Urban Area Extraction Using Optimal Roll-invariant Features and Multi-Aperture Polarimetric Entropy

Inhalt:
Since the variability of urban structures, buildings with substantial cross-polarized scattering may be misclassified as forests. Therefore, urban area extraction is still a challenging problem. In this paper, a new urban extraction method using optimal roll-invariant features and multi-aperture polarimetric entropy is proposed. First, the optimal ratio of correlation coefficient and selected hidden features are used to extract the urban areas and the fusion of correlated probabilities (FCP) algorithm is applied to fuse the urban area candidates. Then, multi-aperture polarimetric entropy is introduced to modify the H/α/A classification method and the classification result is utilized as a branch condition for subsequent extraction processing. Finally, according to the branch condition, G4U decomposition results are applied to commendably enhance the extraction accuracy. Spaceborne Gaofen-3 full PolSAR data is used to evaluate the performance of the proposed method. Experimental results demonstrate that the proposed method is capable of extracting urban areas.