Answers are grounded in this paper and cite their sources.
Underwater photogrammetry based on Structure-from-Motion (SfM) has become an established tool in coastal engineering and environmental monitoring; however, shallow-water applications remain challenging due to turbidity, light attenuation, and reduced visibility, which can compromise geometric reliability. This study presents and validates a low-cost methodology for generating underwater 3D models using a Remotely Operated Vehicle (ROV) equipped with a custom multi-camera system. The configuration was designed to increase angular redundancy and surface coverage using affordable components. Field experiments were conducted at San Julián Dock (Spain), a historical dry dock that can be drained, enabling controlled validation of submerged reconstructions against a high-precision terrestrial laser scanning (TLS) reference. Across two analyzed areas, the C2C analysis yielded mean deviations of 3–4 cm and Root Mean Square Errors (RMSE) values of 4.6–5.6 cm relative to the TLS benchmark, while the complementary M3C2 assessment showed median absolute distances of approximately 3.5–4.6 cm together with larger localized discrepancies. Despite the absence of GNSS positioning and underwater control points, the results demonstrate centimetre-level geometric agreement under favorable shallow-water conditions, supporting the applicability of the proposed approach for engineering inspection, monitoring, and documentation of submerged heritage. The study highlights the potential of low-cost multi-camera ROV systems as a practical and replicable solution for underwater 3D modeling when operational constraints limit the use of high-end instrumentation.
No takes yet. Share an insight, caveat, or question.
Santos et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: