Zaitseva-Pärnaste, I. and Herrmann, H., 2024. Comparison of 3D model generation using Structure-from-Motion by open-source software and validation with LiDAR for revealing changes in eroding coastal cliffs. In: Phillips, M.R.; Al-Naemi, S., and Duarte, C.M. (eds.), Coastlines under Global Change: Proceedings from the International Coastal Symposium (ICS) 2024 (Doha, Qatar). Journal of Coastal Research, Special Issue No. 113, pp. 1054–1058. Charlotte (North Carolina), ISSN 0749-0208.

Precise 3D modeling of coastal areas is a crucial tool for monitoring, forecasting, and thereby facilitating the management of coastal processes. This study explores the efficacy of Structure-from-Motion (SfM) technology, utilizing open-source software OpenDroneMap and COLMAP, for 3D modeling of coastal cliffs, comparing it to the more costly LiDAR method. It assesses model quality, including precision and the ability to detect small, potentially catastrophic faults not visible in 2D orthophotos. The paper evaluates each software's features, such as georeferencing, user-friendliness, and performance on High-Performance Computing systems, against drone-based LiDAR benchmarks. The study reveals that the evaluated open-source packages generate good quality models: OpenDroneMap provides automatic georeferencing, while COLMAP produces a slightly higher quality model but without automatic georeferencing, which is a necessity for natural environments. COLMAP automatically chooses a parameter set to produce a high-quality model with a precision of 10 cm or better, but the parameters remain unknown to the user, making manual adjustment challenging. OpenDroneMap, on the other hand, makes manual adjustment of the parameters easier but also requires more parameter tweaking to achieve the same accuracy as COLMAP. While LiDAR has been extensively utilized for coastal monitoring, SfM is typically considered a secondary method due to the significant effort required to achieve high-quality results, including the need for high-quality photographs, georeferencing, and extended data processing times. However, this paper introduces a paradigm shift by demonstrating the feasibility of using SfM as an independent and effective method for assessing coastal cliff erosion, provided that suitable software is employed. The study evaluates two open-source software options for generating 3D models via SfM, verifying their effective performance. This revelation presents a promising avenue for cost-efficient coastal cliff monitoring, suggesting that with the right tools, SfM can stand on equal footing with LiDAR in terms of reliability and effectiveness.