eTraM
收藏资源简介:
eTraM是由亚利桑那州立大学开发的首个完全基于事件的交通监控数据集,旨在探索事件相机在静态交通监控中的应用潜力。该数据集包含10小时来自不同交通场景的数据,涵盖多种光照和天气条件,提供全面的现实世界情况概览。eTraM包含200万个边界框标注,覆盖八种不同的交通参与者类别,从车辆到行人和微型移动工具。数据集通过使用最先进的方法进行评估,展示了事件相机在交通监控中的强大潜力,为研究和应用开辟了新的途径。
Developed by Arizona State University, eTraM is the first event-based traffic monitoring dataset designed to explore the application potential of event cameras in static traffic monitoring. This dataset includes 10 hours of data collected across diverse traffic scenarios, encompassing a wide range of lighting and weather conditions to offer a comprehensive overview of real-world traffic environments. eTraM features 2 million bounding box annotations covering eight distinct traffic participant categories, ranging from vehicles and pedestrians to micro-mobility devices. Evaluated via state-of-the-art methodologies, the dataset demonstrates the strong potential of event cameras for traffic monitoring, opening up new avenues for both academic research and practical applications.

- 1eTraM: Event-based Traffic Monitoring Dataset亚利桑那州立大学 · 2024年



