DTM Extraction Beneath Forest Canopy at L-Band from an Experimental Single-Pass Airborne Pol-InSAR System
Bryan Mercer(1), Qiaoping Zhang(1), Michael Denbina(1), Marcus Schwaebisch(2) and Shane Cloude(3)
(1) Intermap Technologies Corp., #1200, 555 - 4th Ave. SW, Calgary T2P 3E7, Canada
(2) Intermap Technologies GmbH, Heimeranstrasse 35, D-80339 Munich, Germany
(3) AEL Consultants, 26 Westfield Avenue, KY15 5AA Cupar, Fife, United Kingdom
Polarimetric interferometry (Pol-InSAR) has shown considerable promise in recent years in a variety of applications. One of the applications of interest is the extraction of bare-earth DEMs beneath forest canopy as well as the canopy height itself. The use of L-Band data is of particular relevance since it has reasonable forest penetration capability while being less restricted by bandwidth licensing restriction than longer wavelengths. Thus it is being considered for satellite platforms of the future. However the interferometric phase response recorded from typical forest environments is a mixture of ground and forest contributions. The potential of Pol-InSAR, together with an appropriate model such as the Random Volume Over Ground (RVOG) model, has demonstrated significant success for the recovery of canopy height and, to a lesser extent, ground elevation beneath canopy, as has been documented in the literature. However, the L-Band experience to date has been obtained with repeat-pass Pol-InSAR, so that temporal de-correlation as well as residual motion remains an issue. For this reason, an experimental L-Band, fully polarimetric, single-pass, airborne InSAR system has been developed to assess performance in forest conditions in the absence of temporal de-correlation and residual motion issues. In this paper, we show results from recent tests using this newly-developed single-pass Pol-InSAR system. A series of tests were carried out over the period December 2007 – June 2008 in western Canada. In the early part of the series, only a portion of the swath was available due to multi-path problems. After multi-path mitigation efforts, the test areas were revisited and part of these results are reported here. The particular test results presented are from a pine-forested area near Edson, Alberta with stand heights ranging to 30 meters interspersed with clear-cut areas. Ancillary data for analysis purposes included high resolution color air-photos plus lidar ground and feature data. The RVOG model was applied to the calibrated Pol-InSAR data using optimization techniques including both phase and magnitude diversity in order to extract the parameters of interest. Results presented in this paper include the accuracies achieved for the bare-earth DEMs and forest heights extracted as a function of forest height and other parameters. Also presented are some interesting kv-dependant results as well as apparent winter-summer dependencies.