Forest canopy height estimation using double-frequency repeat pass interferometry

Abstract

In recent years, many efforts have been made in order to assess forest stand parameters from remote sensing data, as a mean to estimate the above-ground carbon stock of forests in the context of the Kyoto protocol. Synthetic aperture radar interferometry (InSAR) techniques have gained traction in last decade as a viable technology for vegetation parameter estimation. Many works have shown that forest canopy height, which is a critical parameter for quantifying the terrestrial carbon cycle, can be estimated with InSAR. However, research is still needed to understand further the interaction of SAR signals with forest canopy and to develop an operational method for forestry applications. This work discusses the use of repeat pass interferometry with ALOS PALSAR (L band) HH polarized and COSMO Skymed (X band) HH polarized acquisitions over the Taxiarchis forest (Chalkidiki, Greece), in order to produce accurate digital elevation models (DEMs) and estimate canopy height with interferometric processing. The effect of wavelength-dependent penetration depth into the canopy is known to be strong, and could potentially lead to forest canopy height mapping using dual-wavelength SAR interferometry at X-and L-band. The method is based on scattering phase center separation at different wavelengths. It involves the generation of a terrain elevation model underneath the forest canopy from repeat-pass L-band InSAR data as well as the generation of a canopy surface elevation model from repeat pass X-band InSAR data. The terrain model is then used to remove the terrain component from the repeat pass interferometric X-band elevation model, so as to enable the forest canopy height estimation. The canopy height results were compared to a field survey with 6.9 m root mean square error (RMSE). The effects of vegetation characteristics, SAR incidence angle and view geometry, and terrain slope on the accuracy of the results have also been studied in this work.

Forest Height Estimation Approach Combining P-Band and X-Band Interferometric SAR Data

“…Based on Equation (11), it can be determined that the upper limit of DSM bias estimated based on the IDUV model is equal to one-fourth of the HoA. Obviously, this is not a reasonable limit for diverse forests.…”

“…Based on Equation (11), DSM bias can be easily calculated using the coherence obtained from X-InSAR data. For the IDUV model, the observed coherence (γ IDUV ) is calculated using the following formula:…”

“…For the IDUV model, DSM bias was estimated according to Equation (11) based on γ IDUV , and the compensated DSM based on IDUV (DSM X-IIDUV ) was obtained by adding DSM bias to DSM X-InSAR . Similarly, for the MLM, the compensated DSM based on the MLM (DSM X-MLM ) was formed by adding the DSM bias of the MLM to DSM X-InSAR .…”

“…The common disadvantage of empirical models is that external data are necessary, and, in most cases, the Several studies have verified the feasibility of estimating forest height based on dualfrequency InSAR data from two different covers. Among them, the most used dualfrequency InSAR combination is P-band and X-band [2,[7][8][9], followed by L-band and X-band [10,11]. The above research results have shown that there is a stable correlation between the height obtained by long-and short-wavelength SAR interferometric difference and the forest height, but the forest height is usually underestimated.…”