RobVision: AI-powered vision system for autonomous automation in the industry and logistics.

With RobVision, we present an advanced system that sets new standards in industrial automation. Our AI-based image processing system enables autonomous work even in unstructured and dynamic environments - where traditional robotics reach their limits.

Use cases & applications

Contact sales

CAD model picking

Object detection and pose estimation models are trained directly from CAD data – ideal for precise handling of complex parts.

Label picking

Models are trained with just a few images – optimized for picking flat objects such as labels or sheets.

Vision-based depalletizing

For custom-colored boxes, models are trained using individual images; for standard brown boxes, pre-trained models are available for immediate use.

Coming soon –
Bin picking / Chaotic picking

CAD-based training will enable efficient handling of randomly arranged objects.

Future Vision –
One-Shot Learning

Object detection and pose estimation without extensive training – enabling instant adaptability.

Experience seamless flexibility and efficiency like never before.

RobVision uses advanced computer vision and machine learning techniques to enable robots to operate autonomously in unstructured and dynamic environments. Whether it's picking up and stowing from an orderly 2D plane or a chaotic assortment of parts, this system increases uptime because it is able to work in dynamic environments. Experience seamless flexibility and efficiency like never before.

Get a RobVision demo
image Robvision camera
camera vision 3d
clamp 0.80
clamp 0.55
clamp 0.27
clamp 0.60
clamp 0.80
clamp 0.55

Automatically adapts to changing scenarios.

Experience a new level of flexibility: RobVision does not require constant reprogramming and automatically adapts to changing scenarios. This significantly reduces implementation time and considerably increases the overall efficiency of your automation processes.

Get a RobVision demo

RobVision - The vision system for smart autonomous robot solutions.

The RobVision System offers unparalleled autonomy, enabling robots to perform complex tasks without human intervention in dynamic and unstructured environments. Leveraging advanced computer vision and machine learning, RobVision continuously adapts to changing conditions, eliminating the need for constant reprogramming. This autonomous capability streamlines operations, significantly reduces implementation time, and enhances overall efficiency in automation processes.

Get a RobVision demo
robot vision solution

User flow - How RobVision works

Purchase RobVision

Receive robot with equipped camera and installed RobVision Software

  • More cost effective than competition

RobCo Studio: Prepare and Train Your Model

Upload CAD Model: Upload your CAD file and select a texture. For label picking, simply upload a single image of the part.

Start Training: Automatic synthetic data generation based on your CAD model or image. AI model training in the cloud (fully automated).

Deploy Trained Model: Seamlessly deploy the trained AI model over-the-air to the RobCo Vision Controller.

  • No real pictures needed

  • Ready-to-use AI model for a specific object with just 2 clicks

RobFlow: Configure and program your application

Camera Calibration: Calibrate the camera for precise detection and pose estimation.

Solution Configuration: Select the camera, trained part model, and define parameters like Region of Interest (ROI).

  • User-friendly interface for fast and easy application configuration

Workflow integration

Program the Automation Flow: Use two intuitive function nodes to connect the vision system to your robot flow — no coding required.

Monitor the System: Visualize detection and pose estimation results directly in RobFlow.

  • Significant reduction of programming complexity

  • Quick implementation

RobFlow interface, user selection of RobVision
Camera calibration in RobVision software
Solution configuration in RobVision software
Analysis of the pose estimation Visualization for debugging purposes

AI Technology Deep Dive

User Upload

A CAD model (or image) is uploaded to the system.

Synthetic Data Generation

Thousands of photorealistic, synthetic images are automatically created based on the CAD model.
‍
This allows the AI to learn how the object looks — without needing real-world images.

Automated AI Model Training

Using the synthetic data, an AI model is trained to classify the object, segment it, and accurately estimate its 3D pose.
‍
The entire process is fully automated in the cloud — no manual labeling or tuning required.

Deployment

The resulting AI model, trained entirely on simulated data, can robustly detect and locate the object in 3D space in real-world environments.

RobVision in the market:
Smarter, faster, more flexible.

Get a RobVision demo
Vertical integration

Seamless hardware and software interaction for maximum efficiency.

Superior performance

AI-driven precision and adaptability.

End-to-end support

Comprehensive assistance from setup to ongoing optimization.

Attractive pricing

Competitive and scalable pricing models tailored to your needs.

Ease of use

Eliminate time-consuming parameter tuning – simply upload your CAD model and start the training process.

Simplified programming

Only two intuitive function nodes needed for quick implementation.

Successful companies rely on our automation solution.

Experience RobCo live.

01

Product demo

02

Technical concept

03

Calculation of amortization and savings

OSZAR »