An Algorithm Measuring Urban Building Heights by Combining the PS-InSAR Technique and Two-Stage Programming Approach Framework

Urban building heights offer critical information in studying urbanization. An approach with adequate accuracy is of significant interest. Spaceborne interferometric synthetic aperture radar (InSAR) techniques and multibaseline (MB) InSAR observations can be an alternative to fill the interest. Since the MB InSAR observations are achieved through multitemporal SAR observations with different baselines, temporal decorrelation is unavoidable, affecting most existing MB-based phase unwrapping (PU) approaches. Also, the quality of the derived heights depends on a PU approach. Thus, by combing the permanent scatterers (PS) InSAR technique and the two-stage programming approach (TSPA) framework, we proposed an algorithm to estimate urban building heights with the multitemporal SAR observations, calling it the PS-TSPA algorithm. It takes advantage of the global phase gradient information used in the single baseline PU approaches, not considering the phase continuity assumption. The algorithm should be robust to the temporal decorrelation or noise effect. With the time series of 17 TerraSAR-X data, we estimated building heights in the southern urban area of Chengdu, China. The derived building heights were satisfactory overall and assessed by in situ measurements. In comparison, the Tomo-PS method estimated the heights in the same area and with the same SAR data. Quantitative analyses indicate that the proposed algorithm outperforms the Tomo-PS method. Finally, with routine SAR observations and SAR data downloadable through the Internet, the PS-TSPA algorithm can be inexpensive in estimating urban building heights with reasonable accuracy.