Experimental validation of sparse sensing technique in subsurface microwave holography

In this paper the implementation of the Sparse or Compressed Sensing (CS) technique to subsurface holography allowing data acquisition on a sparse grid is considered, which raises the scanning speed, reduces the amount of stored and transmitted data. To test various aspects of applying CS technique, a custom-built experimental setup acquiring data at several programmable sampling intervals was used. The setup consists of a vector network analyzer (VNA) with connected antenna and a two-coordinate mechanical scanner that moves the sample in the vicinity of the antenna. This technique allows simulating sparse sampling and testing various scanning and sounding parameters. As the test object a sample of cryogenic fuel tank thermal insulation was selected for the purpose of nondestructive testing (NDT). Upon acquisition, a complex data processing algorithm on the basis of windowing functions, interpolation, and filtering was applied. Then the Fourier-based back propagation algorithm was used for obtaining radar images by complex holograms reconstruction. Further development of the considered sensing technique is suggested, including the adaptive scanning on a non-equidistant grid.