🐾 WildEye - AI-Powered Wildlife Monitoring System

WildEye is a real-time, AI-powered wildlife monitoring and early warning system designed to mitigate human-wildlife conflicts. By employing state-of-the-art computer vision (YOLOv8) and a synchronized web stack, the system identifies dangerous/endangered animals, logs sightings, and instantly alerts field rangers via email dispatch.

---

🌟 Key Features

• Real-Time AI Animal Detection: Custom-trained YOLOv8 computer vision pipeline detecting target species (elephants, bears, leopards, tigers, etc.) with configurable confidence thresholds.
• Live Camera Stream: Smooth, low-latency live camera streaming directly on the dashboard.
• Secured JWT Admin Portal: Protected control panel featuring a modern login/signup layer using JSON Web Tokens (JWT) and bcryptjs password hashing.
• Automated Dispatch System: Nodemailer-driven SMTP engine that emails instant alerts to enabled forest officers upon high-confidence detections.
• Dynamic Analytics & Custom Reports: Automatically compiles historical database entries into dynamic overview statistics (Total Detections, threatened species ratio, active hours) and exports daily, weekly, or monthly logs.

---

⚙️ System Flow

graph TD
    A[IP Camera / Video Feed] -->|Frames| B(Python YOLOv8 Model)
    B -->|Detects Target Animal| C{Backend API Server}
    B -->|Writes annotated frame| D[latest_frame.jpg]
    C -->|1. Saves Sighting| E[(MongoDB Atlas)]
    C -->|2. Broadcasts event| F[Socket.io client]
    C -->|3. Triggers email| G[Nodemailer SMTP]
    F -->|Real-time alert| H[React Dashboard]
    D -->|Poll & Render| H

Loading

---

📊 Dataset

We trained our YOLOv8 model using a custom dataset hosted on Roboflow Universe:

How to Download the Dataset

If you want to retrain the model locally, install the Roboflow Python package:

pip install roboflow

Then download the dataset in YOLOv8 format:

from roboflow import Roboflow
rf = Roboflow(api_key="YOUR_ROBOFLOW_API_KEY")
project = rf.workspace("tejas-eurut").project("animal-detection-85p4k-qmaea")
version = project.version(2)
dataset = version.download("yolov8")

---

📂 Project Structure

WildEye_Project(EDI)/
├── ML_model/
│   ├── wildeye_test.py       # YOLOv8 target detection and frame streaming script
│   └── yolov8n.pt            # Pre-trained model weights
├── backend/
│   ├── server.js             # Node.js / Express API & Socket.io server
│   └── .env                  # Port, database URL, and SMTP key parameters
└── frontend/
    ├── src/
    │   ├── components/       # App pages: Home, LiveMonitor, Reports, Settings, Login
    │   ├── App.js            # Stateful JWT routes & path guards
    │   └── index.js          # React entry point
    └── package.json          # React libraries

---

🛠️ Setup & Installation

Prerequisites

• Node.js (v16+)
• Python (3.8+)
• MongoDB Atlas account (or local MongoDB database)

---

Step 1: Backend Setup

1. Navigate to the parent directory containing the backend dependencies:

   cd backend

2. Create or open the .env file and configure your credentials:

   DATABASE_URL=mongodb+srv://<username>:<password>@cluster.mongodb.net/WildeyeDB?retryWrites=true&w=majority
   EMAIL_USER=your_gmail@gmail.com
   EMAIL_PASS=your_16_character_app_password
   JWT_SECRET=your_secure_jwt_secret_key

   Note: For EMAIL_PASS, you must generate a 16-character Google App Password.

---

Step 2: Frontend Setup

1. Navigate to the frontend directory:

   cd ../frontend

2. Install the frontend dependencies:

   npm install

---

Step 3: Machine Learning Model Setup

1. Install Ultralytics YOLOv8, OpenCV, and Requests:

   pip install ultralytics opencv-python requests numpy

2. Open ML_model/wildeye_test.py and verify these paths match your files:
   • video_path: Path to your input video feed or camera device index.
   • model_path: Path to your trained weights (best.pt).

---

🚀 Running the System

Open three separate terminal instances to run the complete WildEye pipeline:

1. Launch the Backend Server

cd backend
node server.js

Console output should show: SUCCESS: Successfully connected to MongoDB Atlas!

2. Launch the React Web Portal

cd frontend
npm start

This will open the web interface at http://localhost:3000.

3. Start the Object Detection Loop

cd ML_model
python wildeye_test.py

A local OpenCV window will launch displaying the annotated video. Detections will instantly print to the backend and update the web portal.

---

🔒 Security & Roles

• Admins/Control Room Operators: Register an account on the sign-up screen, log in, and monitor sightings, customize active forest officer notifications, and download analytics.
• Field Officers: Registered inside Settings. They do not need to register accounts or access the dashboard. The system automatically routes instant alerts to their registered email addresses.

---