ORPAON AI-SOP
Assembly error-proofing system: turn work SOPs into inspectable processes
Target customers
Discrete manufacturers whose assembly, packing or boxing workstations carry the risk of missing parts, wrong assembly or skipped steps; factories that want vision-based work compliance inspection.
Typical symptoms
- Manual visual inspection misses defects, leading to customer complaints
- New employees' work is not standardized; SOP execution relies on self-discipline
- The traceability chain is missing the "work actions" link
Solution path
Define workstation SOPs → select the visual inspection route (per workstation needs) → inspection judgment and NG evidence capture → vision data into the traceability chain.
Core features
- HkVisionPro (multi-layer assembly, layer-by-layer feature-point inspection): deployed route — automatic flow-through when the current layer is all OK, NG alarms with capture and evidence archiving; usable for assembly error-proofing delivery
- OrpaonVision (motion recognition standalone route): YOLO part detection, hand key-point and sequential action recognition; self-contained project packages and performance regression gates are in place — close to commercial, but still requires scenario acceptance against workstation data, cycle time and acceptance criteria
- OrpaonSOP (VisionMaster plugin route): progressing — joint validation available per workstation
- AI-SOP-main (integration prototype): a prototype used for route validation
Engineering evidence
The HkVisionPro route (multi-layer assembly, layer-by-layer feature-point inspection) has deployment evidence; group-level maturity: extensible.
The motion recognition route is close to commercial and delivered after workstation-scenario acceptance; the plugin route offers joint validation per workstation.
Expansion path
Single-workstation error-proofing → multi-workstation replication → vision data into the traceability chain → the motion recognition route advances with scenario acceptance results.
Applicability conditions
- Workstation action sequences can be defined
- Stable lighting or modifiable imaging conditions
- Advancing by route and by scenario is acceptable
Non-applicability conditions
- Requests for a "one system, plug and play across all lines" promise
- Treating the motion recognition