Real-Time Visual Vessel Re-Identification for Maritime Surveillance on Edge Devices

Maritime surveillance plays a critical role in ensuring the security and safety of coastal regions, demanding reliable identification and tracking of vessels across spatially distributed camera networks. This task is particularly challenging as the system must identify a vessel even if it appears in different lighting conditions, angles, or environments. In this work, we propose a real-time visual vessel re-identification system optimized for deployment on edge devices. We evaluate our method on a custom maritime dataset and benchmark its performance on the NVIDIA Jetson Orin NX platform and a Raspberry Pi 5 with Hailo-8 accelerator. This work highlights the feasibility of deploying advanced Re-ID systems at the edge, enabling scalable, real-time maritime monitoring solutions.