This project was developed as a mini-project for the Sensors course to assist visually impaired individuals using real-time object detection and voice feedback(Text-to-Speech). The system captures live video from a webcam or camera module from the device, detects nearby objects using YOLOv8, and announces their names and positions (left, center, right) through a Text-to-Speech (TTS) engine.

It demonstrates the integration of computer vision, speech synthesis, and real-time sensor-inspired systems — representing how technology can help enhance accessibility and independence for the blind.

Additionally, an ultrasonic sensor measures the distance to obstacles, and a vibration motor alerts the user through tactile feedback when objects are nearby.

Features

• Real-time object detection with YOLOv8
• Voice output using pyttsx3 (offline)
• Lightweight custom tracking based on IOU (Intersection over Union)
• Announces object name and direction (e.g., “person at left”)
• Multi-threaded Text-to-Speech system for smooth audio without lag

Project Structure

OBJECT-DETECTION/
│
├── main.py             # Main script (camera, detection, and logic)
├── Object_det.py       # ObjectTracker class (IOU-based tracking)
├── tts.py              # Threaded Text-to-Speech assistant
├── requirements.txt    # Dependencies
├── .gitignore
├── README.md
│
├── yolov8m.pt          # YOLOv8 model weights

Requirements

• Python 3.8+
• A working webcam or camera module
• (Optional) CUDA-compatible GPU for faster inference

Installation

1️⃣ Clone the repository

git clone https://github.com/<your-username>/OBJECT-DETECTION.git
cd OBJECT-DETECTION

2️⃣ Create and activate virtual environment (optional but recommended)

python -m venv .venv
# Windows
.venv\Scripts\activate
# Linux/Mac
source .venv/bin/activate

3️⃣ Install dependencies Create a file named requirements.txt and add:

ultralytics
opencv-python
cvzone
pyttsx3
torch

Then run:

pip install -r requirements.txt

4️⃣ Add YOLO model weights Ensure that yolov8m.pt is present in the same folder (as shown above). You can also download it using:

yolo download model=yolov8m.pt

Running the Project

Run the main file:

python main.py

The system will start capturing live video and announcing detected objects. Press q anytime to exit safely.

How It Works

Step	Component	Function
1️⃣	YOLOv8 (Object Detection)	Detects objects in each video frame
2️⃣	ObjectTracker (Object_det.py)	Tracks previously seen objects to prevent duplicate announcements
3️⃣	VoiceAssistant (tts.py)	Announces detected objects asynchronously using text-to-speech
4️⃣	Position Detection	Determines if an object is on the left, center, or right of the frame

Future Improvements

🔹 Integrate Google TTS or Coqui TTS for natural voice 🔹 Implement object tracking using DeepSORT 🔹 Port to Raspberry Pi for portable embedded use

📜 License

This project is released under the MIT License. Feel free to use, modify, and share for educational purposes.