The Third Eye

Reference number
Coordinator Lunds universitet - Matematikcentrum
Funding from Vinnova SEK 6 631 000
Project duration October 2018 - October 2022
Status Ongoing
Venture Machine Learning - FFI
Call 2017-03344-en

Purpose and goal

The project aims to build a prototype of a new sensor, a three-eye sensor and additional supportive software based on machine learning and computer vision. The purpose of this technique is to facilitate infrastructure-based information that can be used for traffic analysis, support for connected transport systems including autonomous vehicles, services for vulnerable road users, traffic research and urban planning in general. A common denominator for these tasks is the ability to detect and track road users like cars, pedestrians and bicycles, which will be the focus of the project.

Expected results and effects

The focus of the project is on new knowledge and techniques for using deep neural networks combined with a three-eye camera for traffic and transportation. The project is testing a new concept for a video-based sensor that uses three lenses ("three-eyed stereo camera"). The concept will lead to an industrial prototype and new machine learning methods, based on machine learning, will be developed to analyze road users (cars, bikes, pedestrians). Test results from real, but controlled, environments and exploration of use cases that benefit society are implemented.

Planned approach and implementation

The project consists of five work packages (WP). WP1 handled by Axis, WP2 managed by LTH Traffic and Roads, WP3 managed by LTH Mathematics, WP4 managed by LTH Traffic and Roads and WP5 handled by LTH Mathematics. WP1: Design, prototype and evaluation of three-eye sensor WP2: Data collection and traffic scenario planning WP3: Algorithm design with machine learning WP4: User cases for traffic planners and cities WP5: Project Management

The project description has been provided by the project members themselves and the text has not been looked at by our editors.