Soil Moisture Estimation With SVR and Data Augmentation Based on Alpha Approximation Method
- 26 November 2019
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 58 (5) , 3190-3201
Abstract
Soil moisture content is an important parameter in hydrological, meteorological, and agricultural applications. Balenzano et al. proposed the alpha approximation method in 2011 for solving some complex issues during the retrieval of soil moisture over agricultural crops with synthetic aperture radar data. However, determining the constraints and solving the underdetermined system of equations in this method add new challenges. Considering the questions of constraints and underdetermined system of equations, the alpha approximation method is used to augment the measured data, and can avoid solving the underdetermined system of equations with constraints directly. Then, these data are applied in a support vector regression machine for soil moisture estimation. It is found that when an optimal model is determined, the method proposed in this article is superior to the direct use of the alpha approximation method, and the root-mean-squared error (RMSE) decreased from 0.0775 to 0.0339 and R^2 increased from 0.0467 to 0.6491. In addition, the method obtained a good result from a data set collected that included a different growing period of crops by changing the standardized method from StandardScaler to Scale, where the RMSE is 0.0501 and R^2 is 0.3204. This indicates the good generalization capability of this method. In conclusion, the proposed method solves the two questions effectively and provides a potential way for long-time or large-scale soil moisture monitoring with much less in situ measurements.Keywords
Funding Information
- National Key Technologies Research and Development Program of China (2016YFD0300603)
- National Natural Science Foundation of China (41771371)
This publication has 36 references indexed in Scilit:
- GMES Sentinel-1 missionRemote Sensing of Environment, 2012
- Use of Soil Moisture Variability in Artificial Neural Network Retrieval of Soil MoistureRemote Sensing, 2009
- On the Soil Roughness Parameterization Problem in Soil Moisture Retrieval of Bare Surfaces from Synthetic Aperture RadarSensors, 2008
- SOIL MOISTURE PREDICTION USING SUPPORT VECTOR MACHINESJawra Journal of the American Water Resources Association, 2006
- Understanding c-band radar backscatter from wheat canopy using a multiple-scattering coherent modelIEEE Transactions on Geoscience and Remote Sensing, 2003
- Soil moisture retrieval from space: the Soil Moisture and Ocean Salinity (SMOS) missionIEEE Transactions on Geoscience and Remote Sensing, 2001
- Interpreting ERS SAR signatures of agricultural crops in Flevoland, 1993-1996IEEE Transactions on Geoscience and Remote Sensing, 2000
- Rice crop mapping and monitoring using ERS-1 data based on experiment and modeling resultsIEEE Transactions on Geoscience and Remote Sensing, 1997
- A review of target decomposition theorems in radar polarimetryIEEE Transactions on Geoscience and Remote Sensing, 1996
- Support-vector networksMachine Learning, 1995