🌿 Athena: AI Utility Vegetation Management Platform

"Protecting the grid, preserving the wild."
Athena is an AI-powered Utility Vegetation Management (UVM) platform designed to autonomously identify, assess, and mitigate wildfire risks caused by vegetation encroachment on power lines.

---

Mission

Wildfires caused by power line failures are a catastrophic and recurring problem. Athena bridges the gap between Macro-Scale Analysis (Satellite/OSM) and Micro-Scale precision (Drone/Computer Vision) to provide utility operators with a unified, real-time risk assessment dashboard.

Used Dataset

Used TTPLA: An Aerial-Image Dataset for Detection and Segmentation of Transmission Towers and Power Lines from https://github.com/r3ab/ttpla_dataset. This dataset contains aerial images datset on transmission towers and power lines.

Key Features

Macro Scanner

Identify risks from the sky.

• Geospatial Intelligence: Ingests power line data from OpenStreetMap (OSM) and satellite imagery.
• Risk Heatmaps: Automatically calculates risk scores for thousands of assets based on location and historical data.
• Interactive Map: A high-performance React Leaflet map visualizing critical hotspots across California.

Micro Inspector

Validate threats with precision.

• Drone Simulation: Simulates a live drone inspection feed for high-risk zones.
• Computer Vision: Deployed a custom U-Net Semantic Segmentation model (trained on the TTPLA dataset) to distinguish between:
  • 🔴 Vegetation (Trees/Shrubs)
  • 🔵 Power Lines (Cables)
  • 🟡 Towers
• Real-Time Inference: Detects intersections between vegetation and cables to flag "Critical Encroachment" instantly.

---

Tech Stack

Component	Technologies
Frontend	React, Vite, Tailwind CSS, React Leaflet, Lucide Icons
Backend	Python, Flask, GeoPandas, OSMnx, Shapely
AI / ML	PyTorch, Segmentation Models, OpenCV, NumPy
Data	OpenStreetMap, TTPLA Dataset (Aerial Power Lines)

---

Dashboard Preview

1. The Command Center

A dark-mode, futuristic map interface for monitoring grid health.

2. AI Inference

Micro-Inspector detecting a tree branch touching a high-voltage line.

---

Quick Start

Prerequisites

• Python 3.9+
• Node.js 18+

1. Backend Setup

cd backend
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows
pip install -r requirements.txt
python app.py

2. Frontend Setup

cd frontend
npm install
npm run dev

3. Usage

• Open http://localhost:5173 in your browser.
• Click on any Red (Critical) hotspot on the map.
• Click "Launch Drone" to see the simulated view.
• Click "Run Inference" to let the AI analyze the image!

---

🧠 Model Training

The core AI model is a U-Net architecture with a ResNet-18 backbone, trained on the TTPLA dataset.

# Sample Inference Code
model = smp.Unet(encoder_name="resnet18", classes=4)
mask = model(image_tensor)
risk = check_intersection(mask['vegetation'], mask['cable'])

---

🔮 Future Roadmap

• LiDAR Integration: 3D depth analysis for precise clearance measurement.
• Predictive Growth: Use historical weather data to predict vegetation growth rates.
• Work Order Automation: Automatically dispatch tree-trimming crews to critical zones.

---

Made by Edwin Kim