A real-time object detection system that leverages YOLOv5 to identify specific objects (like people, cats, dogs, and cars) from a live camera feed and sends instant notifications, along with a captured image, to a Telegram chat. The system also includes a Telegram bot that can receive images from users.
- Real-time Object Detection: Continuously monitors a live camera feed for predefined objects.
- YOLOv5 Integration: Utilizes the powerful YOLOv5 model for accurate and fast object recognition.
- Configurable Target Classes: Easily specify which objects you want to detect (e.g., person, cat, dog, car).
- Intelligent Notification System: Sends Telegram notifications only after a specified interval to prevent spam.
- Image Capture & Sharing: Saves detected object frames with bounding boxes and sends them directly to your Telegram chat.
- Telegram Bot Interaction: A Telegram bot that can receive images from users and provide a simple acknowledgment.
- Easy Setup: Minimal configuration required to get started.
| Technology | Description | Icon |
|---|---|---|
| Python | The core programming language for the application. |  |
| OpenCV (cv2) | Library for computer vision and image processing. |  |
| PyTorch (torch) | Deep learning framework for building neural networks. |  |
| YOLOv5 (Ultralytics) | State-of-the-art object detection model. |  |
| python-telegram-bot | Library for interacting with the Telegram Bot API. |  |
| python-dotenv | Tool for managing environment variables. |  |
The system operates with a straightforward architecture, designed for real-time processing and efficient notification.
- Camera Feed: The system continuously captures frames from a connected webcam.
- Object Detection (YOLOv5): Each captured frame is fed into the pre-trained YOLOv5 model.
- Filtering & Analysis: The model's detections are filtered to identify only the
TARGET_CLASSES. A time-based debounce mechanism prevents excessive notifications. - Image Capture: If a target object is detected and the notification interval has passed, the frame (with bounding boxes drawn) is saved locally.
- Telegram Notification: A message, along with the captured image, is sent to your configured Telegram chat ID.
- Telegram Bot Listener: Separately, the Telegram bot listens for incoming messages, specifically photos, and responds accordingly.
Follow these steps to get your Smart Surveillance system up and running.
- Python 3.8+
- A webcam connected to your system.
- A Telegram account.
git clone https://github.com/your-username/smart-surveillance.git
cd smart-surveillancepython -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activatepip install -r requirements.txt(You'll need to create a requirements.txt file based on the provided code. It should contain: opencv-python, torch, ultralytics/yolov5, python-telegram-bot, python-dotenv.)
Create a .env file in the root directory of your project:
TELEGRAM_BOT_TOKEN="YOUR_TELEGRAM_BOT_TOKEN"
CHAT_ID="YOUR_TELEGRAM_CHAT_ID"- Open Telegram and search for @BotFather.
- Start a chat with BotFather and send /newbot.
- Follow the instructions to choose a name and username for your bot.
- BotFather will provide you with an API Token. Copy this token and paste it into your .env file.
- Start a conversation with your newly created bot in Telegram.
- Send any message to your bot.
- Open your web browser and go to:
https://api.telegram.org/bot<YOUR_TELEGRAM_BOT_TOKEN>/getUpdates- Look for the chat object in the JSON response. The id field within this object is your CHAT_ID. Copy this ID and paste it into your .env file.
Once everything is set up, you can run the main script:
python smartBot. py- A window titled "Object Detection" will appear, displaying the live camera feed with bounding boxes around detected objects.
- When a TARGET_CLASS object is detected, and the DETECTION_INTERVAL has passed, a notification and an image will be sent to your Telegram chat.
- You can press
qon the "Object Detection" window to quit the application. - You can send images to your Telegram bot, and it will acknowledge receipt.
You can easily customize the detection behavior by modifying the following variables in the script:
- TARGET_CLASSES: A list of strings representing the object classes you want to detect.
TARGET_CLASSES = ['person', 'cat', 'dog', 'car']- DETECTION_INTERVAL: The minimum time (in seconds) between sending notifications for new detections. This prevents notification spam.
DETECTION_INTERVAL = 5 # seconds- Camera Selection: If you have multiple cameras, you might need to change
cap = cv2.VideoCapture(0)to1,2, etc., to select a different camera.
- Customizable Notification Messages: Allow users to define their own notification message templates.
- Multiple Camera Support: Extend to monitor multiple camera feeds simultaneously.
- Web Interface: Develop a simple web interface for configuration and viewing detection history.
- Cloud Storage: Option to upload captured images to cloud storage (e.g., Google Cloud Storage, AWS S3) instead of local storage.
- Advanced Alerting: Integrate with other notification services (e.g., email, SMS).
- Time-based Scheduling: Allow users to schedule when the detection system should be active.
- Motion Detection Pre-filter: Use basic motion detection to trigger YOLOv5 inference only when movement is detected, saving resources.
Contributions are welcome! If you have suggestions, bug reports, or want to contribute code, please feel free to open an issue or submit a pull request.
This project is licensed under the MIT License - a permissive license that allows for flexible usage, modification, and distribution. For more details, check out the full LICENSE file.
I love collaborating, sharing knowledge, and connecting with like-minded individuals. Feel free to reach out to me through any of the following platforms:
- 🌟 LinkedIn – Connect with me professionally and view my latest career updates.
- 🐦 Twitter – Follow me for updates, tech insights, and occasional musings.
- 🧑💻 GitHub – Explore my repositories, open-source projects, and contributions to the community.
Let's grow together! 🚀