ForestVision: AI-Powered Vegetation Analysis Pipeline

Automated end-to-end forest vegetation segmentation, scoring, and policy insight using AI, enabling scalable environmental monitoring and decision-making.

Comment

What it does

🌳 ForestVision: AI-Powered Vegetation Analysis Pipeline

ForestVision is a modular, multi-agent AI pipeline designed to analyze satellite and aerial images (especially from Google Earth Pro) to detect vegetation coverage, generate meaningful ecological insights, and recommend policy actions — all with a click.

🚀 Live Demo

🔗 Hosted app (via ngrok): https://bf93-206-84-239-130.ngrok-free.app

🧠 What It Does

  • Accepts satellite/landscape images (ideal from Google Earth Pro, any year or timeline).
  • Segments out vegetation (dummy mask + scalable architecture).
  • Computes a vegetation density score.
  • Generates natural language insights.
  • Recommends policy actions (e.g., afforestation, conservation).
  • Displays results on a simple Streamlit interface.

🔍 Example Use Case

Upload a 2002 aerial snapshot from central Nairobi via Google Earth Pro.
ForestVision detects low vegetation density and recommends urban reforestation policies.

📦 Tech Stack

  • Language: Python 3
  • Frontend/UI: Streamlit
  • Agent Framework: Modular Agent Pipeline (custom)
  • Data Sources: Images from Google Earth Pro
  • Libraries:
    • OpenCV / PIL for image handling
    • NumPy for calculations
    • Pydantic for data validation
  • Cloud/DevOps:
    • Ngrok (for live sharing)
    • GitHub (code + orchestration)
  • External Datasets:

🛠️ How It Works

1. DataAgent

Loads raw image and preprocesses it.

2. SegmentationAgent

Performs vegetation segmentation (mask initialized to dummy but easily upgradable).

3. ScoreAgent

Calculates vegetation score based on green pixels.

4. InsightAgent

Converts scores into human-readable interpretations.

5. PolicyAgent

Recommends actions based on ecological insight.

6. Extra Threshold Logic

Displays custom messages:

  • ⚠ Low vegetation → Reforestation advice
  • 🌱 Medium vegetation → Improve cover
  • 🌳 Dense vegetation → Sustain & conserve

🌐 Input Format

Just drag-and-drop satellite images of land from any region, any year, exported via Google Earth Pro.

  • Ideal Resolution: 512×512 or 1024×1024
  • Formats: .png, .jpg, .jpeg

💻 Local Setup

git clone https://github.com/yourusername/forestvision.git
cd forestvision
pip install -r requirements.txt
streamlit run app.py

## How we built it

## Challenges we ran into

## Accomplishments that we're proud of

## What we learned

## What's next for ForestVision: AI-Powered Vegetation Analysis Pipeline

Built With

Share this project:

Updates