Scratches, cracks, chips and surface discontinuities
Use raking, diffuse, bright-field or dark-field illumination as the surface demands, then validate the defined discontinuities across material, finish, tooling and supplier variation.
Metals and machining
How can AI inspection support cast and machined components?
Evaluate surfaces, presence, orientation and optically measurable conditions with qualified plant cameras, factory-owned edge inference and a customer-chosen cloud account.
In brief
The right architecture depends on the visual signal, process risk, line conditions and validation standard, not the industry label alone.
Candidate inspection families
Where plant teams may find a useful starting point.
These are categories to evaluate, not claims of deployed performance.
Feature presence and orientation
Confirm that holes, slots, inserts, threads or other visible features are present and oriented correctly while fixtures hold the part within the qualified view.
Optical dimensional estimates where appropriate
Measure configured edges, gaps or positions only when calibration, lens distortion, working distance, focus and part presentation can be controlled to the required tolerance.
Process evidence and part presentation
Use visible residue, tool marks, orientation or handling evidence to support a defined decision; do not ask the model to infer a condition the image cannot show.
Readiness
Qualify the production environment, not an isolated image.
A credible review covers optics, lighting, field of view, product presentation, cycle time, edge compute, interfaces, response and acceptance criteria.
- Collect representative images across normal production variation
- Document what a false accept and false reject mean operationally
- Agree how uncertain outputs are handled
- Define change triggers for new products, materials and conditions
Use Cases
See inspection workflows connected to this environment.
Each use case connects the inspection workflow to the questions a plant should answer before production deployment.
Use Case
Surface defect inspection
A workflow for learning the appearance of acceptable surfaces and configured discontinuities.
Review the workflow
Use Case
Dimensional gauging
A workflow for estimating configured dimensions when optics, calibration and part presentation are controlled.
Review the workflow
Use Case
Oil-leak detection
A workflow for identifying visible or fluorescent evidence on a component before it moves downstream.
Review the workflowQuestions answered
What metals and machining teams ask first.
Which inspection should this industry start with?
Choose a visible, economically relevant condition with accessible images, a clear decision, manageable integration and a practical validation set. The best first use case proves the complete operating workflow.
Can one model be copied across lines?
Not safely by default. Cameras, lighting, products, fixtures, suppliers and process conditions can differ. Reuse the deployment package and governance, then validate locally.
Does United Vision replace every existing vision system?
No. Stable, deterministic checks can remain an excellent fit for traditional machine vision. United Vision should be evaluated where learned visual patterns and a reusable plant-owned lifecycle are useful.
Can we use our existing cameras and edge devices?
Yes, when the existing equipment is suitable and supported for the inspection. The review qualifies image quality, optics, lighting, timing, interfaces, edge capacity and the plant environment; United Vision does not require one proprietary camera and edge stack.
Where do inference and cloud services run?
Production inference always runs on a supported factory-owned edge device. United Vision cloud services run in a supported cloud account chosen and owned by the customer, keeping the real-time decision close to the line.
Can the factory add more inspection use cases without another software license?
Yes. One factory license supports unlimited United Vision use cases within the licensed factory. Each use case still requires suitable evidence, imaging, integration and production validation.
Map an inspection opportunity in metals and machining.
Bring the product, image source, process condition, decision and current inspection method. We will use those facts to test fit.