The ML Classification Rule wizard will create a rule which includes an ML Classification verification. It concludes with a video demonstrating this verification.

Contents

Configuring the Wizard

1. Choose the ML Classification Rule type and click Next:
   
   

   Open

   

    

2. Select a New Area (in green) or an existing Area (in white).
   
   

   Open

   

    

3. Here a New Area has been selected. This will automatically create an area entitled “ML Classification Rule Area” under the folder Areas in the configuration tree. The area can be edited and moved once the wizard has been completed. Click Next:
   
   

    

4. On the second page of the wizard, you can configure additional properties. None are selected by default, and if not configured at this point they can be added at a later date by editing the rule. Click Next:
   
   

    

5. On the third page of the wizard you can configure Set Break Allowance Properties. In the example below the Priority was changed to High. Finally, click Finish:
   
   

    

6. This is the resulting ML Classification Rule:
   
   

    

7. You can then edit the rule if required. In the example below there is an additional ML Classification:
   
   

Nodes

Input

Track Sighting Input https://navtechradar.atlassian.net/wiki/spaces/TUN/pages/1874853893/Rule+Inputs#Track-Sighting-Input

Check

Area Check https://navtechradar.atlassian.net/wiki/spaces/TUN/pages/1866498140/Rule+Checks#Area-Check

A check if the target enters the specified area.

Filtered Property Check https://navtechradar.atlassian.net/wiki/spaces/TUN/pages/1866498140/Rule+Checks#Filtered-Property-Check

This checks if a property of a Track Sighting meets the defined criteria after previous rule checks have passed.

ML Classification Check (1)

Verify the classification using ML object detection.

ML Classification Check (2)

Actions

Camera

Selecting the Keep Following checkbox will set a camera to keep following the target after the condition clears.

Alarm (1)

This will sound an alarm if the Area is entered. The Priority and Severity of the ingress can also be configured; for example, in a highly sensitive area then you may want the highest threat level to be set. By default, the setting is a Medium Threat.

Alarm (2)

Video

Here is a short video demonstrating an ML Classification Rule being broken; a vehicle in the area will cause an orange warning alarm, and a person in the area will cause a red threat alarm:

Related Information