Surveillance and CCTV

Surveillance Dashboard

The Surveillance Dashboard (/surveillance) submits sensor feeds for analysis, cues one sensor from another, and streams telemetry from the common operating picture (COP). To open it, go to Today’s Command and select Video in the AI Command bar.

Ingest a feed

### Describe the sensor Under **Feed Ingestion**, enter the **Sensor ID** and choose the **Domain**: LAND, AIR, MARITIME or CYBER. Then enter the **Feed URL** and the **GPS Coordinates**. ### Set the threshold Move the **Confidence** slider (0 to 1). ### Submit Select **Ingest Feed**.

The result tile reads Threat Detected — {classification} or No Threat Detected. It lists the damage grade, severity, remediation, any text read from the frames, and how many frames were processed. If the server attached a note saying what it could not measure, the note is shown word for word. That note is different from a clean result.

Cue a sensor

Under Sensor Cueing, enter the Source Sensor ID and choose a Target Sensor Type: INFRARED_CAMERA, YARD_CCTV or PLC_CRANE_LOCK. Enter the Target Coordinates and select Trigger Cueing. The result reads Cue Acknowledged or Cue Rejected, with the target’s status.

Stream COP telemetry

Under COP Telemetry Feed, enter a Subscriber ID and select Start COP Stream. A LIVE badge shows while events arrive, newest first. Each event carries its severity as a glyph and a word, not by colour alone. A SEVERE or HIGH event also raises a Surveillance alert. Select Stop COP Stream to end the stream. Before the first event, the list says No telemetry events yet.

Captured media, on the same page, lists what this browser tab captured. It is held in the tab only.

If a request fails, the page shows the error in place of the cards. For example: Failed to ingest surveillance feed: {error}, Failed to trigger sensor cueing: {error} or Telemetry stream error: {error}.

Behind this page: SurveillanceService.IngestSurveillanceFeed, TriggerSensorCueing, and SubscribeTelemetry (server stream). See the API reference.

CCTV connection

CCTV Connection (/cctv_connection) checks an RTSP or ONVIF camera and samples knowledge from its footage, such as occupancy, safety or inventory. To open it, go to Fetch and select CCTV Connection beside ACTIVE CONNECTIONS. The info button (Info) explains the page.

### Enter the credentials Under **Configure Credentials**, fill in **ACCESS KEY ID**, **SECRET ACCESS KEY** (hidden until you reveal it) and **BASE ENDPOINT URL**. ### Test Select **Test**. The cockpit checks that the endpoint answers, then asks the camera to describe itself. ### Read what the camera reported Under **WHAT THE CAMERA REPORTED**, the state chip reads one of four values (see the table below). When the camera describes itself, each stream shows its codec, width, height, frame rate and stream URL. A value it did not give reads **the camera did not state this**. **Use this stream** copies a stream's URL into **BASE ENDPOINT URL**. ### Extract knowledge Select **Extract Knowledge**. **EXTRACTED KNOWLEDGE** shows a text summary and the sampled frames. **COPY SUMMARY** copies the summary to the clipboard.

State	Meaning
Not asked	Nothing has asked the camera yet. Run Test.
No answer	The endpoint did not answer, so the camera was never asked.
Described nothing	The endpoint answered, but nothing describing the camera came back.
Camera described itself	The streams listed are the camera’s own report. That doesn’t prove anything is publishing on them.

This page tests and samples a camera. To add a video feed as a data source, use the Video Stream card in the connect-a-data-source wizard.

Behind this page: CCTVService.TestConnection, DescribeSource and ExtractKnowledge. See the API reference.

Simulate Calling