Lucentara / Dinosty Fossils specializes in preparing and finishing ammolite gemstones and fossils. A key production step is spraying a 2-component (2K) clear coat to protect and enhance the optical properties of the stone. One of the biggest issues is ensuring full coverage — pinholes and micro-bubbles can form if spray passes are uneven or if substrate conditions vary. These defects reduce yield and increase rework costs.

Develop a prototype quality control system that can detect, predict, or prevent pinholes and bubbles during or immediately after spraying clear coat on ammolite specimens.

By the end of the project, students should provide:

1. Research Report — summary of best practices in coating defect detection (in automotive, aerospace, or jewelry contexts).

2. Proof-of-Concept Prototype (software or hardware mockup):

Computer vision pipeline (using open-source models like YOLO or SAM) to highlight bubbles or voids in test images.

OR process-parameter model predicting bubble risk based on humidity/temperature/spray overlap.

3. Implementation Guide — simple steps for how this system could be deployed in Lucentara’s spray booth with minimal cost (USB camera, LED lighting, Raspberry Pi/Jetson).

4. Presentation / Demo Video — walk-through of how the AI system works with example data.