Safety monitoring system for escalators at Buji Station on Shenzhen Metro Line 3

Project Background

 The trial and trial operation project of the escalator visual recognition monitoring system at Buji Station on Shenzhen Metro Line 3 is implemented by Jiangsu Bamaier Intelligent Technology Co., Ltd. The project was completed and commissioned on September 28, 2024, and passed the project acceptance inspection on December 19, 2025, obtaining the acceptance report from Shenzhen Metro Operation Group Co., Ltd. Operation Company 3. This project aims to enhance the safety monitoring level and emergency response efficiency of automatic escalators through advanced AI visual recognition technology, effectively preventing and reducing passenger injury incidents. A complete "end-edge-cloud" collaborative monitoring system was deployed on two key escalators (FT-04, FT-05) in the transfer corridor between Line 3 and Line 14 at Buji Station. The system integrates high-definition smart cameras, edge computing units, audible and visual alarms, and other equipment, achieving round-the-clock, real-time intelligent detection and analysis of passenger behaviors (such as falling, walking backwards, squatting, and lingering) and potential safety hazards (such as large luggage, baby carriages, and foreign objects on the comb plate). Once a risk is detected, the system immediately triggers on-site voice warnings and simultaneously displays a pop-up window and sounds an audible alarm on the large screen in the station control room, forming a rapid closed-loop early warning mechanism from the scene to the backend. During the nearly 15 months of trial operation, the project team worked closely with the Shenzhen Metro operation department to continuously optimize the algorithm model in response to the recognition challenges posed by the complex on-site environment (such as low ceilings and diverse luggage). The comprehensive recognition accuracy of the system has steadily increased from about 90% in the initial stage and stabilized at over 95%, demonstrating good adaptability and reliability. Practice has shown that the system significantly improves the station's response efficiency to escalator emergencies, enabling staff to intervene quickly, reassure passengers in a timely manner, and effectively avoid secondary injuries. At the same time, the system provides precise video playback functionality, greatly enhancing the efficiency of handling and tracing passenger injury incidents.


Achievements and Value

 This project successfully verified the practical value and promotion prospects of AI visual recognition technology in the field of public safety in rail transit. It provided innovative and efficient safety management tools for Shenzhen Metro operations, achieving the expected trial objectives.