NUDrive Lane Change Subset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/Ekim-Yurtsever/DeepTL-Lane-Change-Classification
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是位于日本名古屋的一辆装备有仪器的测试车辆所捕获的变道视频片段的子集，包含860个带有风险标注的视频片段。这些视频片段由十位标注员对变道行为的风险进行评估，其中风险实例与安全实例的比例为43:817。该数据集的规模为860个视频片段，其任务是对变道行为进行二分类，将其划分为风险与安全两个类别。
This dataset is a subset of lane-changing video clips captured by an instrumented test vehicle operating in Nagoya, Japan, and includes 860 video clips with risk annotations. Ten annotators evaluated the risk of each lane-changing behavior in these clips, resulting in a ratio of 43 risky instances to 817 safe instances. With a total of 860 video clips, the core task of this dataset is binary classification of lane-changing behavior, dividing instances into two categories: risky and safe.
提供机构：
NUDrive
