VISION 2026 returns to Stuttgart from 6 to 8 October, bringing together the technologies shaping industrial machine vision. For manufacturers, this year’s show offers plenty to explore, with AI featuring across quality inspection, robotic guidance, smart cameras and the development of vision applications.
Across the exhibition, suppliers are addressing familiar production challenges: finding defects on variable surfaces, reading difficult markings, handling mixed parts and putting reliable inspection into the hands of operators. Alongside those applications, AI agents, synthetic training data and embedded processing are changing how vision systems are developed and deployed.
That gives visitors several routes into the show. Manufacturers looking to automate an existing quality check can compare smart sensors and inspection software. Those developing robotic cells can investigate perception and picking. Engineering teams building their own systems can explore cameras, lighting, processors and tools for training and maintaining AI models.
The following preview brings together announced show demonstrations and the established AI technologies of participating exhibitors. Products described as prototypes or forthcoming releases remain at those stages.
MVTec: AI agents move into machine vision development
Hall 8, stand 8C56
MVTec will give visitors a substantial look at both the application and development of AI in machine vision, with its HALCON, MERLIC and Deep Learning Tool portfolio alongside the AI Vision Solver prototype.
The prototype tackles the engineering work involved in creating a vision application. A developer describes a task using natural language and images. AI agents interpret that request, select operators from the HALCON library and generate an executable application in HDevelopEVO, MVTec’s next-generation development environment.
At the stand, visitors will be able to select tasks such as barcode reading, matching and optical character recognition through a touchscreen. The system identifies syntax errors and logical gaps, then feeds that information back to the agents so they can refine the application.
The developer remains responsible for evaluating the result. For manufacturers and integrators, the demonstration offers a way to examine how AI assistance might shorten development work while retaining engineering oversight.
There is also fresh product news ahead of the exhibition. HALCON 26.11, announced on 1 October and scheduled for release on 18 November 2026, will introduce oriented bounding boxes and individual confidence thresholds for AI object detection. An ONNX Runtime AI² interface will support deep learning inference on Linux devices using 64-bit Arm processors.
HDevelopEVO will become available as an Early-Adopter version with that release. Its additions include an integrated MCP server giving the AI assistant access to HALCON-specific knowledge, alongside further development and runtime functionality.
For teams seeking a visual development workflow, MERLIC 26.10 is scheduled for release on 15 October, after the show. The update focuses on more secure communication and easier operation, extending the framework for creating and deploying machine vision applications without programming.
MVTec’s Application Tour adds a useful opportunity to see its software working within partner systems. The approximately hour-long English-language tour includes IMAGO’s AI-based PCB inspection, Micro-Epsilon’s 3D inspection, wenglor’s Deep OCR application, Basler’s virtual commissioning and digital twins, and MVTec’s AI Vision Solver. Registration is available through MVTec’s VISION event page.
For more detail on the development approach, read our coverage of MVTec’s AI Vision Solver.
Supar: bringing AR and AI-assisted inspection to the shop floor
Hall 10, stand 10H37
SuPAR, developed by Turkish company CDMVision and CDMTech GmbH, deserves attention from manufacturers inspecting assemblies, sheet metal parts and welded structures.
The platform overlays 3D CAD data onto the physical component through a tablet camera, allowing an inspector to compare the intended design with the part in front of them. Guided inspection workflows and digital reporting provide a way to record findings against the component.
SuPAR AI adds deep learning to selected repetitive checks within that workflow. Its modules cover hole detection, spot weld inspection and verification of stud presence and installation. Targets that appear missing, incorrectly positioned or incomplete can be flagged as the operator progresses through the inspection.
This approach is particularly relevant where an inspector needs to move around a large component or assembly. The operator retains responsibility for the inspection and the final acceptance decision, with performance dependent on training data, imaging conditions and validation for the specific application.
Supar also has two technical presentations scheduled:
- 6 October, 11:40–12:00: “AR and Deep Learning for Visual Quality Control”, Industrial VISION Days, Hall 8, stand 8C70.
- 7 October, 13:00–13:20: “On-Device Deep Learning with CAD-Based AR Tracking”, Scientific VISION Days, Hall 8, stand 8C50.
