Please use this identifier to cite or link to this item:
Title: PolInSAR based modelling for scattering characterization and forest parameter retrieval
Other Titles: 38th Asian Conference on Remote Sensing - Space Applications: Touching Human Lives, ACRS 2017
Authors: Kumar S. 
Joshi S.K. 
Tomar K.S. 
Aggarwal N. 
Khati U.G. 
Chandola S. 
Bharadwaj S.P. 
Agarwal S. 
Kushwaha S.P.S. 
Issue Date: 2017
Publisher: Asian Association on Remote Sensing
Abstract: SAR remote sensing has already proven its capability to retrieve forest structural and biophysical parameters. Prime focus of the present study was to evaluate the potential of PolSAR, PolInSAR and PolInSAR tomography (PolTomSAR) for forest structural and biophysical parameter retrieval. This work includes the utilization of ALOS PALSAR, RADARSAT-2, RISAT-1 and TerraSAR-X data for scattering characterization and coherence estimation of different locations of forest. Tree height retrieval was performed with the model inversion using PolInSAR data and signal compression techniques of SAR tomography. Potential of PolTomSAR was evaluated to retrieve forest height and variation in backscatter power at different height levels. Fourier transform, beamformer and capon algorithms were compared for vertical profile of forest patch. Fully polarimetric capon showed best forest height result with RMSE of 2.58 m and an average accuracy of 88.64%. PolInSAR RVoG modelling based three stage inversion (TSI) and coherence amplitude inversion (CAI) techniques were implemented on PolInSAR data to generate forest height map. PolInSAR data was also used for AGB retrieval of forest with the help of coherence based Interferometric Water Cloud Model (IWCM). PolSAR based EWCM model was developed for L-band ALOS PALSAR data for Dudhwa National Park, India and the modelled output for AGB estimation showed 0.43 R2 and 119 (t/ha) RMSE. IWCM based modelling for forest AGB retrieval showed R2 value of 0.4, RMSE of 62.73 (t/ha) and a percent accuracy of 51%. TSI based PolInSAR inversion modelling showed the most accurate result for forest height estimation. The correlation between the field measured forest height and the estimated tree height using TSI technique is 62% with an average accuracy of 91.56% and RMSE of 2.28m. The obtained results showed that PolSAR and PolInSAR remote sensing based modelling approach have capabilities to provide structural and biophysical parameters of forest with reliable accuracy. © 2017 ACRS. All rights reserved.
Appears in Collections:Conference proceedings

Show full item record

Page view(s)

checked on Jul 25, 2020

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.