Building Height Information Extracting Technique Based on PSInSARLIU

This article proposed a new model and algorithm for extracting the height of buildings which applied on high-resolution SAR images based on time-series PSInSAR from the statistical analysis of the interferometric phase. We then extract the height of typical buildings in Beichen District of Tianjin with the proposed model and 27 high-resolution TerraSAR images. Statistical results showed that: for high height of man-made objects, the height difference between the value accounted for a smaller proportion of the actual height. For smaller, low height of man-made objects, the height difference between the value for a large proportion of the actual height, the main feature is its small height corresponding to the phase component, greatly affected by noise and model solver errors. Overall, the sampling verification for some typical features of the experimental results and the actual height difference between the mean and standard deviation were 0.8m and 2.1m.The error indicated the feasibility and effectiveness of this method, when compared with result of field survey.