ICIEA 2026 Workshop

Workshop 6: From Underwater Object Tracking to Underwater Inspection and 3D Modelling

Speakers:

Speaker 1: Lijun Jiang

Title: Senior Lecturer

Affiliation: Republic Polytechnic, Singapore

Underwater visual sensing remains a challenging frontier due to light absorption, scattering, colour distortion, attenuation, and the refractive properties of water, all of which significantly affect tracking, stitching, and three-dimensional reconstruction performance. This workshop presents recent advances in underwater inspection and modelling through three complementary research directions.

First, it introduces an intelligent hull inspection framework that integrates Simultaneous Localization and Mapping (SLAM), diver monitoring, and a user-friendly graphical interface to support more efficient, accurate, and safer maintenance operations. By enabling real-time data acquisition and visualisation, the system helps maritime operators detect defects and anomalies earlier while reducing inspection time and operational downtime.

Second, the workshop examines underwater video stitching through a comparative study of preprocessing methods evaluated using a task-oriented multi-metric framework. Rather than relying only on perceptual image quality, the study assesses feature matching performance, reprojection accuracy, seam visibility, and panorama quality across clear, turbid, and artificially lit environments, offering practical guidance for method selection under different conditions.

Third, it presents a robust monocular pipeline for underwater 3D reconstruction that combines GLOMAP with 2D Gaussian Splatting to improve pose estimation and suppress rendering artefacts in low-texture, hazy scenes. Together, these topics provide a coherent overview of how modern vision, mapping, and neural rendering techniques can address key challenges in underwater inspection, navigation, and digital modelling, with strong relevance to maritime maintenance, autonomous systems, and intelligent visual perception.