水槽
水下
三角测量
计算机视觉
折射
交叉口(航空)
三维重建
人工智能
光线追踪(物理)
分布式光线跟踪
结构光
光学
地质学
计算机科学
算法
追踪
数学
工程类
几何学
物理
流量(数学)
海洋学
航空航天工程
操作系统
作者
Huitao Sun,Hai Du,Muguo Li,Xinyue He
出处
期刊:Measurement
[Elsevier]
日期:2021-04-01
卷期号:174: 108971-108971
被引量:5
标识
DOI:10.1016/j.measurement.2021.108971
摘要
Stereo-vision-based measurement is affected by refraction of light in hydrodynamics and ocean-engineering experiments that are performed in glass flumes. Using the conventional 3D reconstruction strategy in measurement of underwater objects will lead to considerable errors. In this paper, a new reconstruction method based on ray tracing is proposed for measurement in the flume experiments. The refractive interfaces of the experiments are estimated by a robust algorithm, so that light rays corresponding to the image points can be recovered reliably. In the estimation algorithm, differential evolution is employed to compute globally optimal parameters from general image correspondences. To determine the more reasonable reconstructed points, a modified triangulation approach is also developed. The intersection of the back-projected rays is computed by correcting the corresponding image points. Experiments are conducted with synthetic data and real underwater objects. The results demonstrate the effectiveness and convenience of the proposed method in such underwater measurement tasks.
科研通智能强力驱动
Strongly Powered by AbleSci AI