:: Volume 6, Issue 3 (12-2018) ::
jgit 2018, 6(3): 51-66 Back to browse issues page
Soil Moisture Linear Modeling by Using Decomposition and Selection of Fully Polarized SAR Features
Esmaeil Khedri, Mahdi Hasanlou *
University of Tehran
Abstract:   (1739 Views)
Soil moisture is a key variable in the hydrologic process, which is affected by the exchange of water and energy on the Earth's surface. Precise estimation of spatial and temporal variations of soil moisture is crucial for environmental studies. The Polarimetric SAR (PolSAR) images are a convenient tool for this purpose. These images also guarantee both broad coverage and suitable spatial resolution. In this study, a linear analytical model has been suggested for estimating soil moisture. This model uses data gathered by the AIRSAR sensor in 2003 in C, L, and P bands. For this purpose, with incorporation of a genetic algorithm (GA), sequential forward selection (SFS), and sequential backward selection (SBS), we examine and select appropriate features best fitted for soil moisture modeling. Also in this estimation, soil moisture measurements were compared to in-situ data. The results showed that the proposed method (linear analysis model) had a good efficiency by using GA feature selection compare to both SFS and SBS feature selection. Regarding statistical parameters for proposed method, R2 model is higher than %80 and RMSE is less than 0.027 for P, L, and C bands, which in comparison with other algorithms, the R2 model estimates soil moisture more accurately. Also, the best bands to estimate soil moisture model using proposed model and incorporated PolSAR features is the C band.
Keywords: Linear analytical model, Soil moisture, GA, SFS, SBS.
Full-Text [PDF 1206 kb]   (591 Downloads)    
Type of Study: Research | Subject: RS
Received: 2018/12/25 | Accepted: 2018/12/25 | Published: 2018/12/25

XML   Persian Abstract   Print

Rights and permissions
Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Volume 6, Issue 3 (12-2018) Back to browse issues page