Multi-sensor data capture and fusion for mapping hazardous materials onboard ships

Richard J. D. Moore, Peter Ø. Ivarsen, Ahmed Mohammed

Keywords: SLAM, semantic segmentation, SAM3, gas, 3d, hidden point removal, point cloud denoising, IHM, ship decommissioning

TL;DR: We present our sensor rig design and ship dataset together with initial steps towards fusing semantic information from 3D reconstructions with unstructured IHM data to enable automatic localisation of hazardous objects and materials onboard ships.

Abstract: Ship decommissioning requires accurate, spatially grounded knowledge of hazardous materials within complex vessel interiors, yet this information is often fragmented across heterogeneous sources and disconnected from the ship’s current physical state. We present a multi-sensor data capture and fusion pipeline that combines mobile 3D scanning, gas sensing, and vision-language perception to generate enriched, semantically annotated 3D ship models. The system integrates lidar–inertial SLAM, RGB imagery, and combustible gas sensing, along with post-processing methods for point-cloud denoising and visibility-aware multi-view fusion. Using the museum ship DS Hestmanden as a case study, we demonstrate spatial localisation of gas measurements and 3D semantic mapping of ship-like environments. Our results constitute a first step toward linking unstructured hazard data with spatial models, enabling hazard localisation and safer, more informed ship decommissioning.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 18