MACHINE VISION SOFTWARE
Common Vision Blox (CVB): vendor-independent machine vision software
Common Vision Blox (CVB) is a machine vision software development kit from STEMMER IMAGING. It gives developers vendor-independent control over image acquisition, with processing, 3D and machine learning tools built on top, so applications stay portable across cameras, operating systems and programming languages.
Built for OEMs, machine builders and integrators with their own development teams.
A free CVB Camera Suite licence comes with every GigE Vision and USB3 Vision camera from STEMMER IMAGING. Everything else can be evaluated without a licence, or under a 30-day trial.
STANDARDS
GenICam, GigE Vision, USB3 Vision, RoCEv2
PLATFORMS
Windows, Linux, ARM and embedded targets
APIS
C/C++, .NET, Python
TRACKED RECORD
80,000+ licences in use, in development since 1997
New to CVB
What it is, what it does, and where it fits alongside the tools you already use.
Already using CVB
Downloads, licence key, documentation, sample programs and support.
OVERVIEW
What is Common Vision Blox?
CVB is a programming SDK for machine vision applications. Developers use it to acquire images from any GenICam, GigE Vision or USB3 Vision compliant camera, then process that data with CVB tools, with third-party libraries, or with their own algorithms.
The acquisition core is what most customers come for. It isolates the application from the camera manufacturer, so hardware can change without the software being rewritten. Everything above the acquisition layer is modular, so teams licence only the components they actually use.
What CVB consists of
Component
Function
CVB Image Manager
Core acquisition engine, GenICam, GigE Vision and USB3 Vision compliant camera control
CVB Camera Suite
Full Image Manager functionality, included free with every GigE Vision or USB3 Vision camera from STEMMER IMAGING
CVB Foundation Package
Entry-level acquisition licensing tier
CVB Polimago
Machine learning for search, classification and 3D pose estimation
CVB Match 3D (ICP and DNC)
3D point cloud matching and object positioning
CVB GevServer
Software-defined camera streaming with GEV 3.0 and GenDC compliance
CVB CodeReader
1D and 2D code reading library
Python Scripting Engine
Embeds Python logic inside C++ and .NET applications
MockTL
Camera-free development and reproducible testing
SBOM (CycloneDX)
Automatically generated software bill of materials
THE PROBLEM
Why acquisition independence matters
Every machine vision application needs an acquisition layer, and most development teams inherit theirs from whichever camera vendor they started with. That works until the camera changes.
A discontinued sensor, a supply constraint, a better-suited model from a different manufacturer, or a customer requirement for a second source: any of these means the acquisition code has to be rebuilt, retested and revalidated. The cost is rarely the licence. It is engineering time spent re-implementing something that adds nothing to the product, plus the delay that comes with it.
Vendor-independent acquisition removes the dependency. The application talks to a stable API, and CVB handles the standards, the transport layers and the manufacturer-specific behaviour underneath. Development teams keep their effort on their own image processing IP, which is where their competitive advantage sits.
CAPABILITIES
What CVB does
| Capability | What it does | What it means for your project |
|---|---|---|
Vendor-independent acquisition |
GenICam, GigE Vision and USB3 Vision compliance built into the acquisition core |
Add or swap camera hardware without redesigning the application |
High-throughput transport |
RoCEv2 over standard Ethernet, TurboDrive, MultiStream and GenDC |
CPU utilisation drops from around 50% to under 10% at data rates above 2GB/s |
3D and point cloud processing |
Match 3D ICP and DNC, gap filling, laser triangulation calibration, 2D lens undistortion |
Locate and measure objects in 3D without sensor-specific code |
Machine learning |
CVB Polimago, CPU-based search, classification and pose estimation |
Train usable classifiers from tens of images rather than hundreds, with no GPU required |
Code reading |
CVB CodeReader for 1D and 2D codes |
Reads codes that are rotated, damaged or poorly contrasted, using the same acquisition pipeline |
Software-defined camera |
CVB GevServer with GEV 3.0 and GenDC support |
Present any acquisition source, or several fused together, as one standards-compliant GigE Vision stream |
Rapid prototyping |
Python Scripting Engine and CVBpy |
Prototype processing logic in Python and call it from a C++ or .NET application |
Development without hardware |
MockTL simulates a standard-compliant GenTL producer from stored images |
Start development and run reproducible tests before cameras are on the bench |
Observability |
OpenTelemetry-format tracing, logging and performance metrics |
Feed CVB diagnostics into the monitoring stack you already run |
Compliance |
CycloneDX SBOM generated automatically in the build process |
Machine-readable component documentation for the EU Cyber Resilience Act, without a manual audit |
SCOPE
Where CVB fits
Compared with camera manufacturer SDKs
Vendor SDKs are built to get the best out of one manufacturer's hardware, and they do that well. CVB is built for the case where the hardware is not fixed. It works with any compliant camera, and it can sit alongside a vendor SDK rather than replacing it, so existing software does not need a full redesign.
