The Problem
Climate change is damaging archaeological sites through rising temperatures, flooding, erosion, wildfires, ocean chemistry, and melting permafrost. Fragile artifacts and ancient writing can decay before people can read them.
SPARK Innovation Project
Slides 4-21
Recover Degrading Artifacts
This page follows the slide order from the newest PDF. The full deck is embedded first so all the original pictures, diagrams, tables, and slide layouts stay visible.
Full Slide Deck With Pictures
Project Story in Slide Order
Field Research
At the Metropolitan Museum, we saw ancient Egyptian artifacts where words on stone fragments had degraded.
Limits of Current Technology
Traditional digitization tools often need clean visual patterns. Stains, distortion, missing pieces, cost, and portability make on-site recovery difficult.
Our Solution
SPARK uses a camera-enabled Wi-Fi microcontroller, captures images, reduces noise by thresholding, sends the image to a computer vision model, and returns results for action.
Camera
Capture
Threshold
CV Model
Result
Community Feedback
Feedback led us to move tuning to the cloud, consider drone mounting, add optional LED/IR support, and include microSD offline storage.
Expert Interview
Dr. Brittany Profit recommended offline storage for dig sites and a user correction loop so the model can improve over time.
More Outreach
We shared our learning through the History Guardians FLL YouTube channel and a local public library presentation.
Our Experiment
We simulated an aged artifact with coffee-stained paper, captured it using an ESP32-S3 camera microcontroller, and used Vision Studio to test thresholding and OCR recovery.
Further Enhancements
Short term: shrink the dark-sensor light. Long term: fine-tune an archaeology-specific computer vision model and improve it with user-guided learning.
Thank You
SPARK protects the past by helping recover ancient words before they crumble or fade further.
Appendix and Design Rationale
The appendix covers Vision Studio preprocessing, OCR testing, the edge-device plus cloud-AI design, connectivity options, offline storage, and lighting improvements.
Team Roles
Isabella researched the problem and created experimental artifacts. Mason researched existing solutions. Luke researched computer vision and tested preprocessing and CV models. Claire researched the problem, set up the microcontroller prototype, and integrated the web app with OCR algorithms.
Community Engagement
Special thanks to Mr. Friedman, a history teacher at RCDS, for sharing archaeology experience and suggestions.