Regions of Interest (ROI)

Region of Interest

A camera may monitor areas beyond the ones of actual interest. VIS allows users to define specific zones within the camera view, called Regions of Interest (ROI), and assign actions to them. Multiple regions can be drawn on a single view. Each region can be linked to a detector and can trigger events based on detected objects. Regions are drawn as polygons and may be separated, intersecting, or overlapping. The region’s outline color automatically changes when a detection occurs within it.

Multiple region support

VIS allows for creating theoretically infinite-number of regions of any type. If there are overlapping regions, objects that are detected in the overlapping area trigger both of the regions.

Region groups

VIS can group regions to act as a pooled region to evaluate detected objects, alerts or violations jointly. This is configured in the region and io.yaml files or through the Configuration Manager.

This feature allows two or more unrelated cameras monitoring different regions to trigger a single alert—for example, when a person is standing beside a truck while a forklift is handling it. The forklift operator may not see the person behind the truck, posing a risk of dropping pallets on them. With region grouping, VIS detects a forklift in one region and a person in another, combines the information based on the grouping, and triggers an alert.

Grouping behavior as of V25 and on, assuming we have cameras monitoring one grouped region each:

Incident	Example	Behavior
Incident occurring in a single region	person encounters a forklift	State pool sets outputs State pool signals vislet via vislet state buffer Vislet logs, saves video, optionally sends email
Incident distributed to two cameras	person in one region and a forklift in another	State pool sets outputs State pool signals both vislets All notified Vislets log, save video, optionally send email

Region types

Presence region

A presence region is a designated area within the camera’s view where VIS detects whether relevant objects are present. The system continuously monitors this region and triggers events or alerts based on the presence or absence of objects, enabling safety or operational responses.

Counter Regions

Counter regions are used to quantify the number of detected objects within a specified area. VIS counts how many objects enter, stay, or leave these regions, providing valuable data for monitoring traffic, occupancy, or person flow.

If no regions are assigned to a Vislet, VIS does not perform any detection or counting processing. Assigning regions is essential for enabling targeted monitoring and event generation.