Traffic Light Failure and Anomaly Detection Dataset

The current transportation industry faces challenges in identifying failures and abnormal states of traffic lights. Traditional monitoring methods are difficult to detect signal light status in real-time and accurately, leading to frequent traffic accidents. Existing solutions mostly rely on manual inspection, which is inefficient and prone to errors. This dataset aims to automatically identify abnormal states such as traffic light outages, prolonged flashing, and conflicting signals through video analysis technology, thereby enhancing traffic safety and management efficiency. Data collection is conducted using high-definition cameras at traffic intersections in different cities, recording continuously for 24 hours to ensure diverse scenes and conditions. For quality control, multiple rounds of annotation and consistency checks are employed to ensure accuracy and completeness of the annotations. The data is stored in MP4 format, facilitating large-scale data processing and analysis.