UAV-Based Photogrammetric Inspection of an Urban Bridge in a Lagoonal Environment: Case Study of the Old Cotonou Bridge

Abstract

Bridge infrastructure in sub-Saharan Africa is often monitored with limited resources, leaving many ageing structures without a reliable geometric baseline for tracking deterioration. This paper reports on a UAV photogrammetric inspection campaign conducted on the Old Cotonou Bridge, a two-lane reinforced concrete structure crossing the coastal lagoon of Cotonou (Benin), with the aim of establishing a quantitative geometric reference for deck deformation monitoring. A flight of 573 images was captured at 56.1 m altitude using a DJI Mavic 2 Pro equipped with a Hasselblad L1D-20c 20 Mpx sensor (GSD: 1.28 cm/px), and the dataset was processed with Agisoft Metashape Professional 2.3.1 following a Structure-from-Motion and Multi-View Stereo workflow. Processing yielded a dense point cloud of 26.8 million points at 383 pts/m 2 , a DEM at 5.11 cm/px, and a georeferenced orthomosaic in WGS 84 / UTM zone 31N; six thematic classes were identified by automatic classification, followed by manual verification of the Road and Building classes. Deck deformation was then quantified through 2D polynomial regression of the deck surface, revealing seven statistically significant depression zones (D1–D7) with amplitudes ranging from −17.3 cm to −79.4 cm relative to the reference surface, over areas of 2 to 30 m 2 . The vertical accuracy achieved (RMSE Z = 0.44 cm) confirms that UAV photogrammetry can reliably serve as a quantitative tool for structural deformation detection on bridge decks, despite the use of only three Ground Control Points. The geometric reference dataset (T 0 ) produced here places at the disposal of asset managers a georeferenced database that is immediately usable for prioritising maintenance interventions on this and comparable structures.