Research on passive millimeter wave imaging using compressed sensing architecture

The passive millimeter wave imaging (PMMW) formation has its unique technical advantage which can penetrate the smoke, cloud, drizzle and so on. This paper presents a method of passive millimeter wave imaging using a compressed sensing architecture which can sample the scene signals significantly lower than the Nyquist sampling rate, while the original scene image can be reconstructed with high precision. This method can significantly reduce the amount of sensors and the system complexity. We also adopt the block compressed sensing method for parallel imaging and reducing the data storages. Finally, we propose a non-convex set shrinking iteration (NCSHI) algorithm for reconstruction process. The algorithm uses the adaptive threshholding in the iteration process. In the simulation experiments, our algorithm has faster convergence speed and also leads to smaller reconstruction error compared to POCS. It is more suitable for passive millimeter wave imaging.