This talk summarizes our recent efforts in extracting high-precision motion information from camera images. We present our methodology which is developed to extract 3D vehicle kinematic information from roadside cameras using deep learning. Ground truth data are collected with the help of uncrewed aerial vehicles (UAVs) in terms of top view bounding boxes, providing vehicle position, size, orientation, and velocity information with high precision. These can be converted to roadside view bounding boxes using homography transformation. The ground truth data and the roadside view images are used to train a modified YOLO neural network, and thus, to learn the homography transformation matrix. The output of the neural network is high-precision vehicle kinematic information which can be visualized in both the top view and the roadside view. Once the neural network is trained, only the roadside cameras are needed to extract the kinematic information.

More on this research: https://ccat.umtri.umich.edu/research/u-m/generating-high-accuracy-transportation-datasets-with-unmanned-aerial-vehicles/
---
About the speaker: Gabor Orosz received an MSc degree in Engineering Physics from the Budapest University of Technology, Hungary, in 2002, and a PhD degree in Engineering Mathematics from the University of Bristol, UK, in 2006. He held postdoctoral positions at the University of Exeter, UK, and at the University of California, Santa Barbara. In 2010, he joined the University of Michigan, Ann Arbor where he is currently a Professor in Mechanical Engineering and in Civil and Environmental Engineering. From 2017 to 2018 he was a Visiting Professor in Control and Dynamical Systems at the California Institute of Technology. In 2022 he was a Distinguished Guest Researcher in Applied Mechanics at the Budapest University of Technology and from 2023 to 2024 he was a Fulbright Scholar at the same institution. He served as an associate editor for Transportation Research Part C from 2018 to 2023. He has been serving as an associate editor for the IEEE Transactions on Control Systems Technology since 2021, and for the IEEE Transactions on Intelligent Transportation Systems since 2022. He served as the general chair for the 17th IFAC Workshop on Time Delay System and for the 3rd IAVSD Workshop on Dynamics of Road Vehicles, Connected and Automated Vehicles. His research interests include nonlinear dynamics and control, time delay systems, machine learning and data-driven systems, with applications to connected and automated vehicles, and traffic flow.