Skip to content
VisionObject

Releases

Dated product releases. Each entry says what changed, what it needs, a known limit, and the register entries it stands on. Releases are identified by date: a deployment updated on or after a date has that release, and its documentation is supplied with it.

Availability register

Releases

  1. Module switches, allowances and a licence for each organisation

    Number plates, appearance, AI scene reading and the weekly brief are switched on or off for each organisation. A switched-off module neither runs nor shows what it produced. Each organisation also has an allowance of people and enabled cameras, and a licence expiry after which it is locked until renewed.

    Needs:
    A platform administrator.
    Known limit:
    Switching a module off does not delete what was already stored. Face search is not among the switches: it is not offered in the current release.
    Updating:
    Existing organisations keep their current behaviour until a switch is changed.

  2. Admin console

    Organisations, camera sources and sharing, AI services, the mail relay and retention periods are managed in one console, with a guided first-run set-up.

    Needs:
    A platform administrator.
    Updating:
    Settings previously held in the deployment's configuration file are imported once with a supplied command.

  3. Accounts, organisations, roles and camera scopes

    Everyone signs in. Several organisations can share a deployment, each camera has an owner and can be shared, and each person has one of five roles and an optional list of cameras.

    Needs:
    Accounts created by invitation.
    Known limit:
    Roles are fixed, and sign-in is within the product: there is no single sign-on.
    Updating:
    A deployment without accounts runs a one-time set-up that creates the first administrator.

  4. System log, log export and one-year audit retention

    A system log records operational changes such as cameras going offline. The audit log and system log can be exported, and audit entries are kept for 365 days by default.

  5. Wall mode and HTTPS

    A full-screen wall with large on-picture labels, and HTTPS for the dashboard set up at deployment.

    Needs:
    For HTTPS, a host name for the deployment.
    Known limit:
    HTTPS is off by default and has been verified with a locally issued certificate only. A certificate you supply is not supported yet.

  6. Audit log

    Changes, and searches by plate, scene description and route, are recorded with who made them and when.

    Known limit:
    The log records that an action happened; it is not a tamper-evident seal.

  7. Weekly brief

    A weekly fact sheet for each organisation with a written summary, optionally sent by email. The figures are computed and kept first; if the summary cannot be written, the brief shows the figures and says why.

    Needs:
    An AI service and a mail relay configured for the deployment.
    Known limit:
    Aggregate figures, with the names of your rules, cameras, areas and watchlist labels, are sent to the AI service you configure.

  8. Area filter for Insights and the events list

    Insights and the events list can be filtered by camera, zone or department. An area that matches nothing returns nothing, never the whole estate.

  9. Plate history, near-match search, routes and the plate watchlist

    Each plate read is kept as a sighting. A plate can be searched exactly or allowing for a misread character, its sightings are shown as a route, and a watchlist raises an alert when a listed plate is read.

    Needs:
    The number-plate module.
    Known limit:
    The watchlist alert and a route with more than one stop are verified by automated tests and have not yet been confirmed on a live feed in our own testing.

  10. Map and vehicle tracing

    Cameras on a map, coloured by health, and a plate's sightings as numbered stops in time order.

    Needs:
    Camera locations entered by an operator.
    Known limit:
    A stop without a location is listed but not drawn. Lines between stops are inferred.

  11. Live status for each feed

    Each live feed says whether it is live, behind, stalled or lost, and analysis that falls behind drops old frames instead of running late.

    Known limit:
    A status describes the feed and its analysis, not the scene.

  12. Starter rules, per-camera exceptions and automatic muting

    A catalogue of starter rules, a list of cameras a rule should skip, and automatic muting of a rule on a camera where reviewers keep marking it false.

    Known limit:
    Most starter rules depend on AI scene reading. Automatic muting counts reviewed alerts only.

  13. Plate reads validated and voted across a sighting

    Plate text is read from the full-width frame, checked against valid plate formats, and settled by a vote across the frames of one sighting. A plate that cannot be resolved is skipped, not guessed.

    Known limit:
    On night and dawn street footage only a small share of vehicles produced a usable read in our own testing.

  14. Alert triage and camera health

    Alerts are marked Acknowledged or False positive, assigned and annotated; status can be set for a selection. Each camera is shown as online, degraded, offline or disabled.

    Known limit:
    There are no incident, escalation or resolution states.

  15. AI scene reading

    After a sighting, one still frame is read by an AI vision model that returns a description, up to eight tags and a suggested severity. Rules can fire on a tag, and sightings can be searched by their reading.

    Needs:
    An AI service configured for the deployment.
    Known limit:
    A reading arrives after the sighting, can be wrong, and sends a still of the whole camera view, with the plate as read, the camera's identifier, colours and the time, to the AI service you configure.

  16. Bulk camera registration

    Register a published list of cameras in one step and re-sync it.

    Known limit:
    One list format is supported.

  17. Install script for a cloud virtual machine

    An install script sets up a Linux virtual machine in your own cloud account and reports each part's health.

    Known limit:
    Not yet run end to end on a cloud machine in our own testing: a cloud deployment requires assessment.

  18. Continuous playback of uploaded footage

    All of a camera's uploaded clips play as one timeline, in real time or with the gaps skipped, with detections drawn on, event markers, speed control and keyboard shortcuts.

    Known limit:
    Covers uploaded footage only.

  19. Multi-clip uploads with start times from file names

    Upload several exported clips at once. Start times are read from the file names where the export includes them, and a failed analysis job can be re-run.

    Known limit:
    One export naming pattern is recognised; otherwise the operator enters the start time.

  20. Sightings grouped across cameras and clips

    Sightings that look alike are grouped, on request, as candidates for the same person or vehicle.

    Known limit:
    Grouping rests on overall appearance. Sightings that overlap in time are kept apart.

  21. Recorded footage, placed at the time it was recorded

    Uploaded video is analysed as fast as the hardware allows and placed at the time it was recorded.

    Known limit:
    Uploaded footage never raises alerts.

  22. Rules and alerts

    Zones, lines, dwell limits, counts and items left, with schedules, repeat limits and severity.

    Known limit:
    Zone exit, line and dwell rules are verified by automated tests and have not yet been confirmed on a live feed in our own testing.

  23. Detect, describe, search

    People, vehicles and carried bags followed within a camera, with a still where a frame was captured, colours and tags, and search by type, colour, camera and time.

See it working on footage like yours.

Tell us the workflow you care about. We will walk through it on demonstration footage, say what is available today, and plan a pilot on your own cameras.

Request a demoWatch an investigation