Single-Photon LiDAR System Noise Modeling and Virtual Image Synthesis Technology for Sea Observation

Single-photon lidar has higher resolution, sensitivity, and immunity to interference compared with conventional lidar, and is suitable for low signal-to-noise ratio environments and applications that require high-precision target information. However, noise modeling of the information acquisition process of single-photon lidar systems has been a challenging problem that severely limits the performance of single-photon lidar systems in practical applications such as performance evaluation, design optimization, and prediction. In this way, we modeled the noise of single-photon lidar systems during underwater information acquisition by establishing a simple model of underwater noise, followed by using the model to generate an underwater dataset of single-photon lidar systems on the VOC2007 dataset. The evaluation indicator metrics indicate that the sea observation dataset has high fidelity. The established noise model and the synthesized dataset are of great help to single photon LiDAR system imaging, image enhancement, target detection, and other vision tasks.