PINNS
收藏资源简介:
PINNS是由上海交通大学团队构建的一个专注于非结构化场景下行人-车辆交互的轨迹预测数据集。该数据集覆盖多个国家和地区,包含多样化的典型交通场景,并考虑了季节、光照条件和天气的变化,数据来源于未经校准的监控摄像头视频。数据集构建遵循中国自动化学会标准,通过提出的标注框架从原始视频中提取并注释轨迹数据与场景级信息。其核心应用领域是自动驾驶和轨迹预测研究,旨在为解决复杂混合交通场景中异构智能体(行人-车辆)的交互建模与安全预测提供关键数据支持。
PINNS is a trajectory prediction dataset focusing on pedestrian-vehicle interaction in unstructured scenarios, constructed by a team from Shanghai Jiao Tong University. This dataset covers multiple countries and regions, includes diverse typical traffic scenarios, and considers variations in seasons, lighting conditions and weather. The data is sourced from uncalibrated surveillance camera videos. The dataset is developed in compliance with the standards of the Chinese Association of Automation, and trajectory data and scene-level information are extracted and annotated from raw videos through the proposed annotation framework. Its core application areas are autonomous driving and trajectory prediction research, aiming to provide critical data support for solving the interaction modeling and safety prediction of heterogeneous agents (pedestrian-vehicle) in complex mixed traffic scenarios.




