Verdant

Inspiration

Cities rely on roads, bridges, and public buildings every day, but visible damage is often found too late. Small cracks, potholes, surface wear, and structural issues can grow into expensive repairs that require more materials, more demolition, and more waste. This is not only a safety and maintenance problem — it is a sustainability problem.

We were inspired by the idea that cities could use drone images to inspect infrastructure earlier and make smarter repair decisions. Instead of waiting until damage becomes severe, Verdant helps identify issues sooner and recommends sustainable repair options that reduce waste, cost, and environmental impact.

What it does

Verdant analyzes drone images of roads and buildings to detect visible infrastructure damage. After an image is uploaded, the system identifies damaged areas, estimates severity, and generates a repair report with sustainable recommendations.

The goal is to help city teams answer questions like:

Where is the damage?
How severe is it?
How urgent is the repair?
What would it cost to fix?
Is there a more sustainable repair option?
How much waste or carbon impact could be avoided?

By combining computer vision with AI-generated repair intelligence, Verdant turns inspection images into actionable sustainability-focused infrastructure reports.

How we built it

We built Verdant as a web application using:

React for the frontend interface
TypeScript for safer and more organized code
Vite for fast development and building
React Router for page navigation
Lucide React for icons and UI elements
CSS for custom styling
localStorage mock authentication for the demo login flow

The app includes a landing page, authentication page, and protected dashboard where users can upload inspection images and view results.

The planned AI pipeline connects uploaded drone images to a YOLOv8 damage detection model, which detects visible road or building damage. The detected damage data is then passed into an AI analysis layer that generates severity scores, repair cost estimates, repair timelines, and sustainable repair recommendations.

A simplified version of the pipeline is:

$$ \text{Drone Image} \rightarrow \text{Damage Detection} \rightarrow \text{Severity Analysis} \rightarrow \text{Sustainable Repair Report} $$

Challenges we faced

One challenge was connecting the sustainability impact clearly to infrastructure damage. At first, the project seemed like only a damage detection tool, but we realized the bigger issue is that late repairs create more waste, require more materials, and lead to higher emissions. This helped us focus Verdant around early detection and sustainable repair planning.

Another challenge was designing the product so it felt useful for real city teams. We wanted the output to be more than just a highlighted crack or pothole. The report needed to include severity, cost, urgency, and environmental impact so the user could actually make a decision.

We also had to balance building a clean demo with planning a larger AI system. The frontend was built to show the full user flow, while the detection and analysis pipeline was designed around YOLOv8 and AI-powered repair recommendations.

What we learned

We learned that sustainability is not only about renewable energy or recycling. It also applies to how cities maintain infrastructure. Repairing damage earlier can reduce unnecessary demolition, lower material use, prevent waste, and extend the life of roads and buildings.

We also learned how important it is to make AI results understandable. A city worker or building manager does not just need a model prediction — they need a clear explanation of what was detected, how serious it is, and what action should be taken next.

What's next

Next, we want to connect the frontend directly to the YOLOv8 model and improve the damage detection pipeline. We also want to add clickable damage overlays, real authentication with Supabase, a database for inspection history, and more detailed sustainability metrics such as estimated material waste avoided and carbon savings.

In the future, Verdant could help cities inspect infrastructure faster, prioritize repairs better, and choose greener solutions before small problems become expensive and wasteful.