Robust three-dimensional particle tracking for small-scale dynamics

This dissertation tackles the complexities of characterizing flow and particle dynamics in constrained environments by advancing 3D particle tracking techniques at micro and meso scales. Current methods, such as stereo vision, holographic, and tomographic techniques, face challenges in reconstructing and tracking small-scale particles, largely due to spatial and illumination limitations inherent in single plenoptic camera systems. To address these challenges, this research is structured into two pivotal stages: first, by developing and refining microscale 3D particle tracking methods, and then, by expanding these optimized techniques to mesoscale 3D particle tracking applications. Recent advancements in microlens manufacturing and camera sensing technologies have enabled the integration of a single light field camera into a microscope. This integration facilitates 3D reconstruction of particle positions, albeit with a compromise in spatial resolution. Traditional methods, such as the Multiplicative Algebraic Reconstruction Technique (MART), face notable challenges, including high computational demands and intricate calibration processes. To overcome these challenges, we propose a novel, non-iterative ray tracing method, complemented by a robust post-capture alignment of the microlens array sensor. This technique is designed for the rapid reconstruction of sparse particle concentrations in light field particle tracking scenarios. Our approach uses a kd-tree algorithm to efficiently organize voxels intersected by various rays, thereby reducing both memory requirements and computational time. For particle identification and spatial reconstruction, we deploy a cloud point classification algorithm. The efficacy of this method is assessed through a physically-based and realistic model of a light field camera. Also, we built an optical system within a microscope to capture the 3D velocity field in a fully-developed region. The results obtained demonstrate a good agreement with theoretical solution, showing the potential of this method in practical applications. Expanding from the microscale approach, this study extends to mesoscale measurements (order of 10 mm $\times$ 10 mm $\times$ 10 mm), which are central for investigating flow in porous media, interactions between flow and small biological structures, and other applications. To accommodate mesoscale observations, the non-iterative ray tracing technique is adapted for compatibility with a perspective macro lens. At this scale, challenges arise due to increased optical distortion and a decrease in stability within the cage system, especially when compared to the more stable microscope setup. To counter these issues, a practical calibration method is employed, using blob detection complemented by circularity filters. This approach effectively addresses the challenges posed by the larger scale and different optical characteristics of the macro lens. The accuracy of the mesoscale method is tested using a realistic light field camera model equipped with a macro lens. Finally, the dissertation tackles practical applications of this framework in addressing environmental problems. It specifically examines the dynamics of flexible canopies subjected to low blade deformation or low Cauchy numbers, as well as the dispersion of particles in jet flows. These case studies highlight the framework's capability to provide efficient and rapid 3D particle reconstruction at various scales. Looking forward, the dissertation showcases the potential for future research to explore the motion and behavior of, e.g., microorganisms, which are fundamental to biotechnological advancements and environmental settings.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Theoretical & Applied Mechans

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Hong, Liu

Contributors dc:contributor

• Chamorro, Leonardo
• Best, James Leonard
• Gazzola, Mattia
• Feng, Jie

• 3d Ptv/piv
• Micro Fluid

Statement dc:rights

• Copyright 2024 Liu Hong

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/124171

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F124171