利用高光谱数据估测不同温度胁迫下的水稻籽粒中粗蛋白和直链淀粉含量

Estimating Contents of Crude Protein and Amylose Content in Rice Grain by Hyper-spectral under Different High Temperature Stress

水稻籽粒粗蛋白和直链淀粉含量是评价稻米品质的两个主要指标。本文以不同温度胁迫下的水稻田间试验为基础,分析水稻成熟籽粒粗蛋白和直链淀粉含量与不同生育期冠层反射光谱的相关性。结果显示,水稻籽粒的粗蛋白和直链淀粉含量与不同生育期冠层原始光谱和一阶导数在某些波段达到极显著或显著相关水平。进一步分析比值植被指数、差值植被指数、归一化植被指数和红边参数等光谱参数与成熟籽粒粗蛋白和直链淀粉含量的相关性,用回归分析方法对相关拟合较好的参数进行筛选,建立了水稻成熟籽粒粗蛋白含量(GCPC)和直链淀粉含量(GAC)监测模型,运用独立观测数据对模型进行检验,其预测值与实测值的精确度为0.393-0.683,准确度为0.708-0.923,RMSE值为8.706%-11.296%。表明模型预测值与实测值之间具有较高的符合度,对水稻籽粒粗蛋白质和直链淀粉含量具有较好的预测性。

Grain crude protein and amylose content are two important indexes for evaluating rice quality. The relationships between grain crude protein and amylose content in rice and canopy spectral reflectance at different growth stages were analyzed on the basis of the data from the field experiments under different temperature. The results showed that there were significantly increased grain crude protein content and reduced amylose content under high-temperature, and that there was also significantly or very significantly correlation of grain crude protein and amylose content with canopy spectra and its first derivative at different growth stages in some wavebands. The correlation of grain crude protein and amylose content with the ratio, differential and normalized difference vegetation indices as well as red edge parameters were further analyzed. Those fitting spectral parameters with better correlation were subsequently selected by regression analysis, and the model was established for predicting grain crude protein content （GCPC） and grain amylose content （GAC） in rice. The established model was tested with observed data from other independent experiments with estimation precision of 0.393-0.683, estimation accuracy of 0.708-0.923, and RMSE of 8.706%-11.296%, which indicated a good fit between predicted values in this model and observed values. It was concluded that the established spectral index model can be used for predicting grain crude protein and amylose content in rice with different temperature levels.