Skip to content

AI-powered early detection system for poultry diseases (1st Prize - Data Science, Cornell Hackathon 2025). It uses deep learning to analyze chicken vocalizations and fecal images, helping farmers identify signs of illness before outbreaks occur. πŸ›  Actively being developed with ongoing improvements to backend, models and dashboard interface.

Notifications You must be signed in to change notification settings

farhan-439/avian_alert

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

43 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

πŸ” AvianAlert: Early Detection, Immediate Action

πŸ† First Prize – Data Science Track
πŸ›οΈ Cornell Institute for Digital Agriculture Hackathon 2025

AvianAlert is a real-time AI system that detects poultry diseases early using sound and fecal image analysis.

AvianAlert Logo


🎬 Live Demo

AvianAlert.mp4


πŸ“Š The Problem

Each year, preventable poultry diseases cause devastating losses:

  • πŸ’Έ $600M in economic damage (last quarter)
  • πŸ” 20M dead chickens
  • πŸ₯š 96.4% egg price increase year-over-year
  • 🌾 $100M losses to American farmers

🧠 Solution Overview

AvianAlert is an AI-driven monitoring tool built on three main components:

πŸ”Ή Flock Segmentation

  • Divides poultry facilities into zones
  • Enables localized disease containment

Flock segmentation


πŸ”Ή AI Sound Analysis

  • Classifies poultry vocalizations in real time
  • Detects early signs of respiratory distress

AI Sound Analysis


πŸ”Ή AI Excreta Analysis

  • Analyzes images of chicken droppings
  • Detects diseases like:
    • Salmonella
    • Newcastle Disease (NCD)
    • Coccidiosis
    • Avian Flu

Excreta Analysis


πŸ§ͺ Technical Architecture

🎧 Audio Classification (CNN + Burn Layers)

  • Input: Mel spectrograms of vocalizations
  • Classes: Healthy / Unhealthy / Noise
  • Burn Layers ensure noise resilience

Key Metrics:

  • βœ… Accuracy: 91.38%
  • 🚨 100% sensitivity for unhealthy class

πŸ–ΌοΈ Image Classification (EfficientNetB0)

  • Input: Chicken fecal images (160x160)
  • Model: Transfer learning with EfficientNetB0
  • Output: Multi-class disease prediction

πŸ’» Dashboard

Real-time web dashboard shows:

  • 🧠 Sound-based health indicators
  • 🧭 Zone-specific outbreak tracking
  • πŸ“ˆ Risk scoring and migration alerts
  • πŸ” Chicken count heatmaps

Dashboard UI


πŸ—£οΈ Pitch Video

πŸŽ₯ Watch Our Hackathon Pitch on YouTube


🌿 Sustainability Impact

  • 🌍 Lower environmental footprint
  • 🍳 More stable egg supply
  • πŸ’Ό Improved farmer livelihoods

πŸ›  Market Readiness

Feature Status
AI Model Training βœ… Completed
Real-time Monitoring βœ… Supported
Hardware Requirements βœ… Mic + Phone
Deployment Potential πŸš€ Field-ready

πŸ’° Financial Overview

Metric Value
Cost per Chicken $0.08/year
Potential Revenue $100M
Global Savings $2B
Farm Savings (US) $500M

πŸ† Awards

  • πŸ₯‡ First Prize – Data Science Track
    Cornell Digital Agriculture Hackathon 2025

πŸ‘₯ Team

  • Ahmed Abdulla
  • Farhan Mashrur
  • Suresh Kamath Bola
  • Kiyam Merali

πŸ“œ License

Educational Use License
This project is provided for educational and non-commercial use.
Commercial use requires written permission from the authors.

About

AI-powered early detection system for poultry diseases (1st Prize - Data Science, Cornell Hackathon 2025). It uses deep learning to analyze chicken vocalizations and fecal images, helping farmers identify signs of illness before outbreaks occur. πŸ›  Actively being developed with ongoing improvements to backend, models and dashboard interface.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published