Both will be presented by Celaleddin Subaşı. Visitors can explore the inspection workflow at Supar’s stand throughout the exhibition.
Zebra: inspection, conveyor monitoring and robotic handling
Hall 8, stand 8B40
Zebra’s announcements span AI inspection and the movement of goods through production and logistics operations.
The NS42 AI smart vision sensor will gain deep learning classification and segmentation capabilities alongside anomaly detection and Deep OCR. Its onboard neural processing unit supports AI tasks, including reading embossed characters, laser markings and text with inconsistent contrast.
Zebra will also demonstrate Iris GTX smart cameras working with Aurora software to identify conveyor blockages, providing image confirmation to help operators distinguish genuine obstructions from false alarms.
For robotic applications, the company will unveil the newly branded Aurora Vision Guided Robotics Studio, with demonstrations combining AI and 3D vision for warehouse and logistics operations. Visitors will also meet the Photoneo team, now part of Zebra, and see the new CV70 CoaXPress camera for demanding image acquisition tasks.
For manufacturing readers, the relevant applications include part identification, label verification, anomaly inspection and vision-guided handling between processes.
IDS and DENKweit: AI processing within the camera
IDS: Hall 8, stand 8C60
IDS will present new camera platforms, embedded vision developments and AI-enabled additions to its uEye Live range.
The AI-capable models use neural networks from DENKweit for tasks including object and position detection, parcel handling and code reading. Processing at the edge offers a route to applications where image analysis needs to take place close to the camera.
IDS will also show developments in event-based imaging with Prophesee, alongside its Nion 3D time-of-flight camera and a new version of the Ensenso N active stereo camera family. These technologies address different parts of a vision system, from capturing fast-changing scenes to acquiring depth data for robotics and measurement.
A GMSL multi-camera innovation project will demonstrate processing and combining images within an embedded system before transmitting the results over a single cable.
DENKweit also warrants a separate stop for its DENKnet image analysis software. Its portfolio covers AI-based visual inspection, recognition and measurement, giving visitors an opportunity to discuss the model development behind applications such as the IDS collaboration.
Basler: AI inspection, simulation and agent-assisted engineering
Hall 8, stand 8D50
Basler will demonstrate AI-powered OCR and barcode recognition combined with dimensioning of moving objects, bringing identification and measurement into one application.
Its development technologies are also relevant for teams planning new inspection systems. Vision simulation allows camera, lens and lighting configurations to be tested virtually, with synthetic training images generated before the physical hardware is deployed.
Basler will additionally present an agentic development concept combining a vision simulation agent with a pylon coder agent. Visitors can examine how simulation and AI assistance might support the engineering process.
Compact 2D and 3D systems for autonomous mobile robots and high-throughput camera demonstrations complete the company’s wider show offering.
SICK: AI tools for configurable inspection
Hall 10, stand 10D41
SICK’s show programme combines its 2D and 3D vision portfolio with new AI tools in NOVA.
The software developments include anomaly detection, object detection and classification, alongside AI Blob Finder for identifying relevant image regions using neural networks. These capabilities are relevant to inspection tasks where product appearance varies or defects are difficult to describe through fixed rules.
SICK will also present its Inspector8xx, Ruler3000 and Ranger technologies, allowing visitors to investigate the relationship between image acquisition, AI analysis and integration into an industrial process.
Baumer: local AI inspection in an integrated sensor
Hall 8, stand 8D40
Baumer will present the VeriSens VS200, combining established vision tools with AI in an integrated smart vision sensor.
The company specifies up to 40 inspections per second, with local data processing and configuration through a built-in web interface. For manufacturers looking at a defined quality check, that creates an opportunity to investigate a compact inspection installation and its commissioning requirements.
Baumer will also demonstrate its industrial camera range and high-bandwidth Ethernet imaging. Those systems are relevant to larger inspection installations where several cameras must deliver images efficiently to the processing hardware.
Siemens and 36ZERO Vision: connecting AI inspection to production
Siemens: Hall 10, stand 10F11
36ZERO Vision: Hall 10, stand 10B24
Siemens’ announced VISION portfolio includes Inspekto, Visual Inspection Cockpit, Industrial AI Suite and Industrial Edge. Together, they give manufacturers several routes to discuss AI-based visual inspection and its integration into production systems.
