AI helps airport identify passengers using facial recognition technology

Summary:

By leveraging AI facial recognition technology, the airport significantly enhanced both operational efficiency and passenger satisfaction.

 Client: 

Beijing Daxing International Airport (PKX)

Problem Statement: 

In 2018, Beijing Daxing International Airport (PKX) handled over 100 million passengers annually, posing a challenge in operational efficiency, safety, and service quality. Manual operations led to issues such as mismatched luggage, strict safety management, and a poor travel experience. To address these challenges, Beijing Capital International Airport Company aimed to develop a smart security check-in system for Daxing Airport to enhance operational efficiency and passenger satisfaction.

Results: 

• Achieved over 99.9% accuracy in passenger identification and verification. 
• Increased inspection rate for individual passengers from 180 to approximately 260 per hour, while baggage inspection has reached 792 pieces per hour. 
• Enhanced the overall efficiency and safety of security inspections.
• Greatly reduced passengers’ waiting times.

AI Solution Overview:

The SenseTime Smart Passenger Security Check System automatically verifies passengers’ identity documents and flight tickets before departure. It also operates in the luggage security check area, matching passengers to their luggage through real-time facial and video analytics, which minimizes luggage mix-ups. This system significantly reduces the time needed for security checks, providing travelers with a smooth, stress-free experience.

The ML model behind the software was trained on thousands of images and videos from common open-source libraries, including cloud services, social networks, mobile uploads, and embedded hardware platforms. These sources provided images of people’s faces from various angles and under different lighting conditions, labeled as actual faces, 3D images, drawings, printouts, or disguised faces using wigs and masks. These labeled images and videos were processed through the software’s machine learning algorithm, training it to recognize the sequences and patterns that represent faces, drawings, printouts, or disguised faces.

Users can then upload unlabeled images or videos into SenseTime. The software’s algorithm can identify and detect potentially anomalous images or footage. The system alerts a human employee in real-time about these anomalies for further inspection.

References: 

1. Beijing Daxing International Airport (PKX). https://www.sensetime.com/en/case-detail?categoryId=32014&gioIndex=1 
2. SenseTime AI Serves International Travelers at the Newly-opened Beijing Daxing International Airport. https://www.sensetime.com/en/news-detail/3898?categoryId=1072
3. SenseTime’s AI Technology Enables Intelligent Security Check-in System. https://pandaily.com/sensetimes-ai-technology-enables-intelligent-security-check-in-system/
4. Facial Recognition at Airports – Current Applications. https://emerj.com/ai-sector-overviews/facial-recognition-at-airports-current-applications/

Industry: Public Services

Vendor: SenseTime

Client: Beijing Daxing International Airport (PKX)

Publication Date: 2019