The complete system.
Designed to work together.
Reliova brings vision, mechanics, automation and software into one engineering scope, with clear interfaces from the part to the plant.
Capability across
the inspection system.
Each discipline contributes to a consistent production workflow.
Vision & measurement
Optical inspection, dimensional checks and feature-level decisions matched to the part and its production environment.
Fixtures & station design
Part location, handling, access and changeover engineered as part of the inspection system.
Controls & line integration
PLC and HMI integration, station sequencing, production interlocks and coordinated exception handling.
AI inspection
Application-specific defect recognition, evaluated against quality references and deployed with controlled versions.
Flexible inspection access
Robot and motion-based sensor positioning for complex geometry, product variants and multiple inspection views.
Quality data & traceability
Product-linked results, inspection history and plant-system connectivity for production and quality teams.
Connect the part,
the check and the next step.
An integrated station coordinates inspection with the product and production sequence.
Identify & present
Part identity, product recipe and controlled positioning.
Inspect & evaluate
Visual, dimensional and functional results for the required features.
Release & record
Station decisions, exception routing and a linked quality record.
Engineering for
everyday operation.
Performance includes how the system is used, changed and maintained.
Product changeover
Recipe management, variant identification and application-specific tooling for the intended product family.
Operator & maintenance access
Clear HMI workflows, service access, diagnostics and defined recovery procedures.
Plant connectivity
Interface coordination with customer PLC, MES and quality systems, including data exchange and failure recovery.
Controlled updates
Configuration records, software backups and review of changes that affect the inspection process.
Inspection formats
for the production task.
Configure equipment around access, handling, changeover and the checks required by the program.
Dedicated inspection station
Part presentation, sensing and test equipment arranged for a defined component or assembly family.
Robotic inspection cell
Controlled sensor positioning and coordinated part handling for features that require multiple views or flexible access.
In-line verification
Inspection coordinated with the existing production sequence, station controls and customer release workflow.
Technical depth.
Production context.
The application defines what the system must observe, how a decision is evaluated and how the result is used.
Observation before
interpretation.
Feature visibility, contrast, reflections and depth of field affect the information available to an algorithm. 2D imaging and local 3D sensing are considered against the feature requirement. Sensing, lighting and the fixture are evaluated as one acquisition process.
A datum and a decision
for each feature.
Gap, flush and installed-position checks need an agreed reference, units and tolerance. Reloading and part support are included when evaluating the measurement process.
Develop with scarce
defect examples.
Separate normal product variation from the target defect. Organize representative samples and evaluate data preparation methods while retaining independent reference parts for acceptance.
Small defects in
complex appearances.
Choose the detection or comparison method around feature scale, material texture and expected variation. Review false acceptance and false rejection by defect category.
Performance across
the complete cycle.
Evaluate compute load, image transfer, inference and station communication together. Software optimization is checked against inspection quality and the operating configuration.
A decision the line
can act on.
Product identity, completed-check status and equipment health contribute to release logic. Recovery and reinspection are defined alongside the normal production sequence.
Engineering outputs
that connect the disciplines.
The project scope identifies the documents and records needed for design review, integration and acceptance.
| Engineering area | Reviewable output |
|---|---|
| Vision & measurement | Feature coverage matrix, sensing concept and calibration or verification plan. |
| Mechanical & automation | Station layout, datum and fixture concept, sequence and interface definition. |
| AI & inspection logic | Defect definitions, reference-sample plan and application evaluation report. |
| Quality & software | Acceptance test plan, configuration record, result fields and reinspection workflow. |
Anomaly detection.
Within a complete decision.
A deviation becomes useful when it is evaluated against the product requirement and connected to an action.
Learn the acceptable
production envelope.
Plan representative material, appearance, configuration and presentation conditions. Use anomaly methods where they fit the task and available samples.
Combine the methods
the feature needs.
Use presence and identity checks, defect recognition, local measurement and functional-test integration according to the inspection requirement.
Evaluate the error
that matters.
Keep acceptance references separate from development data. Review missed defects and unnecessary rejects by category, including boundary cases.
Keep production
changes reviewable.
Associate results with the model and recipe used. Review changed materials, optics or configurations before approving an updated inspection process.
Compute where
the process needs it.
The inspection workload determines the local compute architecture, from smart cameras to industrial PCs and accelerator-based systems.
Budget the complete cycle.
Image acquisition, transfer, movement, inference and controller communication are considered together.
Plan failure behavior.
Define the station response to missing images, incomplete checks, lost identity or interrupted plant connections.
Separate learning and release.
Evaluate updated models and configurations before deployment; production release follows the approved decision process.
Make local results useful.
Exchange agreed results with the plant while aligning image retention, access and any further data use with the project scope.
Better observations.
Better-informed decisions.
We evaluate emerging methods against specific inspection tasks and production constraints.
Assisted interpretation
Explore visual and language models for contextual review, knowledge retrieval and explanation, with independent evaluation for the intended task.
A useful next action
Research the connection between uncertain findings, additional viewpoints and controlled robotic observation.
A development environment
Follow advances in scene simulation and synthetic data to explore conditions that are difficult or costly to capture.
A controlled production decision
Evaluate methods before release, preserve approved configurations and define when findings require review or a separate test.