Electromagnetic two-dimensional scattering database for verifying inversion algorithms

The inverse scattering problem has found many applications in the field of remote sensing, non-destructive evaluation, geophysical exploration and biomedical imaging. There are many inversion algorithms developed. However, there are only several measurement datasets available for verifying the robustness of the algorithms. The ill-posed nature of the inverse problem requires multiple incidences and multiple receivers for the measurement data. There are several kinds of measurement methods available to realize this requirement. One way is via a bi-static setup, i.e., to use a single incidence antenna and a single receiving antenna rotating around multiple positions. This method takes a long time but has low mutual coupling between the two antennas. The other way is via a multi-static setup, i.e., to use switches to control an array of antennas. This method is fast but suffers from high mutual coupling. This paper focuses on a bi-static experiment to collect the scattered field data for the purpose of the verification of inverse scattering algorithms. The multiple-incidence and multiple-receiver setup is realized by a two-arm rotating stage. Targets under test are circular cylinders placed at the center of stage. The TM polarization incident wave is considered. Similar with the Fresnel dataset [1], the incident field is collected by the field gathered without any target and the scattered field is obtained by subtracting the incident field from the total field collected with the target presenting. The frequency ranges from 12 to 18 GHz. The accuracy of the measurement is checked by comparison with method of moment. The organization of the paper is as follows: First, the experimental setup is described in detail, including the two-arm rotation stage, motor controller and vector network analyzer. The whole system is controlled by a LABVIEW program. Then the calibration process is demonstrated to compensate for modeling and measurement errors. Finally, the TM polarization measurement is discussed and accuracy of the data is demonstrated.