Poaching Detection System using AI

Overview

This project focuses on developing an AI-powered system to detect poaching activities in real-time. The system leverages facial recognition to identify individuals involved in illegal hunting, using YOLO and OpenCV for object detection. The system includes a PIR sensor that when senses motion activates the Camera attached to Raspberry Pi and records a video of defined length and sends it to AWS EC2, where the code for object detection is stored.

The ultimate goal is to help conservation efforts by providing an automated system to monitor and prevent poaching in wildlife sanctuaries.

Features

• Real-Time Detection: Uses motion sensors to detect motion in the wildlife, and checks if the movement was caused by humans or animals.
• YOLO Integration: Utilizes YOLO (You Only Look Once) for fast and efficient object detection.
• Edge AI: Deploys on devices like Raspberry Pi to process data locally.
• Cloud Computing: Stores captured video and computes the detected code on an AWS EC2 instance.

Motivation

Poaching is a major threat to wildlife and has devastating effects on biodiversity and ecosystems. Current methods of poaching detection are often delayed and require manual intervention. This system aims to:

• Automate the detection process
• Provide real-time alerts
• Minimize response time for anti-poaching teams
• Enable long-term data storage and analysis of poaching activity

Hardware Requirements

To deploy the system in the field, the following hardware components are recommended:

• Raspberry Pi 4 as an edge device.
• Camera Module for capturing images in the field.
• PIR Sensor for detecting motion.
• Battery packs or solar power solutions for remote field deployment.

Software Architecture

The system is divided into several key components:

1. Camera Trap (Edge Device): Captures images in real-time and processes them locally.
2. YOLO & OpenCV Integration: YOLO is used for object detection, and OpenCV handles facial recognition.
3. Image Storage: Stores recognized faces and their metadata securely in a database.

Installation

Clone the Repository

git clone https://github.com/yourusername/poaching-detection-system.git
cd poaching-detection-system

Setting up the Environment

1. Install Python Dependencies

Ensure the following Python dependencies are installed by running:

pip install -r requirements.txt

Key dependencies include:

• opencv-python
• tensorflow
• torch
• pillow
• numpy

2. Download Pre-Trained YOLO Weights

Download YOLOv5 weights from this link and place them in the weights/ directory.

Setting up Raspberry Pi

1. Install TensorFlow Lite and OpenCV

Follow this guide to install TensorFlow Lite and OpenCV on your Raspberry Pi.

2. Connect PIR Sensor and Camera

• Connect and configure the PIR sensor to detect motion and trigger the camera to start recording.
• Ensure that the camera is properly set up and connected to the Raspberry Pi.

Running the Application

1. Run the Poaching Detection System on the Raspberry Pi

Once the PIR sensor detects motion, the camera records a video, which is then sent to an AWS EC2 instance for further processing.

python capture_video.py

2. Process the Video on AWS EC2

Run the object detection script on the AWS EC2 instance to process the video:

python detect_poachers.py

Configuration

Modify the config.json file to set up specific parameters for camera traps, video storage paths, and detection thresholds.

How It Works

1. Motion Detection and Recording: The PIR sensor triggers the camera on the Raspberry Pi, which starts recording upon detecting motion.

2. Video Transfer: The recorded video is automatically sent to an AWS EC2 instance for object detection.

3. YOLO Detection: The YOLO model on the EC2 instance processes the video, detecting human figures and marking them as potential poachers.

4. Facial Recognition: OpenCV processes detected faces for recognition and matches them against a known database of individuals.

5. Data Storage: Processed videos and metadata are securely stored on the EC2 instance for future analysis.

6. Alerts: Real-time alerts can be configured to notify conservation teams of detected poachers for rapid response.

Future Enhancements

• Improved Model Accuracy: Fine-tune the YOLO model with more training data specific to poaching activities.
• Real-Time Notifications: Integrate with AWS services for instant notifications to anti-poaching teams.
• Predictive Analysis: Incorporate analytics to forecast potential poaching hotspots.
• Expanded Sensor Integration: Integrate additional sensors to enhance environmental monitoring.

---

Contribution

Contributions are welcome! Please fork the repository, make your changes, and submit a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for more information.

Contact

For questions or suggestions, feel free to reach out at [email protected] .

---