This project targets approximately 4,000 CCTVs installed on expressways nationwide, combining AI inference infrastructure and video analysis platforms to build an intelligent control system that automatically detects emergency situations such as traffic accidents, stopped vehicles, wrong-way driving, and pedestrians.

The technological core of this project is the linkage structure between Vision AI and VLM (Vision Language Model).

Vision AI conducts first-stage detection of object movement and abnormal situations, while VLM comprehensively analyzes surrounding scenes and road environment to verify whether an actual emergency situation has occurred.

The system transitions from existing visual inspection-focused control to a structure where AI automatically detects abnormal situations and provides information on an event-driven basis, according to the company.

Mobilint supplies AI servers and its own AI accelerator 'MLA400' for this project, taking responsibility for real-time processing of large-scale CCTV video streams.

With a low-power, high-efficiency structure, it functions as AI inference infrastructure that stably processes video data from thousands of CCTVs.

The system can interoperate with legacy equipment such as VMS (Video Management System) and NVR (Network Video Recorder), allowing AI functionality to be introduced incrementally without large-scale CCTV replacement, according to the company.

It is expected to support AI transformation for public institutions while reducing construction costs and timeframes.

The two companies previously conducted joint technology verification on a wildfire monitoring AI project based on domestic AI semiconductors, and this project expands that cooperation to the nationwide traffic safety sector.

Mobilint announced plans to expand its business scope into various public AI video analysis fields including smart cities, disaster safety, and industrial safety.

Kim Seong-mo, Executive Director of Business Development at Mobilint, stated, "This project is a case where domestic AI semiconductors are actually applied to public AI infrastructure to implement large-scale AI control services," adding, "We will expand the use of domestic AI semiconductors in various public sectors including traffic, disaster safety, and smart cities."