Industrial Vision Agent

MODELS · VISION · X-RAY · DEVICE–EDGE

Built on the industrial vision model series and X-ray inspection model, orchestrated by crusbro agent-OS for capture, decisions, review, alerts and QMS/MES writeback. Misses first. Evidence retained. Metrics signable.

Product structure

Domain models are independently acceptable. The station-level vision agent owns recipes, gray-zone handling, dual-sign and system writeback.

LayerWhatRole
Runtimecrusbro agent-OSPermissions, audit, versions, human gates
ModelsVision series · X-rayDetect, segment, measure, film reading
AgentThis pageStation loop, review, writeback, iteration
Industrial vision agent: station loop Edge capture → edge inference → agent orchestration → QMS/MES writeback; gray zone to review; samples feed retrain and edge publish. DEVICE Edge capture Camera · lighting DR · ultrasound · 3D Sorter Alarm MODELS Domain models Vision series X-ray film Edge infer Hot update VISION AGENT Agent orchestration Recipe station · threshold Review gray · dual-sign Alert stop · changeover Writeback QMS · MES EVIDENCE Evidence & writeback Image · params Model version QMS · MES · PLC Sample feedback

Industrial vision model series

Built for high reliability, scarce defects and strong traceability. Select by station; share versioning and acceptance rules.

LineTypical tasksMethod focusAcceptance focus
Surface defectsMicro-cracks, coating, tool marks, edge chipsSelf-supervised pretrain + segmentation + anomaly fallbackRecall first; FN thresholds per defect class
Assembly checksMissing, wrong part, offset, damageDetection + multi-view fusion + CAD geometry checkMissing/wrong as red lines; coverage and cycle time
Internal NDTDelamination, porosity, inclusionsUltrasound / thermography multimodal + segment / anomalyAgreement vs metallography or CT
Connector / cablePin retreat, tilt, sheath damage, shield breakMetric learning + few-shot + defect synthesisModel migration speed; cycle time; first-pass yield
Precision metrologySize / GD&T / profile3D point cloud + registration + auto GD&TAccuracy vs CMM and cycle-time gain

X-ray inspection model

Digital radiography (DR) for welds, castings and structures. AI first pass + certified dual-sign sampling. Reports are archive-ready.

ItemDetail
DefectsPorosity, slag, lack of fusion / penetration, cracks; graded by size and density
StackDR enhance → detect / segment → severity grade → PDF/A report
ComplianceDual-sign: AI read + certified reviewer sampling; escalate on disagreement
ExplainabilityBoxes / masks and attention maps archived with the film
AcceptanceAgreement with certified readers, FN on critical defects, cycle time, report completeness

Agent capabilities

  • Select camera / lighting / model / thresholds per product and station
  • Gray zone → multi-angle re-inspect; still uncertain → human confirm
  • Consecutive anomalies → stop-line / changeover prompts
  • Write decisions, images and model versions to QMS / MES
  • Return corrected samples for retraining and edge hot-update

Constraints

  • Misses first: false positives can be reviewed; FN is a red line by class
  • Scarce defects: good-part pretrain, anomaly fallback, synthetic augmentation
  • Bound evidence: image tied to S/N / lot / station
  • Human co-sign: dual-sign retained on high-risk ops

Station protocol / acceptance

Sign per station. Do not replace with a vague accuracy claim.

DimensionExample
FN / FPPer-class thresholds; red-line items 100% human review
Cycle timeLoad + imaging + inference + unload; inference latency agreed separately
AgreementX-ray scenes vs certified readers
ChangeoverNew model / defect adaptation window by scenario
TraceabilityImage + params + model version + disposition replayable

Deployment

ItemGuidance
SensingIndustrial camera / line scan / DR / ultrasound or thermography by model line
EdgeIPC or inference box; low-latency on the line
CenterTraining, version registry, drift monitoring, canary release
SystemsQMS / MES / PLC / SPC; alerts and stop-line commands configurable

Back to foundation models Whitepaper Book a demo