Early Access

Basler Vision Simulation

Simulation for Faster Vision Solution Development

With over 250 simulated cameras, matching lenses and lighting, you can develop image processing applications virtually – without time-consuming physical prototypes. Thanks to sensor-realistic images, you can generate the image data for training your AI.

Test Vision Simulation
  • Faster

    Vision systems can be created, tested, iterated and optimized virtually – without complex hardware setups
  • Precise images

    Sensor-realistic images respond correctly to changing settings – with exceptional accuracy to real-world setups
  • Train AI

    Neural networks can be fully trained with automatically labeled synthetic images
  • Maximum product selection

    Realistically combine and compare over 250 cameras, 30 lenses and 14 lights

Develop using simulation

Make robust decisions with high correspondence between simulation and reality

Which field of view (FOV) can be best achieved with which camera? How does the scene change under different lighting conditions? Basler Vision Simulation provides the answers.

Optimal design of vision systems

Since our digital twins of vision products both correspond to actual hardware and provide comparable technical image data, we can determine the appropriate camera and optic combination:

  • Is the image sharp? Is the camera resolution sufficient? How large is the field of view coverage compared to the object?

The simulation provides quick and reliable answers while incorporating the ideal lighting products for the scene:

  • What is the light distribution of the selected light? How should the lighting be set up to avoid reflections? At what lighting angle are defects most noticeable?

Realistic image simulation

We use our decades of experience with image processing products to create realistic image simulations:

  • Realistic modeling of cameras in terms of resolution, pixel pitch, field of view (FOV), and exposure time

  • Realistic behavior of optics with regard to focal length, depth of field, aperture, and transmission

  • Simulation of lighting properties such as light distribution, total light emission and control of lighting intensity (Basler SLP controller)

What our customers say

Simulation first-approaches can significantly simplify and accelerate complex system development and deployment. Basler Vision Simulation makes it easier to adopt simulations and increases their precision. This is especially relevant in applications where high levels of vision accuracy and reliability are required.
Falk-Moritz Schaefer
Falk-Moritz Schaefer
Technology Architect, Accenture Innovation

Typical areas of application

Which applications is vision simulation particularly suited for?

Logistics: Code reading

Simulation and evaluation of code readability by image processing systems under various conditions, such as viewing angle, surface, lighting, and edge cases.
Détection de toutes les canettes

Object detection for quality assurance

Simulate object detection and perform presence checks to evaluate reliable OK and NOK decisions. Take into account varying parameters, such as camera position, lighting, and object characteristics.
Robotique

Robotics

Simulation of camera-based robotic applications used to evaluate image acquisition, position detection, and object localization in automated processes.

Synthetic images for AI training

Create high-quality training data in no time

Generating synthetic data is revolutionizing the training of AI models. Instead of collecting real data in a costly and time-consuming manner, you can create large quantities of perfectly annotated training data with the push of a button.

Pixel-perfect and perfectly labeled images

Pixel-perfect, perfectly labeled

Each synthetic image comes with precise annotations – fully automated and without manual labeling effort.

Data sets can be flexibly varied

Maximum diversity

Data sets can be flexibly varied: Both 3D objects and vision system parameters can be randomized.

Flexible expansion of training data

Scale flexibly

Expand your datasets at any time: New scenarios and annotations can be generated quickly, eliminating the need to wait for real image capture.

Early Access

Basler Vision Simulation

Overview

With the Early Access program, you receive free access to our simulation software. You test important features, explore real-world use cases, and provide feedback that directly influences functionality, usability, and performance.

An NVIDIA RTX™-enabled graphics card is required.

The program is limited.

