Sparse sampling-based microwave 3D imaging using interferometry and frequency-domain principal component analysis

摘要

Microwave radar 3D imaging with high resolution generally requires a great number of samples. The authors aim at accurate reconstruction of microwave radar images while significantly reducing the required number of samples. A novel algorithm is proposed which realises sparse sampling with nearly 50% data reduction and high-quality restoration, based on interferometry and principal component analysis (PCA) in frequency domain. Interferometric processing is utilised to concentrate the frequency spectrum into low-frequency stage, thereby reaching an effective sparse representation of radar image. Furthermore, PCA is introduced to reform radar image according to its principal characteristics in frequency spectrum, without side-lobe artefacts and receiver noise. Experimental data in anechoic chamber demonstrates the great potential of the proposed approach.