Solar panel fault detection

The brief

Inspecting thousands of panels on foot is slow. Train image models that classify panel faults (dust, bird droppings, cracks, hot spots) from photos and thermal images, so maintenance crews go straight to the faulty panels. Good looks like: accuracy per fault type, attention to rare faults, and a demo on uploaded images.

What you will build

1. Explore the visual and infrared image sets
2. Fine-tune pre-trained image models
3. Handle rare fault classes
4. Evaluate per fault type
5. Build a demo that tags uploaded images

Free, open data to build it with

• Infrared solar modules ↗ Raptor Maps (GitHub)
• Solar panel images (clean and faulty) ↗ Kaggle

Check each dataset's licence before you publish your work. Your mentor gives your team its own version of the brief (data, scope or region), so every team's project is its own.