In order to overcome the uncertainty in hyperspectral estimation,we establishes a hyperspectral grey correlated estimation model of soil organic matter content based on grey information theory.Based on 76 samples in Zhangqiu District,Jinan City,the spectral data are first transformed by mathematical methods such as logarithmic reciprocal and reciprocal logarithmic first-order differentiation,the correlation coefficient is calculated,and the estimation factors are selected by using the principle of maximum correlation.Then,according to the principle of increasing information and taking large method,the spectral estimation factors of each sample are sorted from small to large,and the grey information sequences are formed,and the grey information relational estimation model of soil organic matter content is constructed based on the information chain.Finally,the estimation results based on different information chains are fused twice,and compared with the commonly used estimation methods.The results show that the average relative error of 12 test samples is 5.576%,and the determination coefficient R2 is 0.934,and higher than that of the common methods such as the multiple linear regression,BP neural network and support vector machine and so on.The results show the grey correlated model based on grey information proposed is feasible and effective,and provides a new way for hyperspectral estimation of soil trait indicators.
Soil organic matterHyper-spectral remote sensingGrey information relationalEstimation model
国家科技期刊平台
NSTL
万方数据
