A frequency-hopping approach for microwave imaging of large inhomogeneous bodies

Microwave imaging involving large inhomogeneous bodies using inverse scattering technique is highly nonlinear. Therefore, as the body becomes larger compared to the wavelength, or when the contrasts of the inhomogeneity become larger, the nonlinear effect is more pronounced. However, this nonlinear effect is less pronounced at lower frequencies as opposed to high frequencies. For many practical applications, high quality data are collected at CW by time averaging. Multifrequency data can thus be collected, but the measurement time is linearly proportional to the number of frequency points at which the data are collected. We show the successful use of a small set of CW data measured over discrete frequencies. By using the image reconstructed from low frequency data as the initial guess to the higher frequency problem, the nonlinear effect can be mitigated. By slowly hopping from lower frequencies to higher frequencies, we can reconstruct objects which are as large as 10 wavelengths in diameter with high fidelity. The image reconstructed is much better than using the high-frequency data directly. Such a microwave imaging algorithm needs no a priori information about the inhomogeneous body.