A Conceptual Approach to Harbor Object Detection: The Potential of 3D-LiDAR-based Sensor Fusion for High Precision ENC

Kurzfassung

In the advancing era of autonomous maritime navigation, the precision of Electronic Nautical Charts (ENCs) is critical. This study conceptualizes a framework to validate and rectify ENCs nearly in real-time, leveraging high-resolution 3D LiDAR, DGPS, and IMU data. The approach of this study includes a domain-specific point cloud filtering, object segmentation, classification in compliance with the S-101 standard classes and a georeferencing strategy. By comparing identified objects against ENC data, this study pinpoints discrepancies and augments the ENCs with up-to-date object information. An evaluation through field tests is proposed, complemented by Traffic Sequence Charts and reference calibration via Realtime Kinematic GPS, ensuring practical relevance of our framework to real-world conditions. The contribution of this paper lies in offering a comprehensive solution for ENC refinement, thereby facilitating safer autonomous navigation by verifying that ENCs are reflective of current environmental conditions.