Русский ↗Discuss your project
Home/Services/Machine Learning/MST ScanVision
MST ScanVision · Computer vision

Codes in motion.
Data in your system.

Automatic QR and Data Matrix recognition in video streams. Connect marking checks with product accounting in manufacturing, warehousing and logistics.

Discuss your requirementsHow it works
QR and Data MatrixVideo streamLocal deployment
Illustration of an industrial camera, conveyor and recognition server
Video stream → Accounting event
QR and Data MatrixFor manufacturing and logistics
Marking becomes an event

Packaging in the frame.
Its code in your system.

ScanVision reads two-dimensional codes from images for use by enterprise systems. We configure the control point for your line, equipment and accounting process.

Cameras and videoExisting or new sources
QR and Data MatrixMarking identification
IntegrationRecognised code delivery
Inside your networkLocal service

Receiving and dispatch

Send product codes to warehouse accounting as items pass the control point.

Marking checks

Recognise codes on a production line and connect them to system operations.

Traceability

Use the identifier, time and control point for further product movement tracking.

Manufacturing
Agriculture
Food industry
Logistics
Construction
Pharmaceuticals
Textiles and consumer goods
Chemical industry
Three stages of an event

See.
Recognise.
Deliver.

The service connects video sources to enterprise information systems. Select a stage to explore its role.

01 · Video stream

Start at the control point.

Connect the video source and configure the image region where marked objects appear.

  • USB and industrial cameras
  • IP cameras with RTSP or HTTP streams
  • Video files for configuration and testing
For multiple cameras, deploy separate service instances with individual settings.
Beyond perfect labels

Real objects.
Real conditions.

Configure image processing for your materials and capture conditions. Challenging cases form part of pilot validation.

Tilt and perspective

Image geometry correction for codes positioned at an angle.

Contrast and print quality

Image preparation for low contrast and blurred printing.

Creases and curves

Recognition testing on folds, deformed labels and curved surfaces.

Your equipment

Select the region of interest, scale and processing parameters for each camera.

Test on your samples. Line speed, code size in the frame, lighting and marking condition affect results. Quality and performance are recorded after testing.
Fits into your process

The recognised code.
The required action.

Send results to ERP, WMS, MES, 1C or a custom service. Multiple consumers can receive events from one control point.

  • Redis Pub/Sub result delivery
  • Separate channels and settings per camera
  • Integration with your information systems
  • Service log and recognised-code log
ERP, WMS, MES and 1C connections are designed for your process. Accounting actions, processing rules and event storage are agreed during integration.
From control point to accountingWithin your infrastructure
CameraVideo stream
→
MST ScanVisionRecognition
→
Redis Pub/SubOutput
↓
ERP / 1CProduct accounting
WMSWarehouse operations
MESManufacturing
What was recognisedCode contents
WhenRecognition time
WhereCamera / Control point
Integration diagram. Accounting system connections form part of the project.
Enterprise deployment

Your cameras.
Your infrastructure.

The software runs as a local service. A container simplifies environment preparation, while Linux tools manage the service.

Linux x86-64Server or workstation
DockerContainer with Ubuntu 22.04 LTS environment
Redis 5.0+Recognition event delivery
4 CPU coresBase requirements
8 GB RAMBase memory requirement
2 GB disk spaceMinimum for the image and logs
Resources depend on camera count, video parameters, line speed and log retention. Decoding settings are selected after testing.

For the pilot:
video from your line.

Test typical objects and difficult cases. We need capture conditions, code samples and the intended use of results.

  • Camera, resolution and connection conditions
  • Movement speed, distance and lighting
  • Code types and typical defects
  • What your system should do after reading

For the solution:
measurable criteria.

Before testing, agree what counts as successful recognition and the performance your process needs.

  • QR and Data Matrix validation programme
  • Quality on an agreed sample set
  • Workload and video source operation
  • Test result delivery to your system
Industrial recognition experience

More than a diagram.
Development in practice.

Two stages of a QR video recognition project from the MST portfolio.

Recognition in streaming video
Stage one · Development

Recognition in streaming video

An application for automatically reading codes from a video stream.

Read the project
Challenging production line conditions
Stage two · Development

Challenging production line conditions

Image processing and decoding improvements for a production workflow.

Read the project
Technical information

Learn more about
MST ScanVision.

Public documents from the product page. Deployment terms and current scope are confirmed for each project.

Functional specificationsWord document · Open original ↗Installation and deploymentWord document · Open original ↗Operating manualWord document · Open original ↗
How we work

From your line
to a configured service.

Solution scope and pricing depend on configuration and are calculated individually.

  1. 01

    Introduction

    Discuss the task and requirements. Sign an NDA if needed.

  2. 02

    Solution selection

    Review initial conditions, agree a pilot and project plan.

  3. 03

    Agreement

    Record scope, terms, timing and responsibilities in the contract.

  4. 04

    Implementation

    Configure the service and integration, then verify agreed workflows.

Questions answered

Discuss the essentials.
Before launch.

Which video sources does ScanVision support?+

USB and industrial cameras supported by Linux, IP cameras over RTSP or HTTP, and video files. Specific source compatibility is checked during solution preparation.

Can we connect several cameras?+

Yes. Each control point runs a separate service instance with its own configuration and camera identifier. Resources are selected for the total workload.

Does the system handle damaged marking?+

Image preparation settings help with tilt, blurred printing, creases and curves. Acceptable defects and recognition quality are tested on your samples.

How are results sent to our system?+

Through Redis Pub/Sub. Consumers subscribe to a camera channel and receive events. Connection, processing rules and accounting system storage are agreed during integration.

Is a GPU required?+

The documented base requirements do not specify a GPU. CPU, memory and service instance counts depend on video sources and test results.

Can the service run in our local network?+

Yes. ScanVision runs on Linux x86-64 in a Docker container, uses Redis for events and is managed as a system service.

Start with your challenge

Shall we examine
your marking?

Tell us about your production line or control point. We will prepare a recognition and accounting system integration test plan.

  • Objects, code types and capture conditions
  • Cameras, line speed and control point count
  • Accounting system and intended result
This email address is being protected from spambots. You need JavaScript enabled to view it.
Solution scope, timelines and terms are agreed for your project.

Loading the contact form…

Your name *
Company name *
Phone number *
+1
Search
    Email *
    Message *

    By clicking the Send Request button, you consent to the processing your personal data.

    Send request