Access and quick start
Access to OMNI Platform
You can work with OMNI Platform in two ways: via the web interface or using Integration GraphQL API.
Access to web interface
Follow the platform URL in your web browser and enter user email and password. As a result, you will be redirected to the Admin panel of web interface. OMNI Platform deployment domain is specified in platform.values.yaml file in ingress.rules.gateway.host
field. The platform URL is http://<ingress.rules.gateway.host>
.
The user password and email are available at platform.secrets.json platform configuration file in the platform-user-secret.default_email
and platform-user-secret.default_password
fields, respectively.
Language and theme selection
By default, the web interface supports English and Russian languages. To change the language, just set the appropriate one in your web browser.
The web interface has a light theme enabled by default. You can change the theme by clicking the Theme icon at the top right side of the web interface.
Admin panel
The main system parameters and functionality are displayed in the Admin panel.
Profiles
The widget displays information about created profiles and watchlists. You can get access to Profile and Watchlist pages through widget elements or side menu.
OMNI Agents
The OMNI Agents widget shows the number and statuses of OMNI Agents connected to OMNI Platform. For more details about Agent, click on this widget or side menu to move to OMNI Agents page.
Activities and events
Widgets show the number of received OMNI Platform events and activities.
Resources
Resources widget provides instant access to special platforms for API testing (Platform API and Event Service API). To test API without special platforms, you need to get an access token (authorization key) and add it to HTTP request header, as in the example { TOKEN: 1689*** }
.
Info
The widget displays information about the version of OMNI Platform, the version of the face recognition model, the allowed number of stored activities, and the sample storage period.
Access to Integration GraphQL API
To get access to Integration API, follow Integrations section.
Quick check
After signing-in to the web interface you can test the functionality of OMNI Platform by simply following the steps below:
Install, run and log in to OMNI Agent – special software that processes video files and video streams from cameras to detect, track and identify faces and silhouettes.
Add profiles to the database and attach them to the watchlist.
noteTo upload multiple face images to create profiles, you can use a special script, which is located in the on_premise/upload_script/ folder of OMNI Platform distribution kit. The script accepts a file with paths to images, and then sends requests to create profiles in the platform. To attach profiles to the watchlist when uploading, you need to pass the watchlist id to the
profile_group
parameter when starting the script. The watchlist id is written in the watchlist card in OMNI Platform web interface.Wait until the profiles are uploaded to OMNI Agent (~1.5 minutes for every 100,000 profiles). The logs at ~/.local/share/OMNIAgent/log (Linux) or %LocalAppData%/OMNIAgent/log (Windows) should contain the line
*DatabaseTDV* database size: <size>
, where<size>
is the number of profiles added to the watchlist.Connect the camera in OMNI Agent web interface. When choosing a camera, try to consider the following parameters:
- The main flow of people should go directly to the camera, and not perpendicular to it.
- The light source should be behind the camera, not in front of it, so that faces are not shadowed in bright light.
After a camera is successfully added, OMNI Agent web interface should display video from the camera, face detections (bboxes) and additional information about people in the camera's field of view (their gender, age, emotions, etc.).
Ask a person who has a profile in OMNI Platform to enter the frame. Make sure that the identification event appeared on Events page of OMNI Platform web interface.