“HIRI FUTURE” Selected as a Typical AI Case in Fujian Province for 2025

Recently, the 2025 Fujian Province Typical Artificial Intelligence Case Exchange and Project Matching Conference was held in Fuzhou. Since its launch in September, the event has attracted extensive attention from all sectors of society, with a total of 62 companies submitting 82 cases covering various fields of social and economic development, including transportation, logistics, and urban governance. These cases include both practical examples of “cost reduction and efficiency improvement” in the intelligent transformation of traditional industries and technological breakthroughs “from zero to one” in emerging fields, showcasing the application achievements of artificial intelligence technology in Fujian Province.

Among numerous outstanding cases, the “Intelligent Agent for Automation in the Vertical Field of Port Equipment” project launched by Fujian Electronic Port Co., Ltd. (hereinafter referred to as “HIRI FUTURE”) stood out and was recognized as a typical AI case in Fujian Province in 2025, demonstrating cutting-edge technological application directions in the field of smart ports.

 

 

 

Given the fragmentation of massive sensor data and business systems in the port industry, as well as the lack of a unified perception and decision-making hub, this project aims to build a “port-wide automated Intelligent Agent” to address the industry’s common pain points—low automation levels, poor system synergy, and high security risks—and thereby drive a revolutionary upgrade of port operations from “point-based intelligence” to “systemic intelligence.”

HIRI FUTURE adopts an integrated technical architecture of “edge perception, cloud decision-making, and data-driven” to build an Intelligent Agent Platform for the port vertical domain. At its core, it achieves intelligent perception and real-time control at the device end, global collaboration and scheduling in the cloud, and continuous capacity evolution through data and model platforms, comprehensively covering the core links of automated port operations.

 

Technical Architecture Diagram of Port Equipment Automation Intelligent Agent

 

Four-tier architecture

01. Intelligent Perception and Edge Control Tier

Intelligent nodes equipped with cameras, LiDARs, and computing units are deployed on key equipment such as Quay Cranes, Yard Cranes, and Container Trucks, integrating multiple sensors to achieve Intelligent Container Recognition, centimeter-level positioning, and local automation operation control.

 

02. Cloud-based Intelligent Decision-making and Scheduling Center

As the “intelligent brain” of the system, it coordinates equipment resources across the entire Procedure through multi-agent collaborative algorithms, achieving seamless collaboration and dynamic optimization among ships, bridges, vehicles, and yards.

 

03. Data- and Model-Driven Platform

Relying on the MLOps system, we have built 63 high-quality datasets and automated model training pipelines, supporting continuous algorithm iteration and edge-cloud collaborative deployment, thereby driving the continuous evolution of system Intelligent.

04. Integrated Industry-University-Research-Application Mechanism

In collaboration with universities, we have undertaken key technological research and development, fostering a closed-loop system that integrates R&D, technology transfer, and real-world application. To date, this effort has resulted in more than 70 patents and software copyrights, with deep compatibility across domestic chips and mainstream port systems.

Five Systemic Innovations

01. Technological Architecture Innovation

It has pioneered a four-tier data closed-loop system of “perception-decision-optimization-iteration”, breaking the limitations of traditional single-point automation and providing end-to-end support from data collection to operation optimization;

 

02. Innovation in Annotation Tools

It was the first to apply weak-supervised learning technology to port scenarios, developing an automation annotation toolchain that increased annotation efficiency by 500% and reduced costs by 80%;

 

03. Algorithm Model Innovation

By independently developing a large-scale model for the vertical field of ports, we have achieved comprehensive self-reliance and controllability from data modeling and network structure design to inference optimization, successfully filling the technological gap in the international community regarding the lack of dedicated large models for fully automated port operation scenarios.

 

04. Business Procedure Innovation

In response to the need for end-to-end automation in ports and following the closed-loop approach of “perception-decision-optimization-iteration,” we have independently developed 63 high-quality sub-datasets covering key scenarios. By integrating algorithmic models with platform-based management, we provide end-to-end support for unmanned port operations, from data collection to practical implementation.

 

05. Compatibility Advantages

Supports mainstream equipment vendors such as Sugon and Siemens, enabling seamless integration with port management systems like TOS and ECS.

 

 

Significant results have been achieved, setting a new benchmark for Intelligent in the industry.

Improved operational efficiency: At a container terminal in Xiamen Port, after implementing this system, the average port stay of ships was reduced by 18%, and the operational efficiency of each Quay Crane increased by 25%, equivalent to handling 10 additional TEUs per hour.

Reduced operational costs: By decreasing staffing levels for roles such as drivers and dispatchers and optimizing equipment energy consumption, overall labor costs in relevant processes were reduced by 35%, while equipment energy consumption decreased by 8%.

A qualitative leap in safety: the system now provides 24/7 uninterrupted automated collision-avoidance warning and intervention, and since its launch, the rate of human-error-related accidents in the relevant operational areas has dropped to zero.

Industry-leading recognition accuracy: In complex environments, the container number recognition accuracy remains stable at over 99.9%, and the license plate recognition rate exceeds 99.5%, providing a reliable guarantee for automation Procedure.

Currently, HIRI FUTURE’s standardized Intelligent agent solution can be flexibly adapted to different port types such as container ports and bulk cargo ports, boasting high scalability and the ability to be replicated on a large scale. Its technological products have been widely deployed in more than 80 major ports both domestically and internationally, including China’s key hub ports such as Tianjin Port, Shanghai Port, Ningbo Port, Xiamen Port, and Hong Kong MTL Port, as well as core ports along the Belt and Road Initiative and in BRICS countries (such as Malaysia, Singapore, Thailand, and Brazil). To date, over 3,000 sets of equipment and system solutions have been successfully implemented and deployed.

 

The picture shows the automation operation Stack yard at Haitong Quay.

 

The successful selection of HIRI FUTURE’s “Intelligent Agent in the Vertical Field of Port Equipment Automation” not only demonstrates the company’s technological strength but also serves as a vivid example of Fujian Province’s efforts to deeply integrate artificial intelligence with the real economy. The project drives the intelligent upgrade of ports through systematic innovation and empowers continuous operational optimization via a data closed-loop, setting a new technological benchmark and development paradigm for smart port construction.