Forest-height inversion using repeat-pass spaceborne polInSAR data

Forest-height inversion using airborne double-antenna synthetic aperture radar (SAR) systems has been widely researched, leading to increasing accuracy. Polarimetric SAR Interferometry (PolInSAR) data from spaceborne single-antenna SAR systems, which are influenced by temporal decorrelation, have difficulty inverting forest height. Given the temporal decorrelation effect, the classical random volume over ground (RVoG) model has been proven to invert forest height with significant errors, using repeat-pass PolInSAR data. In consideration of this problem, the temporal decorrelation RVoG (TD-RVoG; based on the RVoG) model was proposed. In this study, an improved TD-RVoG model is presented, with a new temporal decorrelation function. Compared with TD-RVoG, the new model has fewer unknown parameters and can be applied in a three-stage inversion procedure. Validity of the new model is demonstrated by Advanced Land Observing Satellite/Phased Array type L-band SAR (ALOS/PALSAR) data. Results show that the improved TD-RVoG has better accuracy, with inversion error less than 1.5 m.