PeSOTIF
收藏资源简介:
PeSOTIF数据集是由清华大学汽车安全与能源国家重点实验室创建,专注于自动驾驶系统在长尾交通场景中的感知安全问题。该数据集包含1126帧数据,涵盖多种极端天气和光照条件,以及复杂的交通参与者行为。数据集的创建过程涉及从多个来源收集关键帧,并手动标注关键对象。PeSOTIF数据集主要用于评估和改进自动驾驶系统在不确定环境下的感知性能,旨在解决感知系统在复杂场景下的安全性和可靠性问题。
The PeSOTIF dataset was created by the State Key Laboratory of Automotive Safety and Energy, Tsinghua University, focusing on the perception safety issues of autonomous driving systems in long-tail traffic scenarios. This dataset contains 1,126 frames of data, covering various extreme weather and lighting conditions, as well as complex behaviors of traffic participants. The creation process of the dataset involves collecting key frames from multiple sources and manually annotating key objects. The PeSOTIF dataset is mainly used to evaluate and improve the perception performance of autonomous driving systems in uncertain environments, aiming to solve the safety and reliability issues of perception systems in complex scenarios.

- 1PeSOTIF: a Challenging Visual Dataset for Perception SOTIF Problems in Long-tail Traffic Scenarios清华大学汽车安全与能源国家重点实验室 · 2022年



