AI Video Analytics: From Video to Data to Prediction to Airport Value

Guest Author by Guest Author | Jun 2, 2022

By Maurice Jenkins, Digital Strategy and Innovation Advisor, CIO, Miami Dade Aviation Department

Computer Vision (CV) is an application of Artificial Intelligence (AI) that can process and recognize the content of video or images. A CV model can take an image or video stream input and generate a list of objects or features it recognizes. For example, feed in a picture of an airport gate area and a CV system can output the number of people in the area, the number of bags, and the number of seats occupied. Feed in a video stream of the gate area and you get analytics on how early people turn up for boarding, when boarding starts, and how long the boarding process takes (and if anyone leaves an article behind).

While CV is not new (with research papers dating back to the 1960s), it has however undergone startling progress in the last several years, mostly driven by access to vast datasets, coupled with general advances in AI. Much of the work by companies like Google, Amazon, Microsoft, and Facebook is either open-sourced or made accessible by CV as a service platform. These advances are simplifying access to CV for enterprises that can gain business benefits without needing deep understanding of the technologies.

Why is CV important for airports?

As we all know, digitization of airports and access to data is critical to intelligent airport operations. But data acquisition remains hard, especially for measuring what are inherently offline or human activities – for example how busy a curbside is, or where the wheelchairs are.

This is where CCTV plus CV comes in. All airports have comprehensive closed-circuit television (CCTV) coverage across their campus. When these are integrated with CV systems, they can intelligently observe events at the airport and output telemetry or analytics about what is happening across the campus. In effect, each CCTV camera can become an intelligent edge sensor.

CV is applicable if a human could do the same thing visually, such as looking at a security checkpoint and counting the amount of people there. On the other hand, CV is not applicable if you need technology to complete the task. For instance, it cannot look at a conveyor belt system and tell you when it is overheating (for this, you will still need sensors).

Some specific examples of the use of CV at airports include monitoring curbside traffic; finding spaces in garages; measuring queue wait time; measuring passenger density; monitoring aircraft turn process on the apron; measuring foot traffic at retail areas; and counting the number of people using restrooms to name a few.

Challenges to overcome on the way

Using CCTV plus Computer Vision to measure utilization and performance across the airport can generate a significant amount of data, in turn increasing operational efficiencies, but there are pitfalls to be aware of and hurdles to overcome.

A principal concern is that of privacy, of both passengers and staff. While it is accepted by all that CCTV is necessary and a good thing for airport security, repurposing CCTV for computer vision processing can create some pushback amongst airport stakeholders unless handled sensitively. It is important to be open about intention from the beginning, and to involve airport stakeholder representatives from the on-set.

Integration with the CCTV infrastructure is also not straightforward. There are security/policy issues, along with old fashioned technical integration issues. Depending on your airport, the CCTV infrastructure could be controlled by any number of departments, from security to law enforcement to facilities. They may not be as open to the repurposing of CCTV data for the benefit of improved airport operations. It is important to be able to communicate your goals, and to have both a light touch integration along with clear data processing and retention policies.

Once all this is done, it is important to remember that using CV for data acquisition is just the starting point. Leveraging this data with airport stakeholders to track and optimize performance and reengineer business processes is where the real value is derived.

Value outweighs the challenges

The ACI World IT Standing Committee (WAITSC) that brings together airport CIOs from around the world believes that the potential value that CV can bring to airports, airlines and passengers outweighs the challenges each airport must overcome. The WAITSC is currently combining forces to overcome these challenges and drive our industry forward.

Maurice Jenkins functions as the principal advisor to the Aviation Director on all matters related to information management and information technology. He has full strategic and tactical responsibility for the direction, planning, creation and operation of the Aviation Department’s information technology infrastructure, telecommunications services, application support, end-user support services and training.

Guest Author
156 articles
Share This