aerostack

Search Aerostack

Norwegian

Rydoo Accounts Payable AI invoice automation at Norwegian

MidAmerica St. Louis

GoodMaps indoor navigation at MidAmerica St. Louis Airport

Lyon Saint-Exupéry Airport

Stan robotic valet parking at Lyon Saint-Exupéry

London Heathrow

Facial-recognition eGates at Heathrow

Almaty International Airport

Q-Gate biometric passport-control terminals at Almaty

Pittsburgh International Airport

Arrow Analytics EZ-Board computer-vision carry-on analytics test at PIT Gate A7

Federal Aviation Administration

FAA SMART airspace and trajectory management platform

Dubai International

Dubai Police AI systems for passenger screening, unattended baggage and camera analysis at Dubai airports

Airport intelligence / Chengdu Shuangliu

Chengdu Shuangliu: a mask-era algorithm and a wider smart-aviation trial

Chengdu Shuangliu introduced facial recognition capable of matching masked passengers and later gained regulator approval for an airport operations pilot. Forecast efficiency targets remain targets.

Terminal 2 exterior at Chengdu Shuangliu International Airport
Terminal 2 exterior at Chengdu Shuangliu International Airport. Photograph: AlexHe34 / CC BY-SA 3.0 (cropped to fit). Shows the airport, not an AI implementation.
The picture at Chengdu Shuangliu

One system entered use; another set goals

The pandemic forced airport identity systems to work with covered faces. At Chengdu Shuangliu, a recognition algorithm was introduced at boarding gates in early 2021 to compare passengers despite masks.

In 2024, China's aviation regulator approved a broader smart-aviation operational pilot at the airport, setting targets for taxiing, stand use and departure punctuality. These are proposed results over the following years, not results that can be attributed to the earlier boarding system.

The airport also has an older SenseTime identity-verification installation at security. Its role should remain distinct from the pilot's operating objectives.

Selected implementations

A specific algorithm and a regulator-defined test

01 / Deployed

Boarding passengers wearing masks

The operational problem

Airlines needed workable identity checks during periods when travellers wore face coverings.

What the sources report

China Science Daily reported in January 2021 that Chengdu Shuangliu introduced CloudWalk-based masked-face recognition at boarding, with a reported check taking roughly one second per passenger. In March 2024 the CAAC authorised a smart-aviation trial at Shuangliu.

Why it matters

The historical boarding function has a defined technical purpose; the later pilot addresses much broader operating coordination.

What remains unproven

The regulator's proposed improvements in taxi time, handling and punctuality are targets. The current record does not verify that these were achieved.

Sources and verification (2)

中国民航局 - 关于在成都双流机场开展智慧民航建设全流程运行场景试点工作的通知 · Evidence grade A

科学网(中国科学报)- 双流机场布局多种智能识别技术迎特殊春运 · Evidence grade A

Evidence supports a reported claim; it does not automatically establish a measured operating benefit.

Full implementation record →

The wider register

What else is happening at Chengdu Shuangliu?

SenseTime security identity verification predates the broader pilot. It is kept in the wider register instead of being counted as a second outcome of the 2024 initiative.

SenseTime SenseID face verification at Chengdu Shuangliu security, since 2018 · Deployed · limited scope

SenseTime has supplied SenseID to Chengdu Shuangliu since 2018, checking passengers' faces against their ID cards at security, the company said on 22 January 2019.

Sources and verification (1)

SenseTime · Evidence grade B

Evidence supports a reported claim; it does not automatically establish a measured operating benefit.

A target is a question for the next audit

Chengdu has two documented moments in airport technology development, separated by several years and different operational responsibilities.

The most valuable next evidence would report the actual outcomes of the regulator's operational pilot against its published targets.

Live data from the Aerostack register. Status reflects published evidence only. See Our approach for how records are graded.

Aerostack tracks AI deployments across airports and airlines, with the published evidence behind each record. Photography is real aviation imagery. Record thumbnails may show the named airport or airline; other photographs do not.