Gyu-Sung Cho

Real-time spatial coordination in high-density seaport container yards is a bottleneck in global supply chains. In these environments, GPS signal denial and severe visual occlusion render traditional theoretical scheduling models ineffective. To resolve this operational gap, this paper introduces HY-EnYO, a novel dual-stream ensemble architecture designed to automate the identification of vacant parking bays using monocular CCTV footage without requiring physical sensor infrastructure. Addressing a fundamental limitation of standard single-stream deep learning detectors, which suffer from severe cross-class feature confusion when mapping massive infrastructure alongside dynamic assets within a unified latent space, the proposed framework decouples visual extraction into parallel sub-networks specialized for yard cranes ( M c ) and terminal trucks ( M t ). While this feature extraction remains probabilistic, final occupancy state reasoning is resolved by a deterministic logic engine. This engine projects perspective-aware parking polygons and evaluates Intersection over Union (IoU) overlap against a spatial threshold ( τ = 0.08 ), mathematically counteracting camera perspective distortion and sensor jitter. Compared to standard single-stream baselines, HY-EnYO reduces the computational parameter footprint by over 67% (from 36.49 M to 12.02 M), requires a peak VRAM allocation of only 0.10 GB, and achieves an inference latency of 21.82 ms, ensuring edge-device feasibility. Extensive evaluation on real-world CCTV datasets from the Busan seaport demonstrates peak detection accuracies (maP@0.5) of 0.899 for trucks and 0.999 for cranes. Evaluated across a 316-instance test sequence, the deterministic spatial engine achieves a perfect occupancy precision of 1.000 (preventing false-positive routing deadlocks) and an overall system accuracy of 98.0%, maintaining high operational accuracies of 97.0% and 98.0% under simulated adverse weather and low-illumination conditions, respectively. The framework provides a mathematically rigorous, highly efficient spatial matrix, establishing a data-driven foundation for automated seaport traffic coordination.

This study compared the technical efficiency of container ports in Vietnam to pinpoint areas for improvement, enabling the ports to enhance their efficiency and competitiveness within the Vietnamese port industry. This analysis used cross-sectional data from 13 major container ports and terminals across the country, utilizing the variables related to port infrastructure, including the total quay length (in meters), maximum alongside depth (in meters), and total terminal area (in square meters). The output metric used was the cargo volume, measured in twenty-foot equivalent units (TEU) per year. The port efficiency was measured using stochastic frontier analysis (SFA), with the dataset sourced primarily from the official Vietnam Seaports Association website (www.vpa.org.vn). The methods were used to analyze the variations in port efficiency, including aspects such as distribution, rankings, identification of the best and worst ports, and determinants of the efficiency frontier. The results suggest that the port efficiency, as measured using the SFA models, enhances the technical efficiency scores of the container ports, with the collected scores generally being high among the 13 observations. This highlights the significance of the term error in container port production, showing that the seaport sector is an area where perception errors are highly developed. In addition, ports with higher throughput do not always have the highest efficiency scores. Therefore, the port authorities and managers responsible for port investment and management in Vietnam should consider using the SFA method to estimate and improve port efficiency.

Ensuring the integrity of shipping containers is crucial for maintaining product quality, logistics efficiency, and safety in the global supply chain. Damaged containers can lead to significant economic losses, delays, and safety hazards. Traditionally, container inspections have been manual, which are labor-intensive, time-consuming, and error-prone, especially in busy port environments. This study introduces an automated solution using the YOLO-NAS model, a cutting-edge deep learning architecture known for its adaptability, computational efficiency, and high accuracy in object detection tasks. Our research is among the first to apply YOLO-NAS to container damage detection, addressing the complex conditions of seaports and optimizing for high-speed, high-accuracy performance essential for port logistics. Our method showcases YOLO-NAS's superior efficacy in detecting container damage, achieving a mean average precision (mAP) of 91.2%, a precision rate of 92.4%, and a recall of 84.1%. Comparative analyses indicate that YOLO-NAS consistently outperforms other leading models like YOLOv8 and Roboflow 3.0, which showed lower mAP, precision, and recall values under similar conditions. Additionally, while models such as Fmask-RCNN and MobileNetV2 exhibit high training accuracy, they lack the real-time assessment capabilities critical for port applications, making YOLO-NAS a more suitable choice. The successful integration of YOLO-NAS for automated container damage detection has significant implications for the logistics industry, enhancing port operations with reliable, real-time inspection solutions that can seamlessly integrate into predictive maintenance and monitoring systems. This approach reduces operational costs, improves safety, and lessens the reliance on manual inspections, contributing to the development of “smart ports” with higher efficiency and sustainability in container management.

