Pro
Computer Vision video analytics

Act on the risk your cameras already see, in real time

Connect your existing CCTV cameras to an AI video analytics system that monitors PPE, restricted zones, attendance and vehicle movement.

See which modules fit

An illustration of the control panel: four camera tiles — PPE monitoring, restricted zone, vehicle plate recognition and an access point. Beside them a list of event types: PPE check, restricted zone, vehicle entry. This is a demo interface, not real camera footage.

The problem

When monitoring depends on a person, problems are found too late

The site has cameras, but they only record. Whatever happened is usually discovered after the incident, while reviewing the archive. Here are the five situations most common on industrial sites.

One person cannot watch every zone

Twenty or thirty cameras feed a video wall. The operator sees only a few tiles at a time; the rest go unwatched.

PPE violations are never recorded

A worker entering without a helmet or vest is usually not logged anywhere. Yet an inspection asks for documented proof that the rules were followed.

Attendance is kept on paper

Arrival and departure times are written by hand or entered into a logbook. Working out late arrivals and early departures per shift takes separate effort.

The archive has to be reviewed for hours

Finding when an incident happened means scanning hours of footage. And that is repeated not only for investigations but for routine reports.

Vehicle movement is not centralised

Who entered, how long a truck waited, how long loading took — this data sits in scattered logbooks or does not exist at all.

Platform

One video analytics platform for safety, attendance and vehicles

All three directions run in a single system: one server, one control panel and shared reporting. Modules are selected for your site — you do not have to enable everything at once.

AI video analytics platform

Occupational safety

PPE monitoring, restricted-zone entry, falls and inactivity, smoke and fire, lone working, presence under a suspended load.

Attendance and working time

RFID/NFC card entry confirmed by camera, late arrivals and early departures, reporting by shift and crew.

Vehicles and machinery

Licence plate recognition, entry and exit log, machine-to-person proximity, on-site speed and loading time.

Modules

Choose the modules your site needs

Each module is enabled separately and has its own rules. Below is what each module detects and where it is useful.

Occupational safety

  • PPE — protective equipment monitoring

    What it detects
    Whether a helmet, high-visibility vest and safety harness are present.
    Where it helps
    Entry points, construction sites, production floors, work at height.

  • Unauthorised entry into a restricted zone

    What it detects
    A person or machine crossing a defined zone boundary.
    Where it helps
    Around equipment, conveyors, high-voltage yards, excavation areas.

  • Falls and prolonged inactivity

    What it detects
    A person falling or staying motionless longer than a set period.
    Where it helps
    Lone-working areas, underground work, night shifts.

  • Smoke and fire detection

    What it detects
    Signs of smoke or open flame appearing in the frame.
    Where it helps
    Warehouses, flammable material stores, electrical rooms.

  • Lone working monitoring

    What it detects
    Only one person left in a zone — when the rule requires two.
    Where it helps
    Underground work, inside tanks, around high-risk equipment.

  • Presence under a suspended load

    What it detects
    A person inside the area beneath a crane or hoist load.
    Where it helps
    Loading areas, crane working zones, warehouses.

Attendance and working time

  • Card entry confirmed by camera

    What it detects
    An RFID/NFC card read, matched against the person in the frame.
    Where it helps
    Turnstiles, checkpoints, office and workshop entrances.

  • Late arrivals and early departures

    What it detects
    Deviations from the scheduled shift times.
    Where it helps
    Shift-based production and sites organised into crews.

  • Reporting and export

    What it detects
    An attendance sheet broken down by shift, crew and contractor staff.
    Where it helps
    HR and accounting — exported to Excel or PDF.

Vehicles and machinery

  • ANPR — licence plate recognition

    What it detects
    The registration plate of a vehicle entering or leaving.
    Where it helps
    Checkpoints, warehouse gates, quarry exits.

  • Machine-to-person proximity

    What it detects
    An unsafe distance between moving machinery and a pedestrian.
    Where it helps
    Warehouse aisles, quarry roads, loading areas.

  • Speed and loading time

    What it detects
    Movement speed on site and how long loading or unloading takes.
    Where it helps
    Internal roads, docks, logistics terminals.

Some modules (smoking, phone distraction and similar) are agreed separately, depending on site conditions and camera angle.

Accuracy depends on camera angle, lighting, distance and the real test set. Final figures are determined after the technical audit and the pilot trial.

Industries

A monitoring scenario for your industry

The set of modules differs by industry. Below is the direction most sites enable first.

Mining

In open pits and underground work: PPE monitoring, restricted-zone entry and proximity to machinery. At the exit, haulage vehicles are logged through ANPR.

Construction

Helmet and vest monitoring, safety harness for work at height, the area under crane loads and attendance of contractor crews.

Manufacturing plant

PPE checks at the workshop entrance, restricted zones around conveyors and equipment, attendance by shift and turnstile integration.

Warehouse and logistics

Forklift-to-pedestrian proximity in aisles, loading time at the dock, ANPR at the gate and detection of vehicles without clearance.

Oil and gas

Lone working monitoring, access to tank and process zones, smoke and fire indicators, and vehicle logging at the checkpoint.

Energy

Unauthorised entry onto high-voltage yards, PPE monitoring, prolonged inactivity and an access log at substations.

