🔍 UltraVision

UltraVision is a robotic assistant designed to help the visually impaired navigate their surroundings. The system integrates an ESP32-CAM, ultrasonic sensors, and Google Gemini AI to provide real-time scene descriptions and obstacle detection via a Flask web server.

🚀 Features

• ✅ Live video streaming from the ESP32-CAM
• ✅ AI-powered scene descriptions using Google Gemini API
• ✅ Ultrasonic sensor-based obstacle detection
• ✅ Real-time text-to-speech (TTS) using pyttsx3
• ✅ Web-based interface for image and audio playback

🛠️ Requirements

Hardware:

• ESP32-CAM module
• Ultrasonic sensors (HC-SR04 or similar)
• Speaker (optional for voice output)
• Power supply

Software:

• Python 3.7+
• Flask
• Google Generative AI
• pyttsx3 (TTS engine)
• dotenv (for managing API keys)

🔧 Installation & Setup

1️⃣ ESP32-CAM Setup

1. Open cameraWebServer.ino in Arduino IDE.
2. Update Wi-Fi credentials:

   const char* ssid     = "YOUR_WIFI_SSID";
   const char* password = "YOUR_WIFI_PASSWORD";
   const char* serverUrl = "http://YOUR_PC_IP:5000/upload";

3. Select AI Thinker ESP32-CAM as the board and upload the code.
4. Open the Serial Monitor to verify the connection.

2️⃣ Flask Server Setup

1. Clone the repository:

   git clone https://github.com/yourusername/VisionAssist.git
   cd VisionAssist/server

2. Set up the Python virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt

4. Create a .env file in the server directory and add your Google Gemini API key:

   GEMINI_API_KEY=your_gemini_api_key

5. Run the Flask server:

   python main.py

3️⃣ Accessing the System

Navigate to http://<your_server_ip>:5000 in a web browser to view the live stream and interact with the system.

The Flask server processes images and provides audio descriptions for users.

📂 File Structure

• cameraWebServer.ino → ESP32-CAM sketch for image capture & upload.
• main.py → Flask server for image processing and audio generation.
• server/.env → API keys (ignored by Git).
• server/requirements.txt → Python dependencies.

🔌 API Endpoints

• /upload → Upload an image from ESP32-CAM.
• /stream → Live video feed.
• / → Web UI for live streaming & audio playback.
• /audio → Serve the generated audio file.
• /process-image → AI-powered image processing and voice description.

🔮 Future Enhancements

• 🚀 Voice command support for hands-free interaction
• 🚀 Improved AI accuracy with fine-tuned models
• 🚀 GPS integration for enhanced navigation