扶梯安全隐患行为智能分析数据

通过分析地铁、商城等扶梯监控视频数据，利用人工智能技术，进行智能分析，将视频中大件行李、婴儿车以及跌倒行为进行智能检测，对相关行为进行提醒已经提前应急处理（1）数据采集，通过rtsp对扶梯监控系统的上下摄像头进行拉流，
并可以分割为一帧一帧数据 ；（2）数据处理：对读取到的每一帧数据利用深度学习技术，进行分析，对婴儿车、大件行李、进行检测，对跌倒行为进行分析；（3）数据分析：对图像中的婴儿车、大件行李的检测结果进行分析，结合检测到的位置和扶梯相应的报警区域进行综合分析，进行语音播报予以提醒，对跌倒行为进行连续帧的分析，设定阈值，判定为跌倒行为，进行报警和扶梯急停的操作，对相关检测的发生时间和地点进行概率统计，计算事件发生的危险系数，公式为：事件危险指数=事件检测阈值*日期危险频率权重*时间危险频率权重*位置发生频率权重，做出报表分析；（4）数据应用：通过分析得出相关危险行为发生时间以及地点的规律，对重点日期和时间以及重点地点进行提前布控以及应急预案，可大大减少扶梯中的安全隐患事件的发生。

This dataset leverages artificial intelligence technologies to perform intelligent analysis by processing escalator surveillance video data collected from subways, shopping malls and other public scenarios, intelligently detects oversized luggage, strollers and fall behaviors in the videos, and issues reminders and implements early emergency responses for such risky behaviors.
(1) Data Collection: Pull video streams from the upper and lower cameras of the escalator monitoring system via RTSP, and split the streams into discrete frame data;
(2) Data Processing: Apply deep learning models to analyze each acquired frame, detect strollers and oversized luggage, and conduct fall behavior analysis;
(3) Data Analysis: Analyze the detection results of strollers and oversized luggage in the images, conduct comprehensive assessment by combining the detected positions with the corresponding alarm zones of the escalator, and trigger voice broadcast reminders; for fall behaviors, perform continuous frame analysis, set thresholds to confirm fall incidents, trigger alarms and escalator emergency stop operations; carry out probability statistics on the occurrence time and location of each detected event, calculate the event danger index using the formula: Event Danger Index = Event Detection Threshold * Date Danger Frequency Weight * Time Danger Frequency Weight * Location Occurrence Frequency Weight, and generate corresponding report analysis;
(4) Data Application: Derive the occurrence patterns of dangerous behaviors in terms of time and location through analysis, implement early deployment and emergency response plans for key dates, time periods and locations, which can significantly reduce the occurrence of safety hazard incidents on escalators.

搜集汇总

数据集介绍

特点

扶梯安全隐患行为智能分析数据集包含14611条记录，用于通过AI技术检测扶梯监控视频中的安全隐患行为（如大件行李、婴儿车及跌倒），并进行智能提醒和应急处理。数据集通过算法计算事件危险指数，帮助分析和预防扶梯安全隐患。