Industrial Computer Vision
for metal manufacturing.
Edge AI for automated inspection, measurement and production monitoring — built for demanding metal-industry environments.
MetalEdge AI — developed and supplied by Zero One Labs.
Surface detail
Zero One Labs is a member of ZKP
Zero One Labs, developer of MetalEdge AI, is a member of the Polish Forging Association (Związek Kuźni Polskich, ZKP). ZKP is a member of EUROFORGE.
Learn more about ZKP →Industrial vision across the production line
Inspection, measurement and monitoring for forging shops, foundries, steel mills and metal-component manufacturing. Each use case is qualified against its imaging conditions and acceptance criteria.
Surface & product inspection
Visual inspection of forgings, castings and machined metal components for surface and product defects, with evidence for quality review.
Explore forging inspection →Counting & dimensional verification
Count metal components and verify selected dimensions against agreed tolerances, subject to camera calibration, installed views and measurement qualification.
Process and line monitoring
Monitor product flow and visible process states on metal-production lines. Present events and inspection evidence for operator review and production traceability.
What does MetalEdge deliver?
Reusable platform capabilities, configured for the specific production line, site, cameras, datasets, integrations and acceptance criteria.
Less scrap and fewer claims
Automated defect detection directly on the production line helps reduce defective parts and lower complaint costs.
Greater efficiency
Production monitoring can help operators identify deviations and respond within the agreed workflow.
Edge / on-premise first
Real-time inference can run locally at the plant without depending on cloud availability. Where required, MetalEdge can support customer-controlled hybrid infrastructure for management, reporting or integration.
Auditable evidence
Inspection results, confidence and image evidence support quality review and traceability. Reporting and retention are configured for the agreed use case.
Controlled model lifecycle
Operator feedback and production evidence can be used to improve datasets and models. Updates are validated, versioned and deployed under controlled release procedures with traceability and rollback.
How does MetalEdge work?
Performance is qualified on representative customer data, installed camera views and agreed operating conditions. Acceptance criteria are defined for each project.
Image capture
Industrial cameras capture the agreed product views. Lighting, placement and data quality are qualified for the inspection task.
AI analysis
Models analyze images locally on an edge / on-premise device. Accuracy and processing time are validated for the specific use case and operating conditions.
Reporting
Inspection results, confidence and image evidence are available for review and traceability in the agreed reporting workflow.
Operator workflow & integration
Inspection results, confidence and evidence are presented to the operator and can be integrated with production systems according to the agreed workflow and safety boundary. Human-in-the-loop (HITL) review is used where appropriate.
Engineering behind MetalEdge AI
MetalEdge AI — developed and supplied by Zero One Labs.
Michał P. Wójcik
MetalEdge AI product development
Creator of the MetalEdge AI product at Zero One Labs. Focused on building on-premise AI systems for industrial quality control in metal manufacturing — combining computer vision, edge AI and human-in-the-loop workflows.
LinkedInMetalEdge combines production software engineering with research experience in image-based defect recognition for metal components.
Research publication: M.P. Wójcik, K. Pawlikowski, Ł. Madej (2024). An evaluation of the capabilities of image-based metal component defect recognition with deep learning techniques. Computer Methods in Materials Science 24(3), 33–40.
Want to reduce production losses?
Discuss where visual inspection, measurement and production monitoring could support quality in your metal-manufacturing process.
Book a free demo →Let's talk about your plant
Leave your details and Zero One Labs will contact you to discuss the use case, data and requirements for a MetalEdge AI deployment.
Frequently asked questions
Find answers to the most common questions about MetalEdge.
How long does a MetalEdge deployment take?
A bounded single-station feasibility or pilot can typically start within a few weeks once data access, imaging conditions and acceptance criteria are confirmed. Enterprise deployments begin with qualification and site engineering, with the pilot and production schedule defined for the specific line and use case.
Do I need to replace existing production lines?
MetalEdge is designed to integrate with existing infrastructure. Qualification determines camera, lighting, equipment and integration requirements for the specific line. Replication to further lines or sites is controlled and requires validation of local conditions and acceptance criteria.
Does my data leave the plant?
Real-time inference can run locally at the plant without cloud availability. Optional customer-controlled hybrid infrastructure can support management, reporting or integration. Data flows, access and retention are agreed for each deployment.
How is MetalEdge priced?
MetalEdge is scoped per inspection use case, production line and integration boundary. A standalone inspection cell and a multi-line enterprise programme have materially different delivery requirements. Commercial scope is defined after initial technical qualification.