Research Article
Person Tracking in Heavy Industry Environments with Camera Images
@INPROCEEDINGS{10.1007/978-3-030-51005-3_27, author={Nico Zengeler and Alexander Arntz and Dustin Ke\`{a}ler and Matthias Grimm and Ziyaad Qasem and Marc Jansen and Sabrina Eimler and Uwe Handmann}, title={Person Tracking in Heavy Industry Environments with Camera Images}, proceedings={Science and Technologies for Smart Cities. 5th EAI International Summit, SmartCity360, Braga, Portugal, December 4-6, 2019, Proceedings}, proceedings_a={SMARTCITY}, year={2020}, month={7}, keywords={Heavy industry Industry 4.0 Person tracking Artificial intelligence Image processing}, doi={10.1007/978-3-030-51005-3_27} }- Nico Zengeler
Alexander Arntz
Dustin Keßler
Matthias Grimm
Ziyaad Qasem
Marc Jansen
Sabrina Eimler
Uwe Handmann
Year: 2020
Person Tracking in Heavy Industry Environments with Camera Images
SMARTCITY
Springer
DOI: 10.1007/978-3-030-51005-3_27
Abstract
In this paper, we propose a method to localise and track persons in heavy industry environments with multiple cameras. Using the OpenPose network, we localise the persons feet points on each cameras image individually and perform according 3D transformations. With prior knowledge about the camera settings in the environment, we use a rule-based system to assess which sensor detections to fuse. We then apply Kalman filtering in order to stabilise the tracking. Due to a variable image stack size, our method may increase accuracy if provided with additional computational resources by processing more frames in real-time. We have simulated a heavy industry scenario and use the recorded video material and position data as a basis for our evaluation.
