Research / Python / Simulation
Boiling Bubble 3D Reconstruction
Developed a computer-vision pipeline to segment boiling bubbles and estimate three-dimensional geometry and volume from video.
- YEAR
- 2025
- STATUS
- Research
Objective
Convert high-speed boiling imagery into measurable bubble geometry for heat-transfer research.
Approach
Combined object detection and segmentation models with custom Python geometry tools. The pipeline extracted bubble masks, generated rotational reconstructions, exported meshes, and recorded geometric metadata for later analysis.
Deliverables
- Automated and manual review workflows
- OBJ, XYZ, JSONL, and CSV exports
- Bubble area, centroid, and estimated volume
- Watertight mesh generation and visualization