DCM offers two different versions of ‘People Counting’.
Which option is best for you, will depend partly on what type of counting data you would like, and partly on what camera views are available or can be installed.
Standard Headcount
The standard DCM Dashboard includes an ability to monitor the number of people in a single camera view or combine the headcount from multiple cameras into a real time total. Each camera view can monitor headcount in an area of up to approximately 500 square meters (depending on factors such as camera placement, lighting and obstructions).
If you have sufficient cameras to cover the entire space you are interested in, then standard DCM can provide a report of current attendance. Even if you do not have complete camera coverage of your space, DCM can estimate current attendance based on the percentage of area covered. Of course, the higher the percentage of camera coverage, the more accurate the estimate will be.
Headcount for individual cameras or combined totals can be displayed on the dashboard, and accessed through the reports engine.
This relies on cameras with a view of the area where people are expected to congregate. This may be the area in front of a stage, on a train station platform, or from multiple cameras with views of the main floor of an indoor event space for example.
The flexible DCM reporting engine can also be set up to total the headcounts from multiple cameras to give you a record of attendance over time as a ‘capacity report’.
Cumulative Counting
If however, you have a venue with a limited number of known entry points, and would like to know not only the current count of people within your space, but also their rate of entry and exit – then the cumulative counting option may be of interest.
The Cumulative Counting option measures ‘line crossings’ and thus requires a camera view at each gate or entry point you wish to monitor. If all possible entry points to your event space are covered, then this option can also provide a record of attendance over time.
💡 Summary of Counting Options
Both options can provide monitoring of the number of people within a space.
If you only wish to monitor how many people are within a given space and have camera views of at least the majority of that space or can install them, then use Standard DCM.
If you have defined entry points and are also interested in data about how busy each entry point is, as well as the rate of ingress or egress then use Cumulative Counting.
Cost and Data requirements for Cumulative Counting
Other factors that may influence the decision to use standard headcount or the Cumulative Counting feature include overall cost and data requirements.
The standard DCM headcount uses a less computationally intensive model, that can be run on a standard CPU. The ‘cross the line’ counting is more demanding and requires a GPU based model that is more expensive to run. Both can be implemented as a cloud based or On-premises solution, and the exact costing will depend on your particular circumstances and requirements, but in general you can expect the Cumulative Counting option to cost more.
In addition to being more resource intensive on the processing side, Cumulative Counting also requires a higher volume of data transfer. As a rough guide, the standard DCM headcount, density, flow and mood models combined may require data transfer rates of up to 3 mbps per camera. A camera running the Cumulative Counting model may need up to 10 mbps. Therefore in a venue or event where the network bandwidth available is low, the standard model may be preferable.
As well as being less resource intensive, the standard headcounting model is less vulnerable to network disruption or failure. The standard headcount feature totals detected patrons from all cameras in real time. If a temporary data outage occurs, once a connection is restored, the headcount will resume immediately and is still accurate. Because of the cumulative nature of ‘cross the line’ style counting, if there is a data outage or interruption during patron ingress for example, people entering that gate will not be counted during the outage. Even once restored, the total count from that gate will remain lower than reality in proportion to the length of the outage. Therefore, in an environment where either network bandwidth, speed or reliability is low, standard headcounting would be the recommended option.