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

Single-photon has higher resolution 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. We improve the noise model and build two types of large-scale single-photon LiDAR observation datasets. Firstly, we modeled the noise of single-photon LiDAR systems during information acquisition on land and we also initially developed a simple model of underwater noise, followed by using the models to generate a single-photon LiDAR system land observation dataset as well as an underwater dataset based on the VOC2007 partial dataset. The evaluation indicators indicate that the observation datasets have high fidelity, and the model established and the synthesized datasets are of great help for visual tasks of the single-photon LiDAR system such as imaging, image enhancement, and target detection.