Service Support
What is the practical foundation for advancing future research efforts?
The Group operates 65 productive terminals covering container handling, bulk cargo, liquid chemicals, and cruise passenger transport. It maintains a long-term reserve of professional engineering and technical personnel, with all terminals and ports within the Group fully supporting the innovation team in the R&D, practice, iteration, and application of intelligent port products. Affiliated units of the Group can provide substantial industry-specific data, which serves as the most critical scientific research infrastructure for artificial intelligence algorithm research and engineering practice.
Currently, the team comprises over 30 members, among whom more than 80% are engaged in product design and R&D. The core team consists primarily of professionals with master's degrees or above from top-tier universities (985 Project), many of whom have work experience in renowned enterprises. Additionally, the team includes multiple project managers and product designers with over five years of port industry experience, enabling them to deliver professional and comprehensive industry solutions for port and shipping enterprises. The team keeps abreast of the latest industry technologies and has achieved more than 20 cutting-edge academic results in the field of artificial intelligence, including publications in top journals and conferences such as CVPR and ECCV.
What is the Future Development Plan?
The team's research covers port safety, intelligent supporting of new equipment, intelligent transformation of existing equipment, intelligent port processes, and intelligent port operation and scheduling. Among these key research directions, the integrated intelligent visual solution for ports is more advanced than current domestic and international counterparts' solutions that primarily rely on lidar and millimeter-wave radar. The product technical means adopted in the research and application are at the forefront of the industry, maintaining a leading position among peers in the same technical field and application scenario.
Through intelligent technologies, the team aims to promote the industrial upgrading and development of the port industry. Starting from Xiamen and radiating nationwide, it will improve the quality and efficiency of terminals and enhance the core competitiveness of ports across the country. The team will move towards standardized and market-oriented development, participating in international exchanges and global intelligent port projects every year. The Group will provide comprehensive support in terms of data, application scenarios, research and practice venues, personnel and equipment, as well as funding and policy support.
What are the Advantages of the Deviation Correction and Collision Avoidance System?
RTG (Rubber-Tired Gantry Crane) is a type of gantry crane that travels on a rubber-tired chassis. Boasting high efficiency and strong adaptability, it serves as a crucial loading and unloading equipment in container terminals. However, traditionally, RTGs were not only manually operated but also ranked among the equipment with frequent safety accidents. During loading and unloading operations, RTGs must travel strictly along predetermined tracks with an extremely narrow allowable deviation range— the distance between two cranes must not exceed 10 centimeters. Nevertheless, the complex operating environment of container terminals and uneven pavement make it difficult even for experienced drivers to consistently maintain this tolerance, despite constant focus and careful maneuvering. As a result, RTGs often deviate from their intended tracks. Currently, deviations are primarily corrected manually by drivers who rely on visual observation. However, factors such as driver fatigue, concentration levels, and operational habits can easily lead to human errors, potentially causing collisions between RTGs, container impacts, or even yard collapses.
GCT has introduced an industry-leading AI pure-vision RTG deviation correction and collision avoidance system for on-site RTGs, replacing manual operations to provide reliable, continuous anti-collision protection and automatic deviation correction during operations. In terms of deviation correction, the system eliminates the reliance on sliding contact lines, high costs of magnetic nail installation, and the construction of GPS differential stations associated with traditional solutions, achieving low-cost and reliable deviation correction. For active collision avoidance, unlike conventional millimeter-wave radar or laser-based systems, the AI visual recognition technology enables human-like perception to accurately estimate the type, distance, and movement trajectory of obstacles within a 15-meter range. This effectively prevents false triggers; based on the RTG's traveling speed, the system implements intelligent deceleration and emergency braking controls tailored to the obstacle type and distance.
Technically, the AI vision system captures raw data exceeding 270 million pixels per second in real time through 6 custom industrial cameras mounted on the RTG. Equipped with customized AI computing power, it achieves up to 238 trillion integer operations per second. A deeply customized AI neural network model has been designed and trained specifically for GCT's yard layout, equipment characteristics, and special targets. Combined with high-performance AI computing capabilities, the system delivers efficient recognition, reasoning, and control—achieving response performance comparable to or even superior to human drivers. Integrating cutting-edge control algorithms running on the AI computing platform, it realizes smooth deviation correction with minimal jitter, maintaining a deviation error within 50mm. This system represents a pioneering application of artificial intelligence and machine vision in port RTG operations.
6th Floor, No. 44 Guanri Road, Software Park Phase II, Siming District, Xiamen City, Fujian Province, China
© Copyright 2025 Fujian E-Port Co., Ltd. All rights reserved.
