Camera-based position analysis system for cyclists ordering in bicycle swarms

Camera-based position analysis system for cyclists ordering in bicycle swarms
Vemburaj Yadav, Alain Pagani, Didier Stricker
In: Workshop on Smart Urban Micromobility. Mensch und Computer (MuC-2023), September 3-6, Zürich, Switzerland, ACM, 2023.

Abstract:

Cycling in swarms has gained popularity as a social and fitness activity. To offer enhanced digital services for swarm cycling, it is essential to obtain real-time information about the position of each cyclist within the swarm. While GNSS (Global Navigation Satellite Systems) signals such as GPS, Galileo or GLONASS may not provide precise positioning in such scenarios, this paper proposes a novel approach to address this challenge. By equipping each bicycle with a backward-facing camera and leveraging computer vision and deep learning methodologies, we can achieve the absolute ordering of bicyclists in real-time. This position paper outlines a comprehensive framework that utilizes object detection, monocular depth estimation, and object tracking models to process camera information and obtain accurate positioning within the swarm. The proposed solution also enables the detection of overtakes between cyclists, adding an additional layer of information to enhance the overall swarm cycling experience.