters were improved when plant types were taken into considered. The correlative model of GPC and one of hyperspectral parameters can be constructed indirectly by DCC. The best stage for GPC prediction was at grain filling stage, and the best sensitive parameter is P_Depth560. Furthermore, the models that predicted GPC using P_Depth560 at grain filling stage were constructed and validated preliminarily for two types respectively.