XR-INSPECT

The Challenge

Name:
Remote Inspection Support and Augmented Collaboration between Onboard and Shore-Based Experts

Domain:
Maritime Remote Inspection, Spatial Computing, Maritime Robotics, Augmented Reality, Edge AI

Challenge proposer:

Objectives:
The primary objective is to establish a fully functional system enabling remote expert supervision of onboard inspections through real-time AR interfaces and robotic platforms. Specific goals include improving inspection accuracy, enabling expert guidance without physical presence, prioritizing inspection tasks using operational and historical risk data, validating inspection procedures on operational vessels through live demonstrations, where feasible, with a preference for arranging vessel access independently of consortium member support., and ensuring that all outputs meet regulatory and classification requirements. The project leverages SERMAS’s experience to ensure workflow maturity, technological
readiness, and practical applicability.

The Solution

Name:
XR-INSPECT – Augmented Remote Collaboration via Motion-Compensated XR in Maritime Environments

XR-INSPECT is an Agile “Ship-to-Shore” co-creation framework that deploys a hardware-agnostic Augmented Reality (AR) interface. It empowers onboard crew members to safely execute complex, targeted inspections under the real-time, spatial guidance of remote senior surveyors. By creating a persistent digital Extended Reality (XR) layer that connects robotic platforms with human operators, the system fosters a continuous, data-driven decision-making loop that transitions maritime maintenance into a proactive ecosystem while reducing human exposure to hazardous environments.

XR-INSPECT addresses the analytical flaws of standard 2D video-streaming and Augmented Reality frameworks, which typically fail due to scale ambiguity and the continuous 6-Degrees of Freedom (6-DOF) macro-motion of maritime vessels. To overcome these challenges without relying on heavy and fragile 3D LiDAR systems, the project adopts a dual-path spatial computing architecture. Initially, it utilizes Marker-Assisted Tracking to impose a rigid coordinate system on the environment. Concurrently, it develops a Markerless Vessel-Motion-Compensated Stereo Visual-Inertial Odometry (S-VIO) engine, which mathematically fuses high-frequency kinematic data from an Inertial Measurement Unit (IMU) with stereoscopic visual disparity data to subtract the ship’s global motion and maintain digital anchoring.

To combat the Faraday-cage effect of dense steel bulkheads that sever cloud connectivity, XR-INSPECT deploys lightweight Edge AI Convolutional Neural Networks (CNNs) directly onto the local processing unit. Operating fully on-device, these models execute real-time semantic segmentation in low-light conditions, autonomously differentiating critical features like high-pressure pipelines from structural reinforcements and embedding this metadata into the spatial map.

Furthermore, the system features a novel asynchronous spatial caching protocol that buffers and reconciles data topologies post-network-dropout. This back-end infrastructure is co-architected to automatically format AR annotations and spatial data into Class-compliant digital logs, securely populating central databases to generate Classification Society UTM and Market Condition Survey reports.

The Solution Provider

NI BOC DIGITAL SERVICES LTD is an innovative technology SME based in Cyprus, focused on delivering advanced digital transformation and project management solutions. The company specializes in bridging the gap between rigid financial/regulatory constraints and advanced technical execution. By leveraging spatial computing, Artificial Intelligence, and augmented collaboration tools, NI BOC DIGITAL SERVICES LTD designs and implements robust software architectures specifically tailored for hazardous and high-stress industrial environments, such as the maritime sector.