A Microwave Camera based on Random Projections

The technique proposed in this contribution allows a large reduction of the acquisition time and uses a fixed probe. The method is strictly related to the universal sampling strategy based on random measurement matrices developed by Candes Tao and Donoho and is similar to the single-pixel camera proposed by Rice University for optical and infrared data acquisition. The set-up consists in a reflectarray whose elements are terminated in electronically switchable loads based on PIN diodes or MEMS. The reflectarray is illuminated by the source. A single probe receives the field scattered by the reflectarray with different configurations of the switches, randomly selected. The number of measurements (i.e. the configuration of the reflectarray loads) depends on the sparsity level of the source. In case of high sparse (or sparsificable) source, the number of measurements is very low and allows an almost 'real time' measurement of the field. Numerical examples regarding array diagnosis will be discussed, showing the effectiveness of the technique.