A polarization-power-maximum-based 3D imaging method for Ku-band UAV-borne fully-polarimetric array InSAR

Fully-polarimetric array InSAR (Pol-array-InSAR) can obtain three-dimensional position information and the polarization scattering characteristics of targets. In this paper, we propose a polarization analysis-based processing framework for the synthetic aperture radar (SAR) three-dimensional (3D) reconstruction of complex urban areas. The framework employs multi-baseline coherence optimization and polarization enhancement techniques to project Pol-array-InSAR images into an optimal space. The projected images are then used for the third-dimensional tomographic inversion. After verifying the effectiveness of these techniques in urban SAR 3D imaging, we propose a new polarimetric decomposition method of maximizing polarization power. By rotating the polarization basis, we extract the scattering component with the highest signal energy as the input for the compressive sensing (CS)-based tomographic inversion. Subsequently, we generate 3D pseudo-color point clouds rendered with estimated height, backscattering coefficients, and polarization parameters. To assess the performance of these Pol-array-InSAR 3D imaging methods, we conduct a series of comparative experiments using multiple quantitative metrics on a dataset, acquired by our Ku-band unmanned aerial vehicle (UAV)-borne Pol-array-InSAR system. Ultimately, the proposed method demonstrates improved accuracy and consistency in scene reconstruction, as well as enhanced terrain feature discrimination.