Websocket
Connect to Docker Stream Websocket.
Getting Started
Websockets are a technology that enables real-time communication between a client (such as a browser or application) and a server. They are used to consume real-time event data, particularly useful in applications that rely on immediate data updates like those involving AI model inferences in systems such as VisionAIre Stream.
When an AI model completes an inference task, such as object detection or facial recognition, the results need to be communicated instantly to users or other systems. This is where WebSockets come into play. Unlike traditional HTTP connections that require a new connection for each request/response cycle, a WebSocket establishes a persistent, full-duplex connection over a single TCP connection. This means once the connection is open, data can be sent back and forth between the client and server in real time without the overhead of establishing new connections.
In the context of a dashboard that monitors real-time events, this technology allows for the immediate push of data from server to client as soon as an AI model inference is completed. This ensures that the dashboard can update instantaneously, providing users with real-time alerts and monitoring capabilities. The continuous and instant nature of data transmission via Websockets is crucial for applications requiring immediate data processing and display, enhancing responsiveness and user experience.
To help you understand WebSocket in detail, please check the article reference below:
Websocket Construct
To construct additional parameters for web socket query:
General Analytics
For Multi Logics Analytics (People or Vehicle Analytics)
Example Response
Face Recognition
Face Recognition with Person Attributes
Key
Type
Description
analytic_id
string
Type of analytic. NFV4-FR is Face Recognition
image_jpeg
string
Image of captured object in base64 string.
node_num
int
The node number computation. If you only have a single instance, thennode = 0.
pipeline_data
object
The output of each analytics.
confidence_detection
float
The confidence score of face detection.
face-id
int
Unique identifier of the captured face.
similarity
float
Value of face embedding similarity between detected face and enrolled face.
status
string
The Status information of enrolled photo
variation
string
The unique id of the captured face image.
primary_text
string
KNOWN or UNKNOWN label of the detected face. Known when the face is recognized and UNKNOWN when the face is unrecognized.
secondary_text
string
This key was left empty.
stream_address
string
The original address of stream/camera.
stream_id
string
The unique id of the assigned stream/camera.
timestamp
string
Unix timestamp of the detected face.
event_id
string
Timestamp with 8 char hash
License Plate Recognition 1
License Plate Recognition 2
License Plate Recognition with Multiple Classes
Key
Type
Description
analytic_id
string
Type of analytic.
jpeg
string
Image of captured object in base64 string.
pipeline_data
object
The output of each analytics.
label
string
Class of detected objects, currently support car, motorcycle, bus, and truck
area_name
string
Area name of captured plate and vehicle.
direction
string
Direction of vehicle in ROI: Down, Up, Right and Left
event_id
string
Timestamp with 8 char hash
confidence
float
Confidence the vehicle classification
plate_type
string
Class of vehicle plate number such as APH or public
plate_number
string
The plate character prediction.
stream_address
string
The original address of stream/camera.
stream_id
string
The unique id of the assigned stream/camera.
timestamp
string
Unix timestamp of the detected vehicle and plate.
Vehicle Counting
Vehicle Analytics
Vehicle Analytics with Attributes
Key
Type
Description
analytic_id
string
Type of analytic. NFV4-VC is Vehicle Counting.
image_jpeg
string
Image of captured object in base64 string.
node_num
int
The node number computation. If you only have a single instance, thennode = 0.
pipeline_data
object
The output of each analytics.
area_name
string
Area name of the captured event.
confidence
float
Confidence the vehicle classification.
label
string
Classification of the captured vehicle.
primary_text
string
The vehicle classification prediction.
secondary_text
string
This key was left empty.
stream_address
string
The original address of stream/camera.
stream_id
string
The unique id of the assigned stream/camera.
timestamp
string
Unix timestamp of the detected vehicle.
event_id
string
Timestamp with 8 char hash
People Counting
Key
Type
Description
analytic_id
string
Type of analytic. NFV4-PC is People Counting
image_jpeg
string
Image of captured object in base64 string.
node_num
int
The node number computation. If you only have a single instance, thennode = 0.
pipeline_data
object
The output of each analytics.
area_name
string
Area name of the captured event.
confidence
float
Confidence the person classification
label
string
Classification of the captured person
primary_text
string
The people classification prediction.
secondary_text
string
This key was left empty.
stream_address
string
The original address of stream/camera.
stream_id
string
The unique id of the assigned stream/camera.
timestamp
string
Unix timestamp of the detected person.
event_id
string
Timestamp with 8 char hash
People Analytics
People Analytics with Attributes
Crowd Estimation
Key
Type
Description
analytic_id
string
Type of analytic. NFV4-CE is Crowd Estimation.
image_jpeg
string
Image of captured object in base64 string.
node_num
int
The node number computation. If you only have a single instance, thennode = 0.
pipeline_data
object
The output of each analytics.
area
string
Area name of captured the crowd.
estimation
float
Estimation number of people in the crowd.
interval
int
Interval of data dumping.
primary_text
string
Estimation number of people in the crowd.
secondary_text
string
Area name of captured the crowd.
stream_address
string
The original address of stream/camera.
stream_id
string
The unique id of the assigned stream/camera.
timestamp
string
Unix timestamp of the estimation dumping
event_id
string
Timestamp with 8 char hash
PPE Detection
id
string
Event id where object appear on frame
helmet
boolean
True or False. to check wether helmet appear with object
helmet_location_on_head
boolean
True or False. to check wether helmet is wore on head
vest
boolean
True or False. to check wether vest is wore by object or not
glasses
boolean
True or False. to check wether glasses is wore by object or not
Water Level Monitoring
Pre-Flood Monitoring (NFV4D-FLOD)
Person Smoking Detection (NFV4D-PSMO)
Person with Handphone Detection (NFV4D-PPHO)
Smoke and Fire Detection (NFV4D-FISO)
Person with Firearms Detection (NFV4D-PPFA)
Vandalism Attempt Recognition (NFV4D-VMVD)
ATM Burglary Incident Recognition (NFV4D-VMBG)
Road Crash Monitoring (NFV4D-RCMA)
People Fighting Recognition (NFV4D-VMFG)
Riot Recognition (NFV4D-RIOT)
Tutorial to test WebSocket, you can try here:
Or you can connect via Vanilla Websocket instead
Connect to Vanilla WebsocketLast updated
Was this helpful?