Yongpai: The first in the country! Ningbo Zhoushan Port takes the lead in implementing "embodied intelligence" application for heavy container inspection

On March 2nd, at the heavy container inspection site in Meishan Port Area, Zhoushan Port, Ningbo, a quadruped "robot dog" from Yushu Technology flexibly navigated between containers, aiming at the container number and lead seal of the container to be inspected for precise verification - this is the first time that the inspection assistance robot has undertaken automatic shooting and recognition tasks of container number and lead seal number in ports across the country. This move has taken the lead in realizing the embodied intelligent application of the customs heavy container inspection platform in ports across the country, creating a new paradigm of collaborative intelligence between customs and ports, and accelerating the transition of heavy container inspection business towards intelligence and efficiency.
Embodied intelligence refers to intelligent systems that perceive, understand, and act based on physical entities. Through real-time interaction between physical agents and the environment, it can autonomously obtain information, analyze problems, and perform tasks. The Information and Communication Company, in collaboration with Meidong Company, keenly seized the development opportunities of embodied intelligence, went deep into the front line of the port area, repeatedly debugged and optimized, and successfully built a crane box inspection auxiliary robot system based on the "sail riding" AI model. This system accurately solves the "time difference" problem in the collaborative operation between customs and ports, and promotes a leapfrog upgrade of the inspection mode from the traditional "manual arrival and individual verification" to "intelligent perception and efficient approval".
In recent years, with the continuous increase in container throughput in Meishan Port Area, the average annual growth rate of inspection workload has exceeded 10%. However, the capacity of inspection sites in the port area is close to saturation, and the problem of tight human resources has emerged, which restricts the improvement of port operation efficiency.
In the current heavy container inspection process, customs staff need to take photos of the container number and lead seal number one by one, confirm that they are correct, open the container door and take a photo of the entire cargo before starting the container dumping operation. Nowadays, customs officials can issue operational instructions through relevant systems, and inspection assistance robots can autonomously complete the path planning and safe obstacle avoidance of target containers. With high-definition cameras and AI recognition algorithms, they can accurately collect container number and lead seal number information. The image records of the entire homework process are synchronously sent to the customs, forming a traceable and verifiable complete supervision chain. After the customs officer verifies the quality of the photos and confirms them, the dock operator can cooperate to complete the unpacking and overturning operation.
The robot has already verified all the container information in advance, reducing the opening time to 20 minutes and greatly improving the efficiency of opening and reversing boxes! "Said Zhang Wenzhi, a customs officer from Meishan Customs Inspection Department. Data shows that the current accuracy rate of box number recognition for inspection auxiliary robots exceeds 99%. With three inspection auxiliary robots working simultaneously, the work that used to take 4 to 6 people more than 1 hour to complete can be compressed to be completed within 20 minutes.
The successful application of the "robotic dog" this time is an innovative practice of deepening customs port cooperation and promoting the construction of smart ports by Zhejiang Port Group, Ningbo Zhoushan Port Group, and Ningbo Customs. Next, both sides will continue to deepen the application of intelligent inspection technology, continuously optimize the accuracy of AI recognition algorithms, improve the human-machine collaborative scheduling mechanism, further expand the application boundaries of embodied intelligence in more scenarios of ports, and strive to create a new industry model of intelligent customs clearance collaboration, injecting stronger intelligent momentum into the construction of world-class ports.