36ZERO Vision adds another relevant connection. Its Professional App is available on Siemens Industrial Edge, bringing AI-based defect detection and a no-code workflow to that platform.
The software is designed for integration with new and existing quality processes, with data processed close to the machine. Visitors assessing a factory deployment can discuss camera compatibility, application management and how inspection results connect to the wider production environment.
Exaktera: the lighting behind reliable AI inspection
Hall 10, stand 10C50
Exaktera brings together Advanced illumination, autoVimation, LUMIMAX, ProPhotonix, VisioCablePro and Z-LASER, covering lighting, lasers, optical measurement and protective camera enclosures. The group will introduce its unified brand at VISION 2026.
Its relevance to AI inspection includes an established collaboration with Siemens. LUMIMAX Smart Light technology forms part of the Siemens Inspekto visual inspection system, helping provide the image conditions required by the AI analysis.
For manufacturers inspecting reflective metal, complex shapes or surfaces with changing appearance, this is a useful reminder to examine illumination alongside the inspection model. Exaktera offers a place to discuss that image acquisition work and the practical protection of cameras on the factory floor. The Siemens collaboration is an existing application; a specific Inspekto demonstration at Exaktera’s VISION stand has not been confirmed.
Allied Vision: AI across cameras, 3D sensors and inspection software
Hall 10, stand 10F30
Allied Vision’s stand will bring together several routes to industrial AI inspection.
The Alecs VSWIR smart camera combines visible and short-wave infrared imaging with NVIDIA Jetson Orin processing. Its demonstration is relevant to inspection tasks where material properties or contents require information beyond a conventional visible image.
Other applications include Gocator 2D systems using GoPxL AI and automated optical inspection combining cameras, acquisition hardware and eVision software.
For visitors developing a complete inspection station, the stand offers a view of how sensing, processing and software can be assembled into an application.
Aqrose: industrial AI and vision-language models
Hall 8, stand 8B50
Aqrose will present its industrial AI portfolio alongside vision-language model applications.
Its inspection technologies cover defect detection, anomaly detection and segmentation, with deployment options including industrial PCs and NVIDIA Jetson hardware. The company’s AI Review System adds review and traceability functions around inspection results.
The vision-language work gives visitors another development to investigate: how image interpretation and language-based interaction can support industrial inspection. Manufacturers can use the stand discussions to examine the specific task, deployment requirements and validation behind those applications.
Mech-Mind: perception for robotic picking and handling
Hall 10, stand 10E24
Mech-Mind will demonstrate its “Eye-Brain-Hand” approach, connecting 3D perception, decision-making and robotic action.
Its announced applications include handling challenging objects such as reflective metal, black rubber, transparent items and thin components. These are relevant to manufacturers investigating bin picking, loading and mixed-part handling.
The company will also introduce additional 3D cameras, including the ULTRA M-GL, UHQ-040-GL and RANGE XL-GL. Visitors can discuss which sensing and software combination suits their part geometry, working distance and handling task.
SINTRONES: Computing for AI inspection
SINTRONES will showcase the computing hardware behind demanding AI inspection applications, including electronics and semiconductor defect detection. Its stand will feature the new ABOX-5222E with GPU expansion, liquid-cooled ABOX systems and compact platforms demonstrating multi-camera AI inference, object detection and distance estimation. For machine builders and vision integrators, the focus is on processing camera data and running AI models reliably in industrial environments.
Where: Hall 8, Stand 8D14.
e-con Systems: depth perception and embedded AI
Hall 10, stand 10H60
e-con Systems will launch NetraVis, an active stereo 3D depth camera aimed at autonomous mobile robots, automated storage and retrieval systems, humanoids and pick-and-place robots.
The camera combines two global-shutter sensors with synchronised RGB imaging and an onboard inertial measurement unit. For manufacturing visitors, its relevance lies in the depth information required for navigation and robotic perception.
The company will also show STURDeCAM31FSH, a heated-optics GMSL camera developed for conditions such as cold-storage automation. Its lens heater addresses icing and condensation as robots move between different temperatures.
Separate demonstrations will feature edge AI number plate recognition and pedestrian detection. These provide examples of e-con’s embedded AI capabilities, while NetraVis and heated optics are the more direct fit for many factory and warehouse applications.