Early Access

Basler Vision Simulation

Features

Simulated images can be displayed live, saved, and used for image processing

Simulated images can be displayed live, saved, and used for image processing


Simulated products can be controlled exactly like in the real world

Simulated products can be controlled exactly like in the real world


Detailed information about the simulated products is available

Detailed information about the simulated products is available


Product selection and configuration can be controlled via script

Product selection and configuration can be controlled via script


Support for NVIDIA Omniverse™ Replicator for domain randomization
and synthetic data generation

Support for NVIDIA Omniverse™ Replicator for domain randomization
and synthetic data generation


Interface to Basler pylon Software to test image processing with the
Basler pylon vTools image analysis software

Interface to Basler pylon Software to test image processing with the
Basler pylon vTools image analysis software

On request

Product scope

Simulated cameras

Simulated cameras

252 2D area scan cameras
from the ace and ace 2 series

Simulated lenses

Simulated lenses

30 fixed focal lenses

Simulated lighting

Simulated lighting

14 bar and ring lights
with white LED color

Note:

Customizations of features or additional simulated products are possible within the scope of paid customer projects.

Logo NVIDIA GTC 2026

From physical vision systems to OpenUSD:

Building digital twins for virtual commissioning

Learn how we use OpenUSD and integrate with NVIDIA Omniverse™ to build digital twins of machine vision systems for virtual commissioning. Discover workflows for simulation, configuration, synthetic data generation, AI training, and deployment.

Watch session now

Virtual commissioning

Here's how: Commission your vision system in just a few steps

We offer a comprehensive end-to-end workflow for virtual commissioning:

  1. Procedural asset engineering: A pipeline for generating textured 3D assets from input data.

  2. System design in simulation: Configuration of a vision system in simulation.

  3. Synthetic data generation: A pipeline for creating balanced, large-scale training datasets.

  4. MLOps & production runtime: A trained AI model together with the software runtime for production use.

  5. Deployment: Deployment of the complete solution to real hardware.

Do you still have questions and would like to speak with our project team?


Contact our Vision Simulation project team

FAQ: Answers about Basler Vision Simulation

The Early Access program of Basler Vision Simulation is free of charge.

With the Early Access version of the Basler Vision Simulation, we provide a simulation software for the development of vision applications. The software can be installed and used free of charge and starts with a selection of simulated products. It allows users to get familiar with the features and gain practical experience. At the same time, we aim to collect user feedback as part of the program in order to further develop and improve the software.

There are 252 area scan cameras, 30 fixed focal lenses, and 14 lights available.
The following features are available:

2D area scan cameras
ace and ace 2 models

Fixed folcal lenses

Bar lights and ring lights

3D models

3D models

3D models

Resolution

Focal length

Light distribution

Pixel pitch

Depth of field

Total light emission

Field of view (FOV)

Aperture

LED color: White

Exposure time

Transmission

Control of lighting intensity (Basler SLP Controller)

Balanced, high-quality training data is achieved by systematically varying the relevant parameters during data generation. These parameters include:

  • Object randomization: This involves varying the object’s 3D shapes, arrangement, and materials, as well as introducing procedurally generated defects.

  • Vision system configuration: This involves varying the camera position, light position, light intensity, exposure time, and resolution.

Deliberately combining these parameters creates a diverse dataset that reflects different real-world conditions and improves the robustness of the trained model.

Operating system

  • Windows 11 (64-bit)

  • Linux Ubuntu 22.04, or newer

Hardware requirements

  • GPU: NVIDIA RTX™-capable GPU (GeForce RTX™ 3070 or better recommended), latest version of NVIDIA GPU Driver

  • CPU: Multi-core processor (comparable to Intel i7 or AMD Ryzen 7)

  • RAM: Minimum 16 GB (32 GB recommended)

  • Storage: Minimum 10 GB free storage space for installation

  • Display: Resolution of 1920 px x 1080 px or higher

No, the Early Access version of the Basler Vision Simulation is not intended for productive use in operational business.

You would like to discuss Vision Simulation with us? Our project team looks forward to hearing from you:

Contact our Vision Simulation project team


Do you have technical questions? Please contact our technical support:

Contact support
Nouveau

Develop vision applications virtually:

Test Basler Vision Simulation now!