Use Case for Smart City People Density

[mmccool]

Description: A service that can count people in a given location based on camera images and output the number of people, and/or an estimate of density. This is used to help ensure that social distancing/circuit breaker provisions are followed, among other applications. Requires an AI service whose input is image, identifies people (not personally, just as humans), counts them, and outputs a number. In combination with a camera providing images, this becomes a virtual sensor for density. As extension, the service may measure the distance between people and identify pairs that are too close to each other (note from McCool: want to avoid flagging non-strangers, though).
If density is too high, consequence would be an alert, sent for example to a security guard.

Variant: It might also be used to control access, eg. doors to a mall, to keep the density under some limit.

[mmccool]

This should perhaps be extended and generalized to "Public Health Monitoring" and include variants such as fever tracking, social distancing, mask detection, etc.
Suggest we update the existing use case (https://github.com/w3c/wot-usecases/blob/master/USE-CASES/smartcity-health-monitoring.md) with variants to support these kinds of activities, since they have similar requirements for image processing, alerting, statistical analysis, etc.
Relate to the edge computing use case also.

[mmccool]

As a further extension, there are also public health behaviours like hand washing (20s, all surfaces, etc), face touching (don't), etc. that could also be monitored. These require hand tracking as well.

[mmccool]

To be clear, my proposal is to extend the current smart-city-health-monitoring use case rather than creating a new one for person density... and also to include the other variants mentioned above.