数据堂—1,000人驾驶员行为识别数据

1,000人驾驶员行为识别数据涵盖多年龄段、多时间段、多种光照。驾驶行为包括危险驾驶行为、疲劳驾驶行为和视线偏移行为。在采集设备方面，采用了可见光和红外双目摄像头。本套驾驶员行为识别数据可用于驾驶员行为分析等任务

A 1,000-subject driver behavior recognition dataset covering multiple age groups, time periods and lighting conditions. The dataset includes three categories of driving behaviors: dangerous driving, fatigued driving and gaze deviation behaviors. Visible light and infrared binocular cameras were adopted for data collection. This driver behavior recognition dataset can be applied to tasks such as driver behavior analysis.

搜集汇总

背景与挑战

背景概述

该数据集包含1,000人的驾驶员行为识别数据，覆盖多种年龄、光照条件和行为类型，使用双摄像头采集，适用于驾驶员行为分析任务。数据准确率高，格式为.mp4视频。