【摘要】 为了解稻麦系统作物冠层高光谱特征与其系统产量的关系,进而实时动态监测作物长势与产量情况,分析水稻和小麦系统产量与多类型高光谱特征变量之间的关系.选择相关性较好的光谱波谱与变量参数,并分别建立最佳回归预测模型.结果表明,水稻系统产量的估测以多变量回归模型为最优,红边内一阶导数的总和(Sr)与蓝边内一阶导数的总和(Sb)的归一化值(Sr-Sb)/(Sr+Sb)、Sr和红边位置(λr)3个变量对水稻系统产量有较大的影响(R2=0.739),估测精度为88.24%;小麦系统产量的估测以Sr的指数模型为最优(R2=0.780),估测精度为66.43%.红边光谱区域的高光谱变量与光谱指数用于估算稻麦系统作物长势与产量较为有效,可用于定量、准确监测稻麦系统产量.更多还原

【Abstract】 In order to understand the correlation between rice-wheat system crop’s canop hyperspectrum spectral signature and system yields,then to monitor dynamically the crop growth and yield in real time,the article analyzes the correlation between rice-wheat system yield and multi-type spectral signature variables.The spectral spectrum with preferable correlation and variable parameter were chosen and established optimum regression forecast model respectively.The result shows that the evaluation of rice system yield turns optimum as being multi-variable regression model.The normalized value(Sr-Sb)/(Sr+Sb)、Srof first derivative sum(Sr)inside red edge &sum(Sb)inside blue edge and 3variables at the location of red edge(λr)have significant influence on the yield of rice system(R2=0.739),and the estimation precision is 88.24%;the Srexponential model(R2=0.780)is optimal as the evaluation of wheat system yield,and the estimation precision turns 66.43%.The hyperspectral characteristic variables and spectral index in red-edge spectrum distribution area are quite effective for the evaluation of rice-wheat system crops’ growth and yields,which can be used for the quantitation and precise monitoring of ricewheat system yield.更多还原