Inversion of Soil Moisture from Backscattering Coefficient Using LS-SVM

DOI

10.2991/rsete.2013.91How to use a DOI?

Keywords

Soil moisture, inversion, remote sensing, backscattering coefficient, LS-SVM

Abstract

Inversing soil moisture from remote sensing data is difficult for the problem is usually nonlinear and ill-posed. To enhance the accuracy of this inversion problem and reduce the effect of surface roughness, a least square-support machine (LS-SVM) based inversion approach is used to retrieve the soil moisture from the radar backscattering coefficients. Starting from the generation of data set by using Integral Equation Model (IEM), wide range of soil moisture and surface roughness are simulated. The kernel and capacity parameter of LS-SVM are optimized through the training process. Then, to assess the effectiveness of the proposed approach, testing data added with Gaussian distributed noise is processed by the suitably defined model. Concerning the robustness of the approach, selected training data is applied when the model is established, and the soil moisture is inversed again. Along this process, the comparison between BP neural network and LS-SVM based method is conducted.

Copyright

© 2013, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Cheng Zhihui
AU  - Zhang Qinhe
AU  - Li Jianing
AU  - Lu Wei
PY  - 2013/08
DA  - 2013/08
TI  - Inversion of Soil Moisture from Backscattering Coefficient Using LS-SVM
BT  - Proceedings of the 2013 the International Conference on Remote Sensing, Environment and Transportation Engineering (RSETE 2013)
PB  - Atlantis Press
SP  - 372
EP  - 375
SN  - 1951-6851
UR  - https://doi.org/10.2991/rsete.2013.91
DO  - 10.2991/rsete.2013.91
ID  - Zhihui2013/08
ER  -