An intelligent AI-based surveillance system built using Raspberry Pi, TensorFlow Lite, OpenCV, Flask, and Telegram for real-time object detection and automated security monitoring.
The system detects animals or intruders using computer vision, activates a buzzer alarm, and sends instant alerts with captured images to the user via Telegram.
Traditional CCTV systems require constant monitoring and expensive cloud services.
This project introduces a low-cost smart security solution capable of:
- Real-time object detection
- Automated alarm activation
- Instant mobile alerts
- Cloud image storage
- Live video monitoring
The system processes video locally on a Raspberry Pi, ensuring faster response and improved privacy.
The Raspberry Pi acts as the central processing unit that receives video input from the camera, processes it using TensorFlow Lite and OpenCV, and triggers alerts when an animal or intruder is detected.
- Raspberry Pi 3B+
- Raspberry Pi NoIR Camera
- Buzzer / Siren
- Power Supply
- SD Card
- Jumper Wires
The system continuously captures video frames and processes them using a lightweight object detection model.
When a target object is detected:
- The buzzer alarm is activated
- An image snapshot is captured
- The image is sent to the user via Telegram
- The event is stored for monitoring
The system successfully detects animals or intruders and sends alerts with timestamps.
- Real-time object detection using TensorFlow Lite
- Motion detection using OpenCV
- Telegram alerts with captured images
- Flask-based live video streaming
- Automatic buzzer activation
- Cloud backup for detected events
- Works in low-light conditions using NoIR camera
| Technology | Purpose |
|---|---|
| Python | Core programming language |
| TensorFlow Lite | Object detection model |
| OpenCV | Image processing and motion detection |
| Flask | Live video streaming |
| Telegram Bot API | User alerts |
| Raspberry Pi | Edge computing device |
smart-security-system-raspberrypi
│
├── code
│ └── smart_security_system.py
│
├── images
│ ├── detection_example.png
│ ├── hardware_setup.png
│ ├── real_time_operation.png
│ └── system_block_diagram.png
│
├── web
│ └── index.html
│
├── requirements.txt
├── model_download.txt
└── README.md
pip install -r requirements.txt
Follow the instructions in:
model_download.txt
Place the downloaded model inside:
tflite_model/
Run the main program:
python code/smart_security_system.py
The system will:
- Start camera monitoring
- Detect objects
- Trigger alerts if detection occurs
This system can be used for:
- Home security
- Farm and agricultural monitoring
- Wildlife monitoring
- Warehouse surveillance
- Industrial safety monitoring
This project was developed as a group capstone project during my diploma in Electronics and Communication Engineering.
While the project was completed collaboratively, I was primarily responsible for the technical implementation and system development, including:
- System architecture and workflow design
- Raspberry Pi setup and hardware integration
- Object detection implementation using TensorFlow Lite
- Image processing using OpenCV
- Telegram alert system integration
- Buzzer automation using Raspberry Pi GPIO
- Flask-based live video monitoring interface
- System testing, debugging, and deployment
Mahendra M
Electronics and Communication Engineering
Interested in Embedded Systems, Signal Processing, AI, and IoT systems
This project is shared for educational and research purposes.