How Smart Analytics Work

This is the algorithm our servers follow:

1. We wait for the event from the camera, this can be any event that is configured on the camera. (Motion detection, line crossing, intrusion, etc.)  

2. Once the event is received it is checked against the Detection Rules that were configured in the admin portal on the Analytics tab of the camera properties.  

3. If the event matches the "Act on" requirement than our servers start frame by frame analysis of the footage.   

4. Our servers look for requested objects in all frames for the whole image for the whole duration of the event. The camera does not report to us in what area of the image motion was detected, so we analyze the whole image and report all moving objects. Stationary objects like parked cars are not reported.   

5. Once the requested object is detected the analysis stops and the server waits for the next event start time from the camera.   Multiple detection rules can be configured for the same camera.

Note:

For cameras that support motion detection area configuration over ONVIF the object detection will be looking for objects that are overlapping with the selected detection area and objects will be reported only if they are at least 25% within the area. This is done to decrease the number of unwanted object alerts. For example, if the camera is overlooking both yard and highway but only cars that enter the yard are of interest this logic will ignore all cars on the highway and will produce alerts only when the car enters the yard. If the analysis was done on the whole image every time the camera motion detector is triggered then any car that is present on the highway would be reported to create a true but unwanted alert.

For cameras that do not comply with ONVIF motion detection protocols (Axis and Vivotek are just some examples), the analysis will be done on the whole image and thus an object that is moving will be reported. Such cameras do not provide information about "active" areas thus we cannot filter out true but unwanted events based on the motion mask. 

• Related Articles

• Analytics Categories

  There are three default object categories included with the analytics plans: person, car and animal. If you need to recognize a more specific object please contact your local provider to request this. Some of the objects that can be requested ...

• Thermal Analytics Troubleshooting

  When setting up a Thermal Camera for guard service, a lot of the same standards still apply when it comes to best practices. There are however a few new things that you must take into consideration. 1. Ambient temperature and Background Since the ...

• Individual Cameras - Camera Analytics Setup

  Camera Analytics is an extremely important area of Camera Configuration. This is where the true value is built for a customer by giving him power over notifications, what they are notified about, and what real estate in the camera field of view they ...

• Can I See Camera-Side Analytics Through the Platform?

  Seeing the camera-side analytics from our platform will depend on the camera. If the camera-side analytics is ONVIF compliant than you will be able to control it through our platform without having to log into the camera GUI.

• What is the difference between Basic and Advanced Analytics?

  The main difference is that Statistical Analysis is only available for cameras with Advanced analytics while Basic Analytics will provide only information on the type of object detected, location, size of the object, and color attributes. Advanced ...