摄影测量学
软件
均方误差
过程(计算)
计算机科学
扫描仪
云计算
激光扫描
计算机视觉
人工智能
粒子(生态学)
激光器
光学
数学
地质学
物理
统计
程序设计语言
操作系统
海洋学
作者
Kun Fang,Jiefei Zhang,Huiming Tang,Xiaolong Hu,Hao Yuan,Xiaotao Wang,Pengju An,Bingdong Ding
标识
DOI:10.1016/j.enggeo.2023.107170
摘要
Reliable particle measurements are essential when investigating the size and shape of particles in various fields of study, including engineering geology. Common methods to determine the morphology of particles, such as 3D laser scanners, are typically expensive or require extensive preparation. Therefore, a setup-friendly, low-cost, and accurate approach is required. In this study, a practical Trick-Image-Capturing (TIC) method based on smartphone photogrammetry to obtain size and shape parameters of 3D coarse particles is proposed. The process of image acquisition and particle reconstruction of the TIC method is demonstrated. Then, the performance of the TIC method is comprehensively assessed under various conditions using a 3D laser scanner, including smartphone quality, particle shape, shooting method, and postprocessing software. An optimal setting of the TIC method is selected through the cloud-to-cloud distance. Finally, the size and shape parameters of 32 coarse particles using the TIC method are also evaluated. Results show that the optimum setup of the TIC method is using a smartphone camera with a resolution above 12 MP under the ICS mode with a 30° shooting angle with Metashape postprocessing software. The root mean square error (RMSE) of cloud-to-cloud distances and the average error ratios under the optimum setup are less than 0.075 mm and 3.5%, respectively. Due to its fast image capturing process, easy setup, and low cost, the TIC method is practical and we expect it to become one of popular methods of measuring 3D coarse particles.
科研通智能强力驱动
Strongly Powered by AbleSci AI