Open AccessComparative Analysis for Grey Relation Estimation Models of Soil Organic Matter based on Hyperspectral Data
Author(s) -
Xinhao Li,
Jiangong Li
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/820/1/012002
Subject(s) - hyperspectral imaging , soil organic matter , organic matter , mathematics , logarithm , mean squared error , relation (database) , soil science , soil test , correlation coefficient , environmental science , statistics , soil water , remote sensing , computer science , geography , chemistry , data mining , mathematical analysis , organic chemistry
Rapid and accurate acquisition of farmland soil organic matter content is of great significance for the rapid monitoring soil fertility and the development of precision agriculture. Based on the hyperspectral data and organic matter content data of 74 soil samples in Zhangqiu District, Jinan City, Shandong Province, first transform soil spectral reflectance by the logarithm of the first-order differential, square, reciprocal, logarithm and square root, the estimation factors were selected according to the principle of maximum relation. Second, the hyperspectral estimation of soil organic matter was carried out by using six models such as grey close relation degree, and the estimation results of different methods were combined and analyzed. The results show that the combined grey relation model can effectively improve the estimation accuracy, and the average relative error of 16 test samples is 8.931%. Studies have shown that using the combined model of grey relation degree to estimate soil organic matter content with hyperspectral data is valid.