论文部分内容阅读
土壤含水量的高光谱反演是当今研究的热点,本文以土壤多样化的陕西横山县为研究区,通过野外采集土壤样品,室内利用ASD Field Spec FR地物光谱仪测定土壤样品光谱,采用称重法计算出土壤样品含水量,并分析了不同含水量土壤样品的光谱特性。针对土壤含水量光谱反演中光谱反演因子的构建问题,在研究FD-PCA、WPT-FD-PCA反演输入因子生成方法及存在的不足的基础上,提出了基于谐波分析的WPT-FD-HA-PCA的反演输入因子构建方法。以上述3种反演输入因子为基础,建立了土壤含水量反演的FD-PCA-BP、WPT-FD-PCA-BP、WPT-FD-HA-PCA-BP三种BP反演模型。通过比较土壤含水量实测值与三种反演输入因子的反演结果,得出WPT-FD-HA-PCA-BP模型的反演精度最高,决定性系数R2达到0.9599,均方根误差(RMSE)为1.667%,其反演结果明显优于其他两种模型,说明:通过小波包变换和谐波分析能有效的抑制光谱噪声并压缩信号,在一定程度上明显提高了土壤含水量反演精度。
Hyperspectral inversion of soil water content is the hotspot in this field. In this paper, Hengshan County, Shaanxi Province, whose soil is diversified, was used as the study area to collect soil samples from the field. The spectra of soil samples were determined by ASD Field Spec FR spectrophotometer. Method to calculate the water content of soil samples and analyze the spectral characteristics of soil samples with different water contents. Aiming at the construction of spectral inversion factor in spectral inversion of soil moisture content, based on the research on the generation method of FD-PCA and WPT-FD-PCA inversion input factor and its shortcomings, this paper proposed a harmonic analysis based WPT- Inverse Input Method for FD-HA-PCA Construction. Based on the above three inversion input factors, three BP inversion models of FD-PCA-BP, WPT-FD-PCA-BP and WPT-FD-HA-PCA-BP were established. The results show that the inversion accuracy of WPT-FD-HA-PCA-BP model is the highest, the determinant coefficient R2 is 0.9599, the root mean square error (RMSE) Which is 1.667%. The inversion result is obviously better than the other two models. It shows that the wavelet transform and harmonic analysis can effectively suppress the spectral noise and compress the signal, which can obviously improve the accuracy of soil moisture retrieval to a certain extent.