Towards Robust Multi-Camera Object Tracking: A Framework for Global Coordination and ID Consistency

This study presents a structured framework for a scalable multi-camera tracking system based on ROS2, focusing on architectural design rather than experimental validation. The proposed system is designed to easily integrate different cameras and objects by utilizing the modular structure of ROS2 and can maintain reliable tracking performance even in complex environments. In particular, it adopts a global coordinate system transformation and an identity matching method based on the Hungarian algorithm to effectively maintain the consistency of identities between objects even in the presence of overlapping fields of view (FoV) and an increasing number of objects. This approach is shown to be an efficient and scalable solution for various multi-camera-based applications such as traffic monitoring, and autonomous driving.