The Third Eye
Reference number | |
Coordinator | Lunds universitet - Matematikcentrum |
Funding from Vinnova | SEK 6 631 000 |
Project duration | October 2018 - October 2022 |
Status | Completed |
End-of-project report | 2017-05510sv.pdf(pdf, 327 kB) (In Swedish) |
Important results from the project
The purpose of the project was to demonstrate the usefulness of the new sensor in traffic applications such as traffic safety assessment using traffic conflicts. But also to increase competence in deep learning methods for video analysis and create patents. The prototype, fulfills the intended goals we had for the project initially. During the project the term Traffic Surveillance Vision Pipeline (TVSP) was introduced with six components; video acquisition, camera calibration, object detection, 2D/3D mapping, tracking and trajectory analysis.
Expected long term effects
Several publications, with a machine learning or traffic analysis focus, have been presented at international conferences and workshops. This has led to international contacts and expanded networks. In connection with the publications, a patent has also been accepted and another is submitted as a proposal. Recording in a real, but controlled, environment has been carried out during development. The iterations with the prototype have increased the understanding of both what is technically possible and what the needs of intended customers are.
Approach and implementation
The project has basically followed initial planning and had regular meetings (which switched to the digital format during Corona). Even recording outdoors could be carried out with eased restrictions. The work with the camera prototype was carried out in three cycles. Each cycle consisted of a physical prototype being constructed by Axis which was then shown to the other partners who had to evaluate it and come up with comments and requests for changes. These wishes were then the basis for subsequent cycles.