Open AccessHyperspectral parameters and prediction model of soil moisture in apple orchards
Author(s) -
Wenqian Wang,
Mingyang Gao,
Jiafan Wang
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.179
H-Index - 26
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/687/1/012085
Subject(s) - mean squared error , water content , hyperspectral imaging , environmental science , coefficient of determination , soil science , orchard , spectral index , moisture , correlation coefficient , soil water , adaptability , remote sensing , mathematics , statistics , geology , meteorology , agronomy , geography , spectral line , geotechnical engineering , ecology , physics , astronomy , biology
In this study, the apple orchard of Shuangquan town in mountainous and hilly area of central Shandong Province was taken as the research area. The relationship between the ratio spectral index (RSI), difference spectral index (DI) and normalized difference spectral index (NDSI) and soil moisture content in the range of 400nm-2450nm was explored, and then a quantitative estimation model of soil moisture content was build. The results show that the correlation between soil moisture content and spectral reflectance can be improved by different spectral index calculation. The sensitive modeling areas determined by RSI, DI and NDSI are all near the third water absorption peak, among which RSI (R 2304 , R 2081 ), RSI (R 2306 , R 2080 ), RSI (R 2306 , R 2081 ), RSI (R2307, R2079), RSI (R2307, R2080), RSI (R2307, R2081), RSI (R2308, R2080) and RSI (R2308, R2081) the PLSR regression equation constructed for independent variables has the best estimation effect, the largest determination coefficient (R 2 = 0.7918), and the smaller Root Mean Square Error (RMSE = 0.0054). The selection of full band spectral index can better extract effective information, and significantly improve the accuracy and adaptability of soil moisture spectral estimation model.