Camera teams, image zone partitioning, intelligent targeting.

[Lorenzo]

This would be a real game changer for the smart, combined use of camera sets. Cameras should allow splitting the image into logical zones and, for PTZ models, mapping presets to these zones.
When an alarm is triggered, it can be associated with a specific zone (the AI already calculates internally where the event occurs). This way, the system not only knows that something happened, but also where it happened, even without immediately viewing the video.
In a team setup (e.g. one fixed wide-angle camera plus a PTZ), the PTZ could automatically move to the corresponding preset linked to the zone to verify the event in detail.
Combining this approach with a global alarm scoring/escalation system would dramatically reduce false positives and unlock the potential for selling camera sets that complement each other — fixed cameras for detection, PTZs for confirmation and detail — delivering far more advanced and reliable monitoring.
I believe that not much is missing at the firmware level to enable zone division, since the AI already relies on bounding-box coordinates. The additional functions (mapping zones to PTZ presets, alarm scoring, automation) could easily be implemented in the Reolink app, NVR software, or integrated via Home Assistant.