A Microwave Subsurface Imaging Technique for Rough Surfaces

Subsurface microwave imaging is the image reconstruction of objects buried below the surface of a medium, such as soil. Modeling the interactions between the background medium and objects, as well as artifacts due to the limited view of objects, are the major concerns in the image reconstruction process. In this work, we have presented an efficient subsurface microwave imaging technique for the reconstruction of targets in a two-layered medium with an arbitrary rough interface. The use of a computationally expensive two-layered background Green’s function in the classical integral equation model is avoided by exploiting the boundary data model where only a simple homogeneous Green’s function is required. The contrast source imaging technique is reformulated in terms of boundary data and total variation (TV) regularization is included in the cost function multiplicatively to handle limited view and noisy data in a robust way. Numerical simulations demonstrated that the proposed method dramatically reduces the computational time while keeping the same accuracy in comparison to the classical approach.