5G Enabled Video Analytics for Detecting Container Seals in Port Operations

8th International Physical Internet Conference
Author: P. Basaras (ICCS), K.V. Katsaros (ICCS), G. Kanellopoulos (ICCS), S. Tsagalas (Piraeus Container Terminal S.A), A. J. Amditis (ICCS)
This article is focused on port control, logistics and remote automation, and aims at detecting the presence/absence of seals at cargo containers. The proposed use case is realized by developing a novel computer vision algorithm for the task at hand, enabled as far-edge computing services on board 5G connected Internet of Things (IoT) devices. The proposed service adopts key enabling technologies of the 5G ecosystem such Network Functions Virtualization (NFV) with MAnagement and Network Orchestration (MANO) support for (automated) lifecycle management of the various service components. The service and overall architecture can be deployed on commodity servers to facilitate interoperability across heterogeneous ports, at low cost. Our results, based on real datasets obtained from the Piraeus port operations illustrate that the subject computer vision approach can achieve up to about 94% accuracy for detecting the presence of absence of container seals.