1,000人乘客行为识别数据【数据堂】

1,000人乘客行为识别数据涵盖多年龄段、多时间段、多种光照。乘客行为包括乘客正常行为，乘客异常行为（乘客晕车行为，乘客困意行为，乘客遗落儿童&物品行为）。在采集设备方面，采用了可见光和红外双目摄像头。本套乘客行为识别数据可用于乘客行为分析等任务

The 1,000-person passenger behavior recognition dataset covers multiple age groups, various time periods, and diverse lighting conditions. Passenger behaviors included in this dataset consist of normal passenger behaviors and abnormal passenger behaviors, specifically motion sickness behaviors, drowsiness behaviors, and behaviors of leaving children or items behind. Visible light and infrared binocular cameras were adopted for data collection. This passenger behavior recognition dataset can be used for tasks such as passenger behavior analysis.

搜集汇总

背景与挑战

背景概述

该数据集包含1,000人的乘客行为识别视频数据，涵盖多年龄段、时间段和光照条件，采集于车内场景，使用可见光和红外双目摄像头。数据包括正常行为和异常行为（如晕车、困意、遗落儿童或物品），视频格式为MP4，标注准确率超过95%。