A Real-Time Processed Static Mapping for Surface Vessel Navigation with LiDAR

Autonomous navigation in marine environments relies on accurate mapping for safe operation, yet distinguishing static and dynamic objects using LiDAR data presents challenges, particularly given the limited computational resources on surface vessels. This paper introduces a voxel-based mapping framework designed to generate real-time static maps by integrating lightweight occupancy mapping with cluster extraction techniques, relying solely on LiDAR data. The approach utilizes parallel processing to enhance computational efficiency and employs raycasting to mitigate occlusions and noise typical in marine settings. By tracking and filtering dynamic clusters-such as waves or other vessels—the system aims to provide a repre-sentation of the static environment. Initial testing suggests the framework's potential for reliability, efficiency, and scalability on resource-constrained surface vessels, though additional eval-uation is required to assess its wider applicability.