Abstract

Microwave images of buried dielectric objects were formed from near-field complex-valued reflectance measured with a sampling array. The array contained seven printed circuit antennas in a staggered pattern of two parallel lines. Each antenna was individually and sequentially connected to a transmitter and receiver. The array was translated laterally to scan an area and the data arranged into lines of seven samples. Because the measurements were in the near-field, the spatial distribution of reflectance resembles the object’s shape. The measured data were processed to improve the resemblance. Individual rows of reflectance samples were first Fourier transformed. The transform was propagated to the air–soil interface and then inverted. The final step was forming the difference between inversion results for separate rows and there normalizing to the values for a preceding row. The processed reflectance more nearly resembled the object shape than did the measured data. In addition, subsidiary maxima, or clutter, were somewhat reduced.

© 1991 Optical Society of America