The rapid urbanization phenomenon has introduced multifaceted challenges across various domains, including housing, transportation, education, health, and the economy. This necessitates a significant transformation of seaport operations in order to optimize smart mobility and facilitate the evolution of intelligent cities. This conceptual paper presents a novel mathematical framework rooted in deep learning techniques. Our innovative model accurately identifies parking spaces and lanes in seaport environments based on crane positions, utilizing live Closed-Circuit Television (CCTV) camera data for real-time monitoring and efficient parking space allocation. Through a comprehensive literature review, we explore the advantages of merging artificial intelligence (AI) and computer vision (CV) technologies in parking facility management. Our framework focuses on enhancing container drayage efficiency within seaports, emphasizing improved traffic management, optimizing parking space allocation, and streamlining container movement. The insights from our study provide a foundation that could have potential implications for real-world applications. By integrating cutting-edge technologies, our proposed framework not only enhances the efficiency of seaport operations, but also lays the foundation for sustainable and intelligent seaport systems. It signifies a significant leap toward the realization of intelligent seaport operations, contributing profoundly to the advancement of urban logistics and transportation networks. Future research endeavors will concentrate on the practical implementation and validation of this pioneering mathematical framework in real-world seaport environments. Additionally, our work emphasizes the crucial need to explore further applications of AI and CV technologies in seaport logistics, adapting the framework to address the evolving urbanization and transportation challenges. These efforts will foster continuous advancements in the field, shaping the future of intelligent seaport operations.

Urban expansion has ushered in a landscape of opportunities and challenges across housing, transportation, education, health, and the economy. In response to these evolving dynamics, the application of artificial intelligence (AI) and computer vision (CV) technologies has emerged as a transformative solution. While smart traffic monitoring and advanced parking distribution systems have eased urban pressures, optimizing mobility remains pivotal in the context of burgeoning smart cities. However, the seaport industry grapples with formidable issues in the efficient transport of containers. In light of this, the fusion of AI and CV technology holds promise as a solution. This study pioneers a systematic review, representing a novel exploration, delving into a comprehensive evaluation of the existing literature. It scrutinizes the profound advantages AI and CV offer in constructing sustainable, efficient parking ecosystems within seaports. Our methodological approach encompasses data collection, rigorous quality assessment, and meticulous exploration of the application of CV and AI in the realm of smart parking management. The findings underscore the pivotal role of AI and CV technologies in the development of efficient, sustainable transportation systems, particularly for optimizing container movement within seaports. This research presents a comprehensive analysis of the literature in the area of the application of AI and CV technologies in optimizing parking management at seaports, shedding light on the potential for sustainable transportation solutions in this critical domain. As these technologies usher in enhancements in traffic management, parking space allocation, and container logistics within seaports, this study represents a vital and timely contribution to the field, serving as a cornerstone for future innovations in seaport management and the broader context of smart cities.

Ports are important hubs in logistics and supply chain systems, where the majority of the available data is still not being fully exploited. Container throughput is the amount of work done by the TEU and the ability to handle containers at a minimal cost. This capacity of container throughput is the most important part of the scale of services, which is a crucial factor in selecting port terminals. At the port container terminal, it is necessary to allocate an appropriate number of available quay cranes to the berth before container ships arrive at the port container terminal. Predicting the size of a ship is especially important for calculating the number of quay cranes that should be allocated to ships that will eventually dock at the port terminal. Machine learning techniques are flexible tools for utilizing and unlocking the value of the data. In this paper, we used neighborhood component analysis as a tool for feature selection and state-of-the-art machine learning algorithms for multiclass classification. The paper proposes a novel two-stage approach for estimating and predicting vessel size based on container capacity. Our proposed approach revealed seven unique features of port data, which are the essential parameters for the identification of the vessel size. We obtained the highest average classification accuracy of 97.6% with the linear support vector machine classifier. This study paves a new direction for research in port logistics incorporating machine learning.