Hub Station Stall and Obstruction Behavior Anomaly Detection Dataset

The current transportation sector faces issues with urban traffic disorder, especially at busy hub stations where temporary obstructions like small stalls and parasols severely impact pedestrian and vehicle passage. Existing surveillance systems are mostly static and cannot recognize and handle these dynamic abnormal behaviors in real-time, facing challenges such as low recognition efficiency and slow response. This dataset aims to provide high-quality video data to help researchers and developers enhance their capabilities to detect traffic anomalies, thereby improving urban traffic management. Data collection uses high-definition camera equipment, covering various time periods and weather conditions to ensure data diversity and representativeness. To ensure data quality, we conducted multiple rounds of annotation, consistency checks, and expert reviews to ensure the accuracy of each video and annotation. Data is stored in MP4 format for easy post-processing and analysis, organized by time and location for quick retrieval and application.