Calibration of a Microwave Imaging System Using a Known Scatterer

An increasing number of operational microwave imaging systems have been presented in recent years, especially for medical imaging. This has increased the focus on the practical aspects of microwave imaging, such as the need for calibration, how to decrease measurement time, and how to minimize the effect of noise.

At the Technical University of Denmark, a 32-channel microwave imaging system for breast cancer screening has been under development for some time. In this system, each antenna is equipped with its own transceiver module, containing amplifiers, switches, and a mixer. This design ensures that the low-amplitude RF signals, available at the terminals of the antennas, only need to travel a very short distance to get to the low-noise amplifier, while the RF as well as the IF signals running to and from the transceiver modules all have significant amplitudes.

However, some leakage between the transmission and receiving channels inside of the transceiver modules has been detected. This is most likely caused by the lessthan- perfect isolation of the switches in the modules as well as leakage through the PCB itself. Since the presence of such a leakage signal in the measurements seriously influence the imaging capability of the system, it is of interest to remove it.

In this work, a calibration procedure capable of removing a constant offset, i.e., the leakage, from the measured signals is presented. The calibration procedure is based on a comparison between the relative change observed between a measurement with an empty imaging system and a measurement with a simple known object in the system, i.e., a metal cylinder, and the relative change observed between simulations of the same two situations.

Assuming that the simulation software is capable of accurately modeling the imaging system, some relatively simple considerations on the difference between the measured and simulated results lead to an expression for the offset. Once the offset is known, this can be subtracted from the measurement of the unknown object to be investigated (the breast).

The main advantage of the calibration procedure presented here is that the time needed for doing the calibration is very low compared to a more rigorous and classic type of calibration in which measurements should be made for all combinations of transmitters and receivers with and without the antennas connected to the transceiver modules to allow for determining the leakage signal.

Apart from being a time consuming task, the bending of the cables caused by removing the transceiver modules from the antennas would change the phase of the measured signal, thereby introducing a source of error in the calibration procedure. In the procedure presented here, one can simply remove the object to be imaged (the breast) after the measurement is complete, insert the known scatterer, and do a single standard measurement, thereby allowing for running the calibration after each measurement.