Architecture

Video is analysed inside your site

The camera stream never leaves the site: analysis runs on a server on the premises, and only the notifications and reports you allow go outside.

Architecture

  1. Existing RTSP cameras
  2. On-site GPU server
  3. AI modules
  4. Control panel
  5. Alerts and reports

What on-premise gives you

  • Video never leaves the site

    Camera streams are not sent to a cloud. For industrial and government sites with strict confidentiality requirements this is usually the deciding condition.

  • Not dependent on the internet

    Analysis continues even if the connection drops. The internet is only needed for remote panel access and outbound notifications.

  • Lower latency

    Analysis runs on the same network as the cameras, so an alert arrives faster than a round trip to a cloud service.

YOLO11TensorRTByteTrackPythonFastAPINode.jsRedisReactWebRTCRTSP

Systems connected as the project requires

TelegramSMSSirenaTurniket1CHRSCADAVMS

The list above shows what is possible. Which of these your project includes is decided by the technical audit.

Infrastructure

Are your existing cameras good enough?

A module has to be able to see the object in the frame. That makes distance, angle, lighting and the size of the detected object decisive — resolution alone is not a sufficient criterion.

What else is considered beyond resolution

  • Angle — from which direction the camera views the object
  • Lighting — night, backlight and changing conditions
  • Size of the detected object — a helmet and a licence plate need different framing
  • Frame stability — vibration, dust and weather exposure

Final camera requirements are determined after a site audit and measurements. At some points the existing camera is enough; at others a replacement or an extra camera may be needed.

Control

A single control view for management

Output from every module is collected in one panel — separate programs and logbooks are no longer needed.

Live video

Camera streams are viewed in the browser and the required camera is selected from a list.

Violation log

Every record is stored in a list with its time, zone and module.

Photo and video evidence

Each record carries a frame and a short clip — they can be attached to a report.

Zones and rules

Zone boundaries and when each rule is active are configured in the panel.

Reports

Reports by shift, crew and module, exported to Excel or PDF.

Roles and audit log

Access to data is limited by role and actions inside the panel are recorded.

Rollout

From a pilot point to the whole site

The project starts at one point: the module is tested there in real conditions, and once the result is confirmed it is extended to the remaining zones.

  1. 01

    Technical audit

    We review the site: which process is left unmonitored, where the incident risk is highest and which module delivers value first.

    Stage outcomeA list of monitoring points and the module selection for the first stage.

  2. 02

    Assessing cameras and monitoring points

    Existing cameras are checked: do distance, angle, lighting and frame quality meet the module's requirements.

    Stage outcomeA written statement of which cameras suffice and which need replacing or adding.

  3. 03

    Data and model trial

    The module is tested and tuned on real frames captured at your site.

    Stage outcomeThe result at the pilot point and settings that reduce false alerts.

  4. 04

    Platform and integrations

    The server is installed, modules enabled, the panel configured and the agreed systems connected.

    Stage outcomeA working platform with configured zones and alert channels.

  5. 05

    Launch, training and handover

    Safety and HR staff are trained on the panel and the system is formally accepted.

    Stage outcomeA handed-over system, a guide and the technical support period.

The timeline is set by the chosen modules, the number of cameras and the complexity of the site — an exact schedule follows the technical audit.

Pricing

What determines the cost of a Computer Vision project?

The plans below are a starting point for the software. Cameras and server hardware are quoted separately.

Standard

Basic computer vision solution

15,000,000 UZS

Who it suits

Sites starting with one or two monitoring points.

  • Image recognition
  • Object detection
  • QR and barcode reading
  • Basic image analysis
  • Web interface
  • API integration
  • Basic reports

Premium

Professional computer vision solution

30,000,000 UZS

Who it suits

Sites that need several zones, face recognition and real-time analysis.

  • Everything in the Standard plan
  • Face recognition
  • Video analysis
  • Motion detection
  • Custom model training
  • Extended image analysis
  • Integration with existing systems

Premium+

Enterprise computer vision solution

50,000,000 UZS

Who it suits

Sites with many cameras that require integration with corporate systems.

  • Everything in the Premium plan
  • Deep learning models
  • Real-time analysis
  • Enterprise system integration
  • Custom APIs and integrations
  • High level of security
  • Server built for high load

What sets the final price

  • Number of cameras and monitored zones
  • The set of modules selected
  • Required server and GPU capacity
  • Whether cameras need replacing or adding
  • Integrations: Telegram, turnstiles, 1C, HR, SCADA, VMS
  • The pilot stage and tuning the model to your site
  • Technical support period and terms

The price excludes cameras, installation, network equipment and server hardware — these are quoted separately for the site. An exact estimate follows the technical audit.

See all pricing plans
Questions

Questions about Computer Vision video analytics

Answers to the questions asked most often before a technical audit.

IP cameras that provide an RTSP stream can be connected. But not every camera suits every module: distance, angle and lighting have to meet its requirements. That is why cameras are checked point by point during the audit, and it is stated in writing which are sufficient and which need replacing.

Let's identify which safety process on your site can be automated

During the technical audit we review your existing cameras, distances, lighting and monitoring points. Then we prepare the matching modules and a project proposal.

Discuss the project