Data from: Geomorphological signatures of known hurricanes and validation of theoretical emplacement formulations: coastal boulder deposits on Cuban low-lying marine terraces

Here we show the models made from SfM photogrammetry for 5 boulders found on the Cuban island emplaced by three hurricanes on the low marine terrace. To generate a 3D-scaled SfM point cloud of each studied boulder, we adapted an easy SfM workflow using Agisoft Metashape software (version 1.7.2). We surveyed each CBD obtaining pictures between 0.5 m - 3 m distance from the boulder and also climbing on the boulder to maximise the diversity of picture angles. The survey was carried over a time lapse of ~30 minutes and under cloudy conditions to avoid shadows created by direct sunlight. For scaling SfM results and assessing the quality of measurements, we used six wooden balls as photogrammetric targets. The diameter of the balls is 30 ± 0.2 x 10-3 m, as verified by repeated caliper measurements. We complemented these targets with two 70 x 10 cm wooden scales. All these targets were used as reference scale bars to correct the spherical distortion inherent to the camera optics and perform an auto-calibration. We further used these targets to check possible distortion after post processing and the generation of the resulting 3D model. In order to create a point cloud and a solid volume, we followed the software workflow: we first aligned the images to identify tiepoints, which were then used in bundle adjustment to compute the 3D dense clouds in a second step.本研究展示了基于运动恢复结构（SfM, Structure from Motion）摄影测量技术构建的5块巨砾模型，这些巨砾发现于古巴某岛屿的低海蚀阶地上，系三次飓风搬运沉积而成。为生成每块研究巨砾的带尺度三维运动恢复结构点云，我们采用了基于Agisoft Metashape软件（版本1.7.2）的简易运动恢复结构工作流。我们对每一处古巴巨砾沉积（Cuban Boulder Deposit，以下简称CBD）开展外业测量，在距离巨砾0.5米至3米的范围内拍摄照片，同时攀爬至巨砾之上以最大化拍摄角度的多样性。本次测量耗时约30分钟，且选择在多云天气下进行，以避免直射阳光产生的阴影。为校正运动恢复结构的测量结果尺度并评估测量精度，我们使用6个木质球体作为摄影测量标志点。经反复游标卡尺测量验证，这些球体的直径为30 ± 0.2 × 10⁻³ 米。我们额外增设2块尺寸为70×10厘米的木质比例尺作为补充标志点。所有上述标志点均用作参考比例尺，用于校正相机光学系统固有的球面畸变并完成自动校准。此外，我们还利用这些标志点核查后处理阶段及最终三维模型生成后可能存在的畸变问题。为生成点云与实体体积模型，我们遵循该软件的工作流程：首先对图像进行配准以识别连接点（tiepoints），随后将其用于光束法平差，于第二步计算得到三维密集点云。