抄録
In this paper, a novel signal processing method for microwave radar is proposed, which enables to estimate the internal cross-section of concrete structures. By using proposed method, enhanced information such as position, surface shape and material (between metal and non-metal) of buried objects can be discriminated. As the objects are buried near the concrete surface, signal observation is difficult because a reflected wave is often interfering with a surface wave in an observed signal. In order to reduce an influence of the interference, nominal surface waveform identified from the observed waveform without buried objects is subtracting from other observed waveforms. As a result, boundary and material of the buried objects are discriminated more precisely. Next, a boundary model fitness whose shape is the normal distribution function with characteristics of the original model fitness is newly introduced to enhance a vertical resolution of reconstructed images for buried objects. In the back projection images calculated by using them, the boundaries of the buried objects are discriminated more clearly. These proposed methods have applied to experimental application to several concrete test specimen, and position and diameter estimation and material discrimination performance have confirmed.
