01658381

Component

Congestion in transport stations could result in stampede development and deadly crush situations. Closed circuit television (CCTV) cameras enable station managers to monitor the crowd and reduce overcrowding risks. However, identifying congestion conditions is a very laborious task for a human operator who has to monitor more than ten locations at the same time. Level of service (LOS) is the most widely accepted standard to measure congestion. Existing methods to measure LOS based on crowd density estimation from images have the disadvantages that, crowd density cannot be estimated accurately. In addition, the complex calculation process of flow parameters is not indicative of congestion in real-time. This paper proposes a novel method to directly classify LOS based on a convolutional neural network (CNN) and support vector machine (SVM) without calculating flow parameters, which can greatly simplify the measurement process. A second contribution of this research is to generate spatial-temporal LOS maps to visualize pedestrian distribution and variation of distribution in time. Experimental evaluation at Flinders Street Station in Melbourne shows that this method can achieve an accuracy of 80.6% in LOS classification using CCTV images.

This paper was sponsored by TRB committee AP015 Standing Committee on Transit Capacity and Quality of Service.

18-06670

English

11p

2018

Transportation Research Board 97th Annual Meeting

Digital/other

Figures; Photos; References; Tables

Closed circuit television; Crowds; Level of service; Mapping; Neural networks; Pedestrian density; Pedestrian flow; Rail transit stations; Real time information; Spatial analysis

Temporal analysis

Melbourne (Australia)

Pedestrians and Bicyclists; Planning and Forecasting; Public Transportation; Terminals and Facilities

Transportation Research Board Annual Meeting 2018 Paper #18-06670

TRIS, TRB, ATRI

Jan 8 2018 11:43AM