3D computational imaging for photonic LiDAR with high noise resilience and photon efficiency

In this work, we introduce a new computational imaging pipeline for a single photon sensitive LiDAR system, which incorporates an image reconstruction module with a point cloud processing module to improve noise resilience and photon efficiency. The image reconstruction module is based on a 3D extension of an iterative image restoration algorithm in a constrained maximum likelihood framework. The point cloud processing module consists of a hierarchical point cloud denoising and segmentation scheme, which produces a refined point cloud data for 3D image reconstruction. Experimental results demonstrate the effectiveness of the proposed computational imaging method under robust operating conditions.