Non-line-of-sight imaging based on Archimedean spiral scanning

Non-line-of-sight imaging can reconstruct objects hidden out of sight, which has drawn great attention in recent years. However, long data acquisition time is required since most of transient non-line-of-sight imaging need multi-point raster scanning of the visible wall. Here, we propose an Archimedean spiral scanning method based on confocal non-line-of-sight imaging, which greatly reduces the data acquisition time. Fewer scanning points are sufficient for imaging owing to accurate extraction of effective information. The Archimedean spiral scanning method is experimentally verified with published data and real scenes, then the curvature regularization algorithm is used to process data to achieve higher quality imaging. In the case of guaranteeing the imaging quality, the number of scanning points is reduced by an order of magnitude with the proposed method compared to traditional multi-point raster scanning method in this experiment. As a comparison, the imaging quality of Archimedean spiral scanning method is higher than those of the circular and confocal non-line-of-sight scanning methods at the same number of scanning points. This scanning method is expected to have a certain remarkable effect for real-time NLOS imaging.