Empirical Assessment of Geometric Accuracy of Underwater Lidar in Tropical Shallow Waters

Mentari K. Azzahra, Fickrie Muhammad, Arnadi Murtiyoso, Annette Scheider, Harald Sternberg, Gabriella Alodia

Abstract. This study empirically evaluates the geometric accuracy of point cloud data acquired using an underwater lidar (ULi) system in a tropical shallow water environment. The field test was conducted in the tropical waters of the Seribu Islands, Indonesia, characterised by relatively low turbidity. Terrestrial laser scanning (TLS) and close-range photogrammetry were employed as independent reference datasets. Geometric discrepancies between datasets were quantified using the multiscale model-to-model cloud comparison (M3C2) algorithm, and errors were statistically characterised using the median and median absolute deviation (MAD) to ensure robustness under non-normal distributions. The results indicate that the error of ULi relative to TLS is 0.008 ± 0.012 m, while the error relative to photogrammetry is 0.006 ± 0.013 m. In comparison, the discrepancy between photogrammetry and TLS is smaller, at 0.002 ± 0.004 m. Dimensional analysis of an acoustic Doppler current profiler (ADCP) frame further shows that ULi agrees with TLS and photogrammetry within the millimetre to centimetre range (0.000–0.015 m). Larger deviations in specific segments are attributed to local effects, including edge-related artefacts. Overall, the results demonstrate that ULi provides reliable geometric measurements in shallow-water conditions with low turbidity. Despite slightly lower accuracy compared to terrestrial methods, the system shows potential for underwater mapping applications, particularly in shallow water environments.