论文部分内容阅读
利用统计分析方法选取了土壤N、P、K元素含量近似而有机质含量差异较大的样本60个,通过高光谱探测分析获得样本反射率对数的一阶导数光谱,采用Bior1.3函数进行多层离散小波分解,剔除低频近似信号和高频噪声信号,得到反映土壤理化参数的特征光谱曲线;采用相关分析筛选土壤有机质含量的显著相关波段,基于显著相关波段和特征光谱曲线分别构建土壤有机质含量高光谱多元回归估测模型;通过比较分析,确定了提取土壤有机质特征光谱的最佳小波分解尺度并构建了最佳预测模型.结果表明:提取土壤有机质特征光谱的最佳小波分解层数是9层,其次是8层和10层;基于小波9层分解特征光谱曲线的有机质含量估测模型最佳,其决定系数(R2)为0.89,比基于显著相关波段构建模型的R2增加了0.31,比基于原始光谱所构建模型的R2增加了0.10.
Sixty samples with similar content of N, P and K in the soil but different content of organic matter were selected by statistical analysis method. The first order derivative spectra of logarithm of sample reflectance were obtained by using hyperspectral detection and analysis. The Bior1.3 function Layer discrete wavelet decomposition, excluding the low-frequency approximation signal and the high-frequency noise signal to obtain the characteristic spectral curve reflecting the soil physical and chemical parameters; using correlation analysis to screen the significant correlation bands of soil organic matter content; building the soil organic matter content based on the significant correlation band and the characteristic spectral curve Hyperspectral multivariate regression estimation model was established.The optimal wavelet decomposition scale of soil organic characteristic spectrum was determined and the best prediction model was established.The results showed that the optimal wavelet decomposition level of soil organic matter was 9 Layer, followed by 8 layers and 10 layers. The organic matter content estimation model based on wavelet decomposition of 9-layer decomposition characteristic curve is the best with the determination coefficient (R2) of 0.89, which is 0.31 higher than R2 based on the significant correlation band construction model R2 based on the model built with the original spectrum increased by 0.10.