Three-dimensional reconstruction of bubble geometry from single-perspective images based on ray tracing algorithm

Yuchen Song,Guangyuan Huang,Junlian Yin,Dezhong Wang

DOI: https://doi.org/10.1088/1361-6501/ad7e43

IF: 2.398

2024-09-25

Measurement Science and Technology

Abstract:An improved shadow-based imaging method combined with a ray-tracing algorithm is proposed for three-dimensional (3D) reconstruction of bubble mesostructures. The 3D micro-scale interface measurement of bubbles categorized as transparent objects is challenging at the forefront of research in the field of multiphase flows. In the study, a reconstruction algorithm is developed using a single high-speed camera under a small-angle diffused backlight source based on Helmholtz reciprocity, Gaussian optics, and Snell's law. The accuracy of the algorithm is assessed by comparing the reconstruction results of the rendered images generated from the 3D model using Blender. Standard models include a transparent hollow sphere model and several irregular bubble models. The accuracy of the algorithm is determined by analyzing and comparing the contour lines along the edges of the bubble models. The error of the method in the paper can be improved below 5% with the ray tracing optimization. The average uncertainty Uc is 2.8%, according to the evaluation of the models in Blender. The extreme deformation of the bubbles in a turbulent flow field enhanced by ultrasonic is performed to evaluate the 3D reconstruction algorithm. Prior to the fragmentation of bubbles into micro/nanobubbles by ultrasonication, the bubble surface undergoes severe deformation under continuous bombardment by small-scale turbulence eddies. The proposed algorithm exhibited high accuracy and easily performing in experiments.

engineering, multidisciplinary,instruments & instrumentation