Comparative study of sampling strategies for sparse photon multispectral lidar imaging: Towards mosaic filter arrays

In this paper, we investigate the recovery of range and spectral profiles associated with remote three-dimensional scenes sensed via single-photon multispectral lidar (MSL). We consider two different spatial/spectral sampling strategies and compare their performance for a similar overall number of detected photons. For a regular spatial grid of pixels, the first strategy consists of sampling all the spatial locations of the grid for each of the L wavelengths. The second strategy is consistent with the use of mosaic filter-based arrays and consists of acquiring only one wavelength (out of L) per spatial location. Despite the reduction of spectral content observed in each location, the second strategy has clear potential advantages for fast multispectral imaging using only a single frame read out. We propose a fully automated computational method, adapted for each of the two sampling strategies in order to recover the target range profile, as well as the reflectivity profiles associated with the different wavelengths. These strategies were also assessed with high ambient background. The performance of the two sampling strategies is illustrated using a single-photon MSL system with L = 4 wavelengths (473, 532, 589 and 640 nm). The results presented demonstrate that although the first strategy usually provides more accurate results, the second strategy does not exhibit a significant performance degradation, particularly for sparse photon data (down to 1 photon per pixel on average). These results suggest a way forward for the integration of single-photon detector arrays with mosaic filters for use in a range of emerging photon-starved two-dimensional and three-dimensional imaging applications.