Translator Disclaimer
12 March 2019 Comparison of hyperspectral retrieval models for soil moisture content
Author Affiliations +
In order to improve the estimation accuracy of Soil Moisture Content (SMC) by hyperspectral technology, the paper used actual measured spectral data to study quantitative relationship between soil hyperspectral reflectance and the SMC. All total of soil samples from the Puzhehei Scenic Spot in Qiubei County, Yunnan Province in December 2016 were measured in the lab with the spectrometer. This paper used the original reflectivity of the soil samples and its four mathematical transformations as the inversion indicators to construct Unary Linear Regression Model (ULRM), Multiple Stepwise Regression Model (MSRM) and Partial Least Squares Regression Model (PLSRM) aiming to compare the performance and inversion accuracy of these three models, and find the best performance model to inverse the SMC. The results showed that: (1) It was determined that 1350nm, 1450nm, 1841nm, 1897nm, 1905nm, 1935nm and 2146nm were the hyper-spectral characteristic bands of SMC by analyzing the correlation between soil moisture content and reflectance. (2) The coefficients of determinations R2 varied between 0.73 and 0.91 and the Root Mean Square Error (RMSE) ranged from 1.51 to 1.86 with the best performance obtained with the PLSM, and the LRM had the lowest inversion accuracy. (3) The PLSRM established by the logarithm of the reflectivity of 1450nm, 1841nm, 1897nm, 1905nm, 1935nm and 2146nm was the best model of the 15 models by comparing the inversion precision of the samples in each model in this study
© (2019) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Mingli Hao and Wenying Hu "Comparison of hyperspectral retrieval models for soil moisture content", Proc. SPIE 11023, Fifth Symposium on Novel Optoelectronic Detection Technology and Application, 110232E (12 March 2019);

Back to Top