Specim: AI inspection using material information
Hall 10, stand 10D80
Specim will show how hyperspectral imaging can add material information to machine vision.
Its RETEX laboratory demonstration, using the FX17 camera and AI analysis, will address textile fibres, blends and impurities. The company will also demonstrate mineral applications with the SX25 and preview further camera and software developments.
For manufacturers, the attraction is the ability to investigate differences that may be difficult to identify from colour and shape alone. Applications can include material sorting, contamination detection and verification of product composition.
More AI inspection suppliers to include in the visit
The wider exhibition includes specialist software developers, smart camera suppliers and application partners. Several offer established AI portfolios alongside their confirmed show participation, so visitors should discuss the exact products and demonstrations available at each stand.
| Exhibitor | Relevant technology | Why it matters to manufacturers |
|---|---|---|
| wenglor | AI Lab, uniVision 3 and AI-capable smart cameras and controllers | Cloud-based model training connected to execution on industrial vision hardware. |
| IMAGO Technologies | Embedded vision hardware and HALCON-based AI PCB inspection | A confirmed MVTec tour application combines deep learning with rule-based checks for electronic components and connectors. |
| NeuroCheck | Industrial vision software incorporating AI and deep learning | Relevant to configuring inspection applications and integrating cameras into production systems. |
| Neurocle | Neuro-T model training and Neuro-R deployment software | A visual workflow for developing and deploying deep learning inspection models. |
| KEYENCE | IV4 vision sensor with built-in AI | An established option to discuss for compact inspection applications; KEYENCE lists VISION among its events. |
| SmartMore | AI-based visual inspection and manufacturing systems | Relevant to defect inspection, including battery applications combining AI, 2D and 3D analysis. |
| EVT Eye Vision Technology | EyeVision applications and AI-assisted 3D inspection | Its show programme includes a technical presentation on intelligent weld inspection. |
| Teledyne Vision Solutions | Astrocyte AI training, Sapera and Sherlock | An established software route covering anomaly detection, classification, object detection and segmentation. |
| Inventio.ai | Industrial App | Local, no-code AI quality control connecting cameras, automation equipment and enterprise systems. |
| TorqueAGI | Physical AI for perception, manipulation and inspection | Relevant to visitors exploring AI reasoning for robotic handling and assembly. |
| STEMMER IMAGING | AI vision components, software and application engineering | A place to discuss component selection and the engineering of a complete inspection system. |
| RAUSCHER | Vision software, acquisition hardware and systems for deep learning | Included in the official AI guided tours, with an established portfolio supporting AI inspection. |
| Hikrobot | VisionMaster deep learning tools | Established functions include difficult character reading and surface defect inspection. |
Synthetic data and specialist AI developers
Training data remains a practical consideration for manufacturers developing inspection models, particularly when defects are rare or a new production line has yet to generate representative images.
VISION’s Start-up World and specialist exhibitors offer several approaches to that work, alongside tools for model evaluation and assistance at manual workstations.
| Exhibitor | Technology to investigate |
|---|---|
| Sightwise | Synthetic-data-based visual inspection, with applications including forged rims and die-cast components. |
| Medabsy | Virtual development of vision systems and controlled generation of synthetic training data. |
| Artificial Pixels | Digital twins, simulated vision applications and synthetic datasets for AI training. |
| Tensorleap | Analysis and validation of industrial AI models, including failure modes, edge cases and changes in production data. |
| Memorence AI | Memorence Operagents and Memorence Suite, with AI support for manufacturing operations and manual workstations. |
| Rosepetal AI | Classification, detection, anomaly inspection and OCR trained around production applications. |
| ONE WARE | ONE AI, a visual development platform that creates application-specific models for deployment across PC, embedded and FPGA hardware. |
| 24 VISION | Neural-network-based automotive quality control, including product configuration checks. |
| AI4IV | Bio-inspired sensing and perception technology designed for AI applications in robots and industrial automation. |
These suppliers address different stages of an AI project. Synthetic data providers can help prepare training examples. Model analysis tools can expose weaknesses before or during deployment. Workstation systems can apply vision to tasks that continue to involve an operator.
