Research on Identification of Abnormal Behavior of Subway Passengers

As a vital component of urban public transportation, subway operations directly impact the safety of passengers' lives and property. Through analyzing extensive real-world cases and safety regulations, this study establishes a classification system for abnormal subway passenger behaviors and proposes a visual-based behavior recognition method. By utilizing yolov8 to develop algorithmic models that identify key behavioral characteristics, the method effectively distinguishes various actions. Experimental results demonstrate its exceptional performance in detecting common anomalies, particularly in identifying incidents such as falls and vandalism. These findings provide reliable technical support for subway safety monitoring, offering significant practical value for accident prevention and passenger protection.