This research integrates multimodal large language models with instance segmentation to create more intelligent and intuitive traffic monitoring systems for smart cities.

• Combines LLaVA visual-language capabilities with computer vision for comprehensive traffic analysis
• Enables real-time tracking of vehicles, traffic flow optimization, and congestion reduction
• Provides natural language interfaces for more intuitive traffic management operations
• Enhances road safety through improved monitoring and collision prevention

This engineering advancement represents a significant step toward more adaptive, responsive traffic control systems that can interpret complex urban environments and communicate insights in human-understandable formats.

Leveraging Multimodal-LLMs Assisted by Instance Segmentation for Intelligent Traffic Monitoring