A Novel Computational Imaging Method Using Reconfigurable Intelligent Surfaces

This paper proposes and numerically demonstrates a novel method for computational imaging using reconfigurable intelligent surfaces (RISs) that combines the benefits of compressive imaging using random patterns with spotlight imaging. In our approach, RISs are programmed to direct their beams toward a region of interest (ROI) where they interfere with each other, resulting in spatially diverse patterns that can multiplex the information within the ROI into a few measurements. Since each RIS produces a directive beam, the resulting speckle pattern is only confined to near the ROI, helping to reduce clutter, a typical issue with using random patterns. Furthermore, our method requires fewer measurements than raster scanning and can work even when RISs are separated, which is useful in practical implementation. The proposed RIS-enabled imaging system can find applications in security screening, user tracking, and activity recognition.