Share Email Print

Proceedings Paper

Simulation of SAR backscatter for forest vegetation
Author(s): Richa Prajapati; Shashi Kumar; Shefali Agrawal
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Synthetic Aperture Radar (SAR) is one of the most recent imaging technology to study the forest parameters. The invincible characteristics of microwave acquisition in cloudy regions and night imaging makes it a powerful tool to study dense forest regions. A coherent combination of radar polarimetry and interferometry (PolInSAR) enhances the accuracy of retrieved biophysical parameters. This paper attempts to address the issue of estimation of forest structural information caused due to instability of radar platforms through simulation of SAR image. The Terai Central Forest region situated at Haldwani area in Uttarakhand state of India was chosen as the study area. The system characteristics of PolInSAR dataset of Radarsat-2 SAR sensor was used for simulation process. Geometric and system specifications like platform altitude, center frequency, mean incidence angle, azimuth and range resolution were taken from metadata. From the field data it was observed that average tree height and forest stand density were 25 m and 300 stems/ha respectively. The obtained simulated results were compared with the sensor acquired master and slave intensity images. It was analyzed that for co-polarized horizontal component (HH), the mean values of simulated and real master image had a difference of 0.3645 with standard deviation of 0.63. Cross-polarized (HV) channel showed better results with mean difference of 0.06 and standard deviation of 0.1 while co-polarized vertical component (VV) did not show similar values. In case of HV polarization, mean variation between simulated and real slave images was found to be the least. Since cross-polarized channel is more sensitive to vegetation feature therefore better simulated results were obtained for this channel. Further the simulated images were processed using PolInSAR inversion modelling approach using three different techniques DEM differencing, Coherence Amplitude Inversion and Random Volume over Ground Inversion. DEM differencing technique calculates tree height by generating Digital Elevation Models (DEM) from interferograms in different polarizations and differences in DEM estimates the vegetation height. In CAI technique the phase of coherence is ignored and volume scattering is mainly considered for estimating height. The RVoG model considers both vegetation layer and ground interactions. In this model, the vertical distribution of scatterers do not change with the change in polarization. It was found that with vertical wavenumber values between 0.2113 to .2249 rad/m for mean incidence angle 34.226 degrees the range of tree height achieved by Coherence Amplitude Inversion and RVoG was better among the three inversion techniques.

Paper Details

Date Published: 2 May 2016
PDF: 12 pages
Proc. SPIE 9881, Earth Observing Missions and Sensors: Development, Implementation, and Characterization IV, 98811T (2 May 2016); doi: 10.1117/12.2224036
Show Author Affiliations
Richa Prajapati, Indian Institute of Remote Sensing (India)
Shashi Kumar, Indian Institute of Remote Sensing (India)
Shefali Agrawal, Indian Institute of Remote Sensing (India)

Published in SPIE Proceedings Vol. 9881:
Earth Observing Missions and Sensors: Development, Implementation, and Characterization IV
Xiaoxiong J. Xiong; Saji Abraham Kuriakose; Toshiyoshi Kimura, Editor(s)

© SPIE. Terms of Use
Back to Top