v2.0.0
·
120 commits
to main
since this release
Blind Navigation - Tactile Paving Navigation Assistant System v2.0.0
🎉 What's New in Version 2.0.0
We're excited to announce the major release of Blind Navigation v2.0.0! This update brings significant architectural improvements and new features to enhance the navigation experience for visually impaired individuals.
🏗️ Major Architectural Refactor
Complete Backend Restructuring:
- 🔧 Modular Architecture: Split the monolithic
app.pyinto organized, maintainable modules - 📁 Structured Package Layout:
routes/- Dedicated route handlers for authentication, video processing, and map servicesservices/- Business logic layer for Baidu Map and AI integrationsutils/- Reusable utility functions and decoratorsmodels/- Database models and data access layer
- 🎯 Separation of Concerns: Clean separation between routes, business logic, and data access
- 🔒 Enhanced Security: Improved authentication and authorization mechanisms
🗺️ New AI-Powered Map Features
Baidu Map AI MCP Integration (Beta):
- 🤖 Intelligent Map Assistant: Initial integration with Baidu Map's AI capabilities
- 🗣️ Natural Language Processing: Ask questions about locations, routes, and nearby places
- 📍 Smart Location Queries: "Find coffee shops near me" or "How to get to the train station?"
⚠️ Note: This feature is in early beta and will be fully implemented in the next update
🚀 Key Features in v2.0.0
- 🎯 Real-time Tactile Paving Detection: Using pre-trained YOLOv8 model
- 🔊 AI-Personalized Voice Guidance: Ollama-powered contextual prompts
- 🤖 Intelligent Map Assistant: Baidu Map AI MCP integration (beta)
- 📍 Location Sharing: Real-time location tracking for family members
- 👤 Personalized Settings: Customizable voice, preferences, and user modes
- 🔐 Secure Authentication: Complete user registration and login system
🛠️ Installation & Upgrade
For New Users:
git clone https://github.com/wink-wink-wink555/blind_navigation.git
cd blind_navigation
pip install -r requirements.txtFor Existing Users:
- Backup your configuration files
- Pull the latest changes
- Run the application as usual - the new structure is backward compatible
📋 Requirements
- Python 3.8+
- MySQL Database
- Ollama with qwen2.5:3b model
- Baidu Map API Key (for map features)
- DeepSeek AI API Key (for intelligent assistant)
⚠️ Important Notes
- Baidu Map AI MCP is in beta phase - full functionality coming in next update
- Ensure Ollama service is running for personalized voice prompts
- Update your
config.pywith new configuration options if needed - The pre-trained YOLO model (
yolo/best.pt) remains unchanged and ready to use
🐛 Known Issues & Next Steps
- Baidu Map AI MCP integration requires additional refinement
- Some edge cases in real-time video processing being optimized
- Enhanced error handling for network-dependent features
- Next Update: Complete Baidu Map AI MCP implementation and performance enhancements
🙏 Acknowledgments
Special thanks to our contributors for their valuable feedback that helped shape this release.
Download: GitHub Repository
Documentation: README
Demo Video: Bilibili Demo
Blind Navigation - Empowering independence through AI-assisted navigation