There are two different ways that Milestone can receive metadata from EdgeVis:
Embedded within the video stream
(This is known as Metadata 1 within Milestone)As its own separate metadata data channel
(This is known as Metadata 2 within Milestone)
While this can be confusing, they are used for different use cases, and each offers its own pros/cons. There are three aspects to analytics metadata that are useful:
EdgeVis per-frame metadata
There is per-frame information supplied by EdgeVis (e.g. video timestamp, selected camera input, full-res metadata). This is useful for live video plugins, such as the Milestone full-res plug-in.EdgeVis integrated analytics metadata
For our built-in analytics (Object Detector, Safezone 2D, ONVIF Profile S/T camera triggers), when an alarm triggers, there may be additional metadata (e.g. object type, bounding box) that is recordable and searchable within Milestone's recordings.ONVIF Profile M camera analytics metadata
EdgeVis supports the passing of ONVIF Profile M data directly from your IP camera to your VMS. This can contain a variety of data that your VMS can use (e.g. images of face detections or licence plates). This is used to allow Milestone to record and search the metadata within recordings.
To access any of the above you must use the correct Metadata stream:
| Metadata 1 | Metadata 2 |
EdgeVis per-frame metadata | Always available | Not-available |
EdgeVis integrated analytics metadata | Available (if not configured on Metadata 2) | Available if enabled* |
ONVIF Profile M camera analytics metadata | Not-available | Available if enabled |
(* see section Enabling Profile M metadata for built-in analytics)
The biggest difference between Metadata 1 and 2 is that 1 requires the video to be transmitted and decoded by Milestone, while 2 utilises a separate data channel and is independent of the video.
Metadata 1 is required when you need the live video to go alongside the metadata.
Metadata 2 is more useful when your Profile M metadata contains information (e.g. licence plates) or imagery (e.g. detected faces) that is independent of the video stream. You can save a significant amount of CPU/bandwidth if you do not need to transmit the video.
You can use either or both metadata channels - both are selected automatically when adding your encoder to Milestone (please see section below on completing Metadata 2's setup).
Note that recorded metadata will not be duplicated if both Metadata 1 and Metadata 2 are enabled.
How to ensure Metadata 2 is enabled when adding your encoders
In both cases, the metadata can be recorded at the Milestone Recording Server and when recorded, will be searchable within the Milestone Smart Client.
When adding your encoder to Milestone, it should automatically select both Metadata 1 and Metadata 2:
However, Metadata 2 is not automatically associated with the video stream, so the association must be added manually:
In the Milestone Management Client, select the camera to associate Metadata 2 with and then select the Client tab in the Properties pane to the right.
Click the ... button to the right of Related metadata, select Metadata 2 from the panel to the left and click Add to associate Metadata 2 with the video stream. Click OK, then click the save icon to the top left of the Milestone Management Client to save the configuration.
Enabling Profile M metadata for built-in analytics
By default, both Safezone 2D and Object Detector transmit their metadata within the EdgeVis video stream. You can also enable a secondary metadata channel that will transmit the analytics. This will allow Milestone to receive the metadata without having to stream the video.
To enable this secondary channel:
Open EdgeVis Client and view the desired encoder.
If necessary, change the current camera to the desired camera.
From the Settings menu, select Change analytics.
If you haven't already, select and configure your preferred analytic.
Before hitting APPLY, select the Configure button.
Ensure Publish ONVIF Profile M output stream is ticked.
Now save your changes using the APPLY button.
Notes:
This will not send any images for the detected object
It will have a small increase in bandwidth used - you can see the exact amounts on the Performance page on your encoder's local web interface.