For cameras that support motion detection configuration over ONVIF, only those objects that intersect with the selected detection area will be reported. A minimum of 25% of the object has to be inside the detection area to trigger the alert. Objects that are outside of the detection perimeter will be ignored and no alerts will be sent for such objects.

In the above example, only objects inside of the highlighted area will trigger the alert. People walking on the sidewalk or cars on the road will be ignored.

Examples of scenarios that are known to result in poor object detection results:

1. Small and scattered detection areas. If multiple small areas are scattered the probability that the object will have at least 25% of its area intersecting the motion detection area is small. Only smaller objects will be recognized since big objects will have less than 25% of overlap with scattered detection areas. Such practice can help cut down on the false positives from the camera detector but this will cause missed object detection events due to not enough overlap.

2. Designating the whole area of the image as a motion detection region is not always a good idea as well. Even though this will ensure that all objects will be detected this, in turn, may result in unneeded alerts due to bypassing people and cars. For example, on the bellow image if the main purpose of the alerts is to warn about people at the front door sets the whole image area as a detection region will result in alerts for all moving objects. The better choice would be to set a detection area that covers just the stairs and front porch.