Building Height Estimation from High Resolution SAR Imagery via Model-Based Geometrical Structure Prediction

Height extraction by radar remote sensing is an attractive issue for the building detection and recognition. According to the analysis on the building geometrical properties in the SAR imagery, a novel height estimation algorithm is proposed following a model-based geometrical structure prediction and matching strategy. The range Doppler equation is introduced and simplified for the building 2D geometrical structure prediction in the slant image plane. An evaluation function implementing the ratio of exponentially weighted averages (ROEWA) is also established for the matching between the predicted structure and the observed SAR image. By incorporating the genetic algorithm (GA), the evaluation function is maximized to get the optimal height parameter. The experimental results with the simulated and real airborne and spaceborne SAR images show that the proposed method could efficiently estimate building height from single SAR imagery, and achieve better performance than two popular algorithms with the partial occlusion case.