GPR imaging algorithm of targets located in layered mediums based on CS

Previous studies in compressive sensing (CS) based subsurface imaging methods depended on two important assumptions; namely, that the imaging scene is modeled as a single homogeneous medium and that the relative permittivity of background environment is known. However, these assumptions are not always valid for the practical application of underground imaging with ground penetrating radar (GPR). To this end, a new GPR imaging algorithm based on CS is proposed in this paper. The algorithm attempts to image the point like targets located in unknown layered mediums subsurface. A 2-D underground objective vector firstly obtained includes both unknown layered mediums vector and unknown targets vector. Then the scattering and imaging model is formed by the ray-based simulation methods. This model is designed under the framework of CS theory. Finally, the GPR data can be reconstructed from an extended dictionary by solving a convex l₁ minimization problem. Experimental results of GPR simulation data validate the effectiveness of the proposed algorithm.