传统光谱变换与连续小波耦合定量反演潮土有机质含量

以北京地区的96个潮土土样的有机质含量为研究对象，以传统光谱变换为参照，研究分析传统光谱变换与连续小波的耦合在估测土壤有机质含量的可行性；首先采用传统光谱变换与连续小波处理土壤光谱数据，然后将处理后的光谱数据与土壤有机质含量进行相关性分析，提取敏感波段，并采用偏最小二乘法构建土壤有机质含量估测模型。结果表明：耦合传统光谱变换技术与连续小波技术可大幅提升光谱对有机质含量的敏感性，其相关系数R2最高可达0.714，这表明耦合传统光谱变换技术与连续小波技术可深入挖掘光谱内的有益信息；与传统光谱变换技术相比，基于耦合传统光谱变换技术与连续小波技术构建的模型精度更高，稳定性更好，其中以微分变换构建的模型最优，其R2=0.772，RMSE=0.223，这表明耦合传统光谱变换技术与连续小波技术可有效压制噪声的负面影响，提升光谱的稳定性。

Quantitative Inversion of Organic Matter Content Based on Interconnection Traditional Spectral Transform and Continuous Wavelet Transform

In this study, the soil organic matter content of 96 alluvial soil collected from Beijing area was taken as the object of study; Compared with the traditional spectral transform technology, this paper studied on the analysis of the traditional spectral transform and continuous wavelet technology coupling in the feasibility of estimating soil organic matter content. Firstly, the traditional spectral transform technique and the continuous wavelet transform were used to deal with the soil spectral data. Then The correlation between the spectral data and the soil organic matter content was analyzed, and the sensitive bands were extracted. Finally the estimation model of soil organic matter content was constructed by partial least square method. The results showed that the coupling of traditional spectral transform and continuous wavelet technology can greatly improve the spectral sensitivity of organic matter content, and the correlation coefficient (R2) was up to 0.714, which indicate that the coupling of the traditional spectral transform and continuous wavelet technology can dig the useful signal of the spectral information; Compared with the traditional spectral transform technology, the accuracy of the model based on the interconnection of traditional technique and continuous wavelet transform was higher and better stability; Among of the model based on the interconnection of traditional technique and continuous wavelet transform, the model construct by the differential transform was the Optimal model; Its coefficient of decision and root mean square error were 0.774 and 0.223 respectively, which indicated that the interconnection of traditional technique and continuous wavelet transform spectral technique can effectively suppress noise, improving the spectral stability.