The hardware that brings AI into industrial systems
Processing hardware is another substantial part of the exhibition. For teams building an inspection station, robot or embedded product, the choice of computing platform affects camera connectivity, latency, power consumption and integration.
| Exhibitor | Relevant AI and vision technology |
|---|---|
| Beckhoff | Neural networks and machine learning algorithms executed within TwinCAT runtime, alongside controller-integrated machine vision. |
| Vision Components | VC EvoCam with MediaTek Genio processing for embedded AI, alongside its MIPI camera portfolio. |
| AAEON | Announced demonstrations include physical AI for quality control with Awentia, AI video analytics and Jetson-based systems. |
| ARBOR Technology | Edge AI demonstrations using the ARES-1983H-AI platform and accelerators from MemryX, DEEPX and Axelera AI. |
| Avnet Silica and DEEPX | On-device inspection, vision-language processing and deployment to embedded systems, with demonstrations at stand 10D17. |
| Macnica Europe | Announced robotic vision and edge AI demonstrations drawing on its sensor, processor and engineering partners. |
| PHYTEC | phyVIP, an embedded vision integration platform supporting camera connections, scalable processing and AI acceleration. |
| Alkeria | Announced demonstrations of AI processing directly within camera systems. |
| SiMa.ai | Physical AI processing and development technologies; its established portfolio includes Modalix and Palette Neat. |
| Connect Tech | NVIDIA Jetson platforms and camera integration for embedded vision, robotics and industrial AI. |
| Advantech | Edge AI computing and integrated vision systems for image acquisition, analysis and industrial deployment. |
| Neousys Technology | Rugged embedded computers and GPU-based edge AI systems for machine vision and automation. |
| Vecow | Industrial vision computers supporting AI processing, camera interfaces and inspection workloads. |
| Emergent Vision Technologies | Announced AI-powered PCB anomaly detection within its high-speed imaging demonstrations. |
| LUCID Vision Labs | High-bandwidth camera acquisition and GPU connectivity for demanding processing systems, alongside its wider smart-camera portfolio. |
| Pleora Technologies | Imaging transport, embedded vision and connectivity for AI-enabled systems. |
Imaging partners including FRAMOS and Opto are also included in the organiser’s AI guided tour programme, providing further opportunities to discuss practical applications and the components behind them.
Planning a useful visit
For a manufacturer arriving with an inspection problem, a representative part and a clear description of the defect will help make stand discussions productive.
Useful questions include how many training images are required, how the system handles new product variants, where images and models are stored, and how inspection results are passed to the PLC or production software. Ask to see difficult examples, including acceptable variation and defects close to the acceptance limit.
For robotic applications, bring information about part materials, presentation, cycle time and the required handling accuracy. For embedded systems, discuss camera count, interfaces, environmental conditions and the processing budget.
The official VDMA AI guided tours run on all three show days, beginning at 14:00 and finishing at approximately 16:00. They are conducted in English, are free for exhibition visitors and require registration because places are limited. The published routes include MVTec, IDS, Siemens, Allied Vision, IMAGO, NeuroCheck, RAUSCHER, FRAMOS, Opto and EVT across the three days.
With inspection software, operator-assisted AR, robotic perception and embedded AI represented across the halls, VISION 2026 gives manufacturers an opportunity to compare several ways of solving the same production problem and to examine the engineering required to make each one work.
Frequently asked questions
When and where is VISION 2026?
VISION takes place at Messe Stuttgart, Germany, from 6 to 8 October 2026.
Is MVTec’s AI Vision Solver a released product?
MVTec is presenting it as a prototype at VISION 2026. HALCON 26.11 and the HDevelopEVO Early-Adopter version are scheduled for 18 November.
What does VISION cover?
The exhibition covers industrial machine vision, including cameras, sensors, lighting, optics, image processing software and complete inspection systems. Applications include quality control, robotic guidance and AI-based image analysis.
Are guided tours available at VISION 2026?
Yes. Tours cover electronics and semiconductors, aerospace and defence, automotive and batteries, and AI in machine vision. The VDMA AI tours run daily from 14:00 to approximately 16:00, are conducted in English and are free for exhibition visitors. Advance registration is required as places are limited.
Further reading
