By Zhao Yu

An AI-powered quadruped robot—dubbed a "robodog"—has been put into service at the Meishan Port Area of Ningbo Zhoushan Port, marking a new step in Ningbo Customs' push toward smart, technology-driven inspection.

At the empty-container inspection site, the robot moves autonomously among stacked containers, stopping to scan exteriors or entering units to carry out patrols, recognize container numbers, and detect foreign objects.

The robot is now assisting with quarantine inspections of inbound empty containers, introducing what customs officials describe as an "AI-assisted smart inspection" model.

Meishan Port Area, one of the fastest-growing and most automated container terminals at Ningbo Zhoushan Port, handled more than 13 million TEUs in 2025. Inbound empty containers totaled 3.858 million TEUs, raising demands for faster turnover and more efficient supervision.

Traditional manual inspections have struggled to keep pace. Daily inspection volumes average around 200 containers, while labor shortages and exposure to extreme summer and winter conditions add operational strain.

To address these challenges, Ningbo Customs deployed the AI-powered quadruped robot. Equipped with optical and thermal imaging systems, multiple industrial-grade cameras, laser radar, and high-precision navigation modules, the robot can independently plan patrol routes, identify container numbers, capture images, and transmit data in real time.

"Using laser measurement and image-computing technology, the robot can determine whether containers contain concealed compartments," said Zhang Jian, deputy director of the No. 2 Logistics Monitoring Division of Meishan Customs. "High-definition cameras and fill lights also improve the detection of disease vectors and harmful organisms."

Designed for complex environments, the robot can work around the clock with consistent inspection standards, significantly improving efficiency and reliability.

According to Zhang Xiaorong of Unitree Technology, the robot has been customized for customs use and features advanced motion control, allowing it to navigate narrow spaces, climb uneven surfaces, and operate in all weather conditions.

As of Jan 29, the robodog had assisted in inspecting 1,655 containers, achieving a 92% accuracy rate in container number recognition and a 95% success rate in foreign-object detection. During testing, it inspected up to 100 containers in a single day, with overall efficiency expected to more than double compared with manual inspections.