There are many environmental and use case requirements that will impact the sizing of a VMS plus CV application solution. Many of these factors cannot be simulated in a lab environment, however, with the right source videos and camera simulators a relatively large implementation can be modeled closely enough to gather useful sizing data.
Our design evaluation was targeted at collecting data for three sizes of transportation systems based on the total number of cameras and the number of cameras that require processing with a CV application, primarily for alert generation. The table below shows our three category definitions.
|Size||# of cameras||# analytic cameras|
|Small||250 to 500||Up to 125|
|Medium||501 - 1000||Up to 250|
Our sizing categories above reflect the reality that most customer systems will need a mixture of cameras with some near real-time analytics as well as those that are intended for archive only in the event there is a need for historical video for investigations. In this testing, we are simulating a ratio of 1 analytic camera for every 5 total cameras installed. This will vary across customers and workload needs but 5:1 is the baseline used for all our sizing data collection.
The design can be applicable for environments with greater than 250 cameras (small) but should be useful for system designers working with airports with greater than 5000 cameras by scaling the guidelines for a large (>1000 camera) environment.