Home' RTCA Documents for Review : DO-230H FRAC Contents 196
© 2017, RTCA, Inc.
images (quarter frame) and at reduced frame rates, typically 7 to 12 fps, are more susceptible to false alarms,
but this solution is less computationally intensive and also less expensive. When video analytics programs
run CIF frames, the frame rate should be at least 7 fps and preferably 10 or more fps - the required minimum
frame rate for video analytics to function reliably should be verified with the camera and VMS supplier(s0
and then be demonstrated in the airport’s actual environment.
The best place to perform live video analytics is at the camera, at the edge of the network, where the
analytics can be applied to the video signal before it is compressed or has its frame rate reduced to conserve
transmission bandwidth. This also provides a reliability advantage, since a camera failure only impacts that
camera whereas if a central server fails multiple cameras become inoperative. Camera-based analytics,
however, may not be optimal functionally, may not even be made by the camera manufacturer, and commit
the user to one source of analytics whereas server-based analytics offer more vendor choices.
In both cases, it is important to check that camera firmware updates do not adversely impact video analytics
The best place to locate post-event processing is on a central server, so that recorded video can be searched
many times and using different parameters.
Testing the performance of video analytics is complicated by scene variations, object variations, and
scenario variations among other factors. One approach which an airport should consider is to digitally
scenes at selected sites under the range of operating conditions which will actually be encountered. These
recordings can serve as performance baselines for vendors to assess whether their analytics will perform
adequately. Candidate cameras can then be tested at the selected sites using pass-fail criteria approved by
the airport, and below which the analytics are not worth the airport’s investment.
Artificial Intelligence (AI)
Artificial intelligence (AI) is a collection of analytical means of assessing and applying data using powerful
software algorithms. Depending on the application, it may involve techniques such as “machine learning",
or “deep learning", or "neural networks", or "natural language processing", or “cognitive technology” - all
of which vary by manufacturer, lack transparency in now the functions are performed, and are usually
In a generic sense, video analytics is applied AI because it uses analytics to extract "intelligence" from
video streams from which decisions can be made. VA technology continues to improve. This is evident in
biometric analytics, and especially in facial recognition which is now ready for practical applications in
many instances and, which show how far VA technology has advanced in recent years.
For airport physical security systems, with the exception of some (but not all) VA applications, AI is not
likely to have a large impact in the near future. The future, however, is moving rapidly, fueled by
commercial developments vastly larger than the security market including attempts to realize self-driving
automobiles and airborne drones, also known as Unattended Airborne Vehicles (UAVs), by improving
medical diagnostics, and in other areas.
System and Subsystem Integration
There are no standards for integrating video surveillance systems with other elements of a physical access
control system, or the extent to which such integration should be implemented. The criteria for making
Links Archive Navigation Previous Page Next Page