论文部分内容阅读
利用高光谱探测数据,对黄河三角洲盐碱土盐分含量进行定量估测,为土壤盐分含量提供一种快速、实时、准确的估测方法。以黄河三角洲为研究区,在分析盐碱土光谱曲线变化规律的基础上,进行光谱反射率一阶微分和倒数的对数与土壤盐分含量进行相关分析,筛选敏感波长,运用主成分回归分析方法,建立土壤盐分含量估测模型,并对模型进行精度验证。结果表明,一阶微分变换可以扩大样品之间的光谱特征差异,提高了相关性;通过相关分析选出的敏感波长为388 nm、515 nm、962 nm、1164 nm、1623 nm和2150 nm;建立了土壤盐分含量的主成分回归估测模型y=4.389-798x388′-632x515′-151x962′-184x1164′-138x1623′-217x2150′,经精度验证,模型(R2=0.8405,RMSE=1.152)拟合效果较好。利用主成分回归分析方法估测黄河三角洲地区的土壤盐分含量精度较高,为实现土壤盐分的定量估测提供了理论依据和技术支撑。
Based on the hyperspectral data, the salt content of saline-alkali soil in the Yellow River Delta was quantitatively evaluated to provide a rapid, real-time and accurate estimation method for soil salt content. Taking the Yellow River Delta as the research area, the logarithm of the first order differential and reciprocal of spectral reflectance was analyzed based on the analysis of the spectral curve of saline-alkali soil. The correlation analysis was conducted between the logarithm of the spectral reflectance and the soil salinity. The sensitive wavelength was screened. The soil salinity content estimation model was established and the accuracy of the model was verified. The results show that the first-order differential transformation can enlarge the difference of spectral characteristics between samples and improve the correlation; the sensitive wavelengths selected by correlation analysis are 388 nm, 515 nm, 962 nm, 1164 nm, 1623 nm and 2150 nm; The principal component regression model for estimating soil salinity y = 4.389-798x388’-632x515’-151x962’-184x1164’-138x1623’-217x2150 ’, and the model was validated by the accuracy (R2 = 0.8405, RMSE = 1.152) better. Using principal component regression analysis method to estimate the soil salt content in the Yellow River Delta is more accurate, which provides a theoretical basis and technical support for the quantitative estimation of soil salinity.