Multichannel and Multiview Imaging Approach to Building Layout Determination of Through-Wall Radar

Yong Jia,Guolong Cui,Lingjiang Kong,Xiaobo Yang

DOI: https://doi.org/10.1109/lgrs.2013.2283778

IF: 5.343

2014-01-01

IEEE Geoscience and Remote Sensing Letters

Abstract:This letter considers the problem of building layout determination using through-wall-radar imaging technology, which employs multiple transmit-receive channels to implement multiview synthetic aperture imaging. In each view, the phase errors of multichannel data introduced by the unknown walls deteriorate significantly the performance of the coherently data-combined algorithm by Le et al. Herein, we first obtain multiple single-channel building layout images for all independent channels of each view and propose a novel noncoherent fusion method named multiply-subtract-add to combine them into a single-view layout image. Then, we present an M-N-K detector plus median filtering to fuse multiple single-view layout images and reduce the existing cavities and burrs of wall images. The experimental results reveal that the presented noncoherent image fusion method gives the single-view layout images with higher signal-to-clutter-and-noise ratio than the conventional coherent algorithm based on data combination and a near-tidy panorama layout image is generated almost without the cavities and burrs.