Compared with image processing libraries
CVB is not a comprehensive processing library in the way OpenCV or HALCON are. It provides acquisition plus targeted processing tools, and it interfaces cleanly with any processing SDK. A common setup is CVB for acquisition with a processing library of the customer's choice running on top.
Compared with low-code and GUI-driven toolboxes
CVB is a programming SDK for teams with development capability. There is no click-and-configure environment. The Python Scripting Engine and the sample programs lower the entry point, and code generators cover much of what a visual editor would, but CVB assumes a developer.
CVB is a good fit when
- You have in-house or contracted development capability
- Camera hardware may change, or a second source is required
- You are building a product rather than a one-off installation
- Acquisition performance, throughput or latency is a constraint
CVB is not the right tool when
- You need a ready-to-run inspection application out of the box
- You are committed to one camera manufacturer for the product lifetime
- You need a GUI-driven configuration environment with no coding
- The application is simple enough for a smart camera or vision sensor
Download Common Vision Blox (CVB)
CVB supports both Windows and Linux for 32 and 64 bit. The fully implemented GenICam standard in conjunction with GigE Vision and USB3 Vision on all platforms provides complete hardware independence.
Key Advantages and Benefits
Seamless Integration
Interoperability to combine open-source libraries, multiple tools, custom code, and interfaces to build the best solution.
Future-proof Flexibility
True independence across platforms, hardware, programming languages and APIs enables maximum customisation and allows to cost efficiently evolve with technology.
Superior Acquisition Capabilities
Support for the latest acquisition technologies to maximize hardware capabilities, enabling state-of-the-art vision systems.
Cost-effective and Scalable
Accelerates your projects by eliminating repetitive coding tasks and utilizes a flexible licensing model customized for your specific requirements.
Advanced Machine Learning Capabilities
Fast AI-based OCR, search and classification tools requiring small training sets for fast and efficient development.
Extensive Development Resources
Comprehensive programming examples in multiple coding languages to help you develop your own application quickly and efficiently.
Get a 30 Day Software Trial License
Product Overview
Empowering Advanced Machine Vision Development
True Hardware and Platform Independence
Beyond a Typical SDK
Cutting-Edge Features and Capabilities
Flexible Architecture
High-Performance Processing
PRODUCT FAMILY
CVB modules
CVB Image Manager
The acquisition core behind every CVB application. Manufacturer-independent camera control, multi-stream acquisition and optimised transport.
CVB Camera Suite
The full Image Manager feature set, free with every GigE Vision or USB3 Vision camera from STEMMER IMAGING.
CVB Foundation Package
Entry-level licensing for acquisition-focused applications.
CVB Polimago
CPU-based machine learning for search, classification and pose estimation, trained from small image sets.
CVB Match 3D
Locate objects in point clouds and determine their position, with ICP for precision and DNC for CAD-based detection.
CVB GevServer
Turn any acquisition source, or several fused together, into a GEV 3.0 and GenDC compliant camera stream.
CVB Code Reader
1D and 2D code reading that holds up against rotation, damage and poor contrast.
Python Scripting Engine
Run Python processing logic inside C++ and .NET applications, with access to the wider Python ecosystem.
How CVB's RDMA Integration Delivers 25Gbps+ Performance
CVB Integrations
Sherlock 8 GenICam Driver
- Support for all GenICam-compliant cameras without vendor lock-in
- Flexible configuration with automatic or manual camera setup
- Robust acquisition reliability with CVB's optimized performance
- Available with any valid CVB Image Manager or free CVB Camera Suite license
Third-Party SDK Compatibility
- OpenCV: Direct image exchange with OpenCV for advanced image processing algorithms
- TensorFlow/PyTorch: Seamless handoff between CVB acquisition and AI inference engines
- Halcon: Integration through GenICam standards for high-level machine vision processing
- Other GenTL Consumers: Compatibility with any GenICam-compliant software environment
Custom Application Integration
- C++/C#/.NET: Comprehensive libraries for Windows and Linux development
- Python: Full Python bindings with type hints for improved development experience
- Qt: Seamless integration with Qt for cross-platform GUI applications
Industry Solutions
- Manufacturing and Industrial Automation
- Logistics and Packaging
- Food and Beverage
- Healthcare and Medical Imaging
- Security and Surveillance
- Agriculture and Environmental Monitoring
- Robotics and Automation
- Machine vision specialists
- OEM customers for vertical applications
- Universities and research organisations
- End-users with a requirement for high performance vision tools
- Inexperienced end customers with the need for programming support from machine vision experts
Get started right away!
CVB Camera Suite is a powerful image acquisition software that covers the entire Image Manager feature-set. It supports multiple operating systems, programming languages and advanced features such as RoCEv2, multi-stream acquisition, and TurboDrive technology.
CVB empowers machine vision professionals to create high-performance, future-proof applications with unparalleled flexibility, reliability and efficiency.
Latest Release
Common Vision Blox 15.0, October 2025
Common Vision Blox 14.01, July 2024
TECHNICAL
Platforms, standards and compatibility
Operating systems
Windows, Linux
Architectures
x64, ARM, embedded targets
Deployment
Native installation, Docker containers on Linux guests, Debian packages, RPM on request
Programming languages
C, C++, .NET (C#, VB.NET, Visual C++), Python
Camera standards
GenICam, GigE Vision, USB3 Vision, GenDC, GenTL
Transport technologies
RoCEv2 (RDMA over Converged Ethernet), TurboDrive, MultiStream, MultiPart, FlowSet Pools
Camera compatibility
Any GenICam, GigE Vision or USB3 Vision compliant camera, regardless of manufacturer
Interoperability
Combines with third-party processing libraries, open-source components and existing vendor SDKs
API stability
Stable C ABI and object-oriented APIs maintained across major releases
Documentation
Online documentation, sample programs in all supported languages, user forum
AUDIENCE
Who works with CVB
Organisations
OEMs and machine builders who ship vision as part of a product, machine vision integrators with their own development teams, and universities and research institutes building experimental setups. CVB suits organisations with dedicated development capability rather than end users looking for a finished application.
Roles
Engineering leads and product owners weighing vendor risk, implementation effort and long-term maintenance cost, alongside the developers and systems engineers who implement against the API.
Industries
Factory automation, automotive, semiconductor and electronics, logistics and intralogistics, print and packaging, food and beverage, pharmaceutical production, metrology, traffic, sports and broadcast, agriculture, and defence.
EVIDENCE
Proven in production
80,000+
licences in practical use worldwide
Since 1997
continuous development, with APIs maintained across major releases
50% → <10%
CPU utilisation with RDMA, at sustained data rates above 2GB/s
~80%
CPU utilisation with RDMA, at sustained data rates above 2GB/s
80,000+
licences in practical use worldwide
Since 1997
continuous development, with APIs maintained across major releases
50% → <10%
CPU utilisation with RDMA, at sustained data rates above 2GB/s
~80%
CPU utilisation with RDMA, at sustained data rates above 2GB/s
COMMERICAL
Licensing
CVB is licensed modularly, so you pay for the components you use and add more as the application grows.
CVB Camera Suite
Free with every GigE Vision or USB3 Vision camera supplied by STEMMER IMAGING. Covers the complete Image Manager feature set.
Evaluation
CVB can be evaluated without a licence
Trial licence
30-day full trial
Licence types
Hardware-bound, floating development, or runtime
Terms
Permanent or subscription
Binding
Camera-bound for GigE Vision (MAC address) and USB3 Vision (serial number plus vendor ID). Not tied to the PC, and no dongle required.
Camera-bound licensing means the licence travels with the camera. Replacing a host PC, moving a system between lines, or re-imaging a controller does not require a new licence.
NEXT STEPS
Get started with CVB
01
Download CVB
The full SDK, including sample programs in C++, .NET and Python.
02
Request your licence key
Free Camera Suite key for STEMMER IMAGING cameras, or a 30-day trial licence.
03
Run the Python samples
The fastest way to get images on screen and judge whether CVB fits your architecture.
04
Talk to an application engineer
The Technical Competence Centre supports feasibility questions, acquisition setup and architecture reviews.
RESOURCES
Documentation and support
Use-case driven structure with code examples for every supported language
Sample programs
Working examples in C++, .NET and Python, shipped with the SDK
Peer and STEMMER IMAGING support
First and second level support, with direct development team access for key accounts
Including RDMA in GigE Vision applications
Release notes
What changed in each version, including deprecated and removed features
Feature requests
Submitted through the regional wishlist process
Monthly newsletter covering CVB releases and machine vision technology
FAQ
Frequently asked questions
What is Common Vision Blox used for?
CVB is used to build custom machine vision applications. Developers use it to acquire images from industrial cameras and to process them, either with CVB's own tools or with third-party libraries. Typical applications include inspection, measurement, code reading, robot guidance and 3D analysis.
Which cameras work with CVB?
Any camera compliant with GenICam, GigE Vision or USB3 Vision, regardless of manufacturer. CVB does not require a vendor-specific driver or SDK.
What is the difference between CVB and a camera manufacturer's SDK?
A manufacturer's SDK is optimised for that manufacturer's cameras. CVB is manufacturer-independent, so the same application code works across camera brands. This matters when hardware is discontinued, when a second source is required, or when a better-suited camera comes from a different vendor.
Does CVB support Linux and ARM?
Yes. CVB runs on Windows and Linux across x64 and ARM, including embedded targets, and supports Docker deployment on Linux guests.
Which programming languages does CVB support?
C, C++, .NET (C#, VB.NET, Visual C++) and Python.
Is CVB free with STEMMER IMAGING cameras?
Every GigE Vision and USB3 Vision camera supplied by STEMMER IMAGING includes a free CVB Camera Suite licence, which covers the complete Image Manager feature set.
Can I try CVB before buying?
Yes. CVB can be evaluated without a licence, and a 30-day full trial licence is available on request.
Does CVB include deep learning?
CVB includes CVB Polimago, a CPU-based machine learning module for search, classification and pose estimation that trains from small image sets and needs no GPU. For deep learning frameworks such as TensorFlow, CVB interfaces with them through the Python Scripting Engine rather than replacing them.
Does CVB provide an SBOM?
Yes. A CycloneDX-format software bill of materials is generated automatically during the CVB build process and delivered as a single machine-readable file per release, supporting compliance with the EU Cyber Resilience Act.
Do I need to change my application to use RDMA?
No. Existing CVB applications benefit from RDMA performance once the feature is enabled, with no code changes required.
Start building with CVB
Download the SDK, request a licence key, or bring an acquisition problem to the Technical Competence Centre.
Talk to an application engineer
Benefit from our expertise
Turn our industry expertise and comprehensive services into your competitive advantage. Our experts will be happy to show you the potential for your company and work with you to realise tailor-made solutions to suit your requirements. We look forward to hearing from you and will get back to you promptly.