Automation of Pedestrian Tracking in a Crowded Situation

Saadat, Saman; Teknomo, Kardi

doi:10.1007/978-1-4419-9725-8_21

Cited by 8 publications

(2 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, for a high-density crowd, optical systems fail because the lines of sight between camera and the body parts to be tracked are hidden by other persons or body parts. Therefore, most experimental data provided for analysis and modeling pedestrian dynamics are limited to trajectories of the head of every single pedestrian [ 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ]. Only some datasets include additional global (e.g., distribution of gender or age) or individual (e.g., body size, and head or shoulder orientation) information [ 32 ].…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Tracking System of Full-Body Motion Inside Crowds

Boltes

Adrian

Raytarowski

2021

Sensors

View full text Add to dashboard Cite

For our understanding of the dynamics inside crowds, reliable empirical data are needed, which could enable increases in safety and comfort for pedestrians and the design of models reflecting the real dynamics. A well-calibrated camera system can extract absolute head position with high accuracy. The inclusion of inertial sensors or even self-contained full-body motion capturing systems allows the relative tracking of invisible people or body parts or capturing the locomotion of the whole body even in dense crowds. The newly introduced hybrid system maps the trajectory of the top of the head coming from a full-body motion tracking system to the head trajectory of a camera system in global space. The fused data enable the analysis of possible correlations of all observables. In this paper we present an experiment of people passing though a bottleneck and show by example the influences of bottleneck width and motivation on the overall movement, velocity, stepping locomotion and rotation of the pelvis. The hybrid tracking system opens up new possibilities for analyzing pedestrian dynamics inside crowds, such as the space requirement while passing through a bottleneck. The system allows linking any body motion to characteristics describing the situation of a person inside a crowd, such as the density or movements of other participants nearby.

show abstract

Section: Introductionmentioning

confidence: 99%

A Hybrid Tracking System of Full-Body Motion Inside Crowds

Boltes

Adrian

Raytarowski

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…The camera for each of these experiments was not calibrated. This paper is a substantial extension of the previous work by [12] in terms of describing the feature selection and transformation as well as in incorporating some revisions of the tracking algorithm. We describe our system to automate such tracking.…”

mentioning

confidence: 96%

Automation of Tracking Trajectories in a Crowded Situation

2010

View full text Add to dashboard Cite

Studies on pedestrians using microscopic simulation require large amounts of trajectory data from real-world pedestrian crowds. The collection of such data, if done manually, involves tremendous efforts and is very time-consuming. Although many studies have asserted the possibility of automating this task using video cameras, we have found that only a few have demonstrated good performance in very crowded situations or from a top-angled view scene. This paper deals with tracking pedestrian crowd under heavy occlusion and from an angular scene using only a single non-stereo video camera. Our automated tracking system consists of three modules that are performed sequentially. The first module detects moving objects as blobs. The second module computes feature values from the blob information in order to generate what we call a possibility matrix. The third module is a tracking system, which employs a Bayesian update of the probability tree derived from the possibility matrix and from the detection of each pedestrian, in order to track the next position of the pedestrian. The result of such tracking is a database of pedestrian trajectories over time and space. With certain prior information, we show that the system is able to track a large number of people under occlusion and clutter scene.

show abstract