Search

The random volume over ground (RVoG) model has been widely applied to estimate forest tree height from polarimetric synthetic aperture radar (SAR) interferometry (PolInSAR) data for the past two decades. Successful application of the RVoG model requires certain assumptions to be valid for the imaged forest and the acquisition scenarios in order to avoid large errors in height estimates. Quantification of errors and uncertainties of RVoG-estimated heights have typically been limited to comparison against external validation data, such as lidar or field measurements. In this paper, we present a straightforward approach to simultaneously estimate height and height uncertainty from PolInSAR data using a Bayesian framework that accounts for errors in the data, as well as errors due to incorrect RVoG modeling assumptions, such as those caused by temporal decorrelation effects and errors in ground phase estimation. We apply our method to synthetic data to study how forest height uncertainty depends on modeling assumptions and PolInSAR acquisition parameters. We also compare our estimated Bayesian uncertainties to PolInSAR-derived and lidar-derived height RMS deviations observed over Gabonese tropical forests during the joint NASA-ESA 2016 AfriSAR campaign. Our results show good correspondence between uncertainties and deviations, as well as a strong correlation between uncertainty and estimated tree height. Furthermore, we demonstrate that we can associate specific areas of high uncertainty to confounding effects, such as temporal decorrelation and noncanopy related scattering. [ABSTRACT FROM AUTHOR]

This paper presents a new damage-mapping algorithm based on coherence images estimated from multitemporal polarimetric-interferometric synthetic aperture radar (SAR) data. The interferometric coherence has been restricted in the conventional damage-mapping approaches because the decorrelation sources are too complicated to interpret accurately and temporal decorrelation effects caused by slowly occurring natural changes and disaster events are often coupled together. To overcome these limitations, we formulate a coherence model that accounts for temporal decorrelation in two simplified layers, ground and volume layers, for long-temporal repeat-pass scenarios with zero spatial baseline. The model parameters include: 1) ground-to-volume ratio, a factor to determine the relative scattering contribution of ground and volume layers; 2) temporally correlated change, which captures the exponentially decaying behavior of coherence with time; and 3) temporally uncorrelated change, which is associated with random temporal changes. We estimate the model parameters in three steps: coherence optimization, interferometric pair-invariant parameter estimation, and interferometric pair-variant parameter estimation. To isolate the effects of disaster events from background natural changes, we calculate the probability density functions of historical change pixel by pixel and produce a probability map of damage. We tested the algorithm with uninhabited aerial vehicle data acquired from 2009 to 2015 for mapping the area damaged by the 2015 Lake Fire in California. Based on performance evaluation using receiver operating characteristic curves for optimized coherences and averaged probability maps, the proposed method reduced the false alarm from 0.25 to 0.07 when the probability of detection was 0.85 compared to coherence products alone. [ABSTRACT FROM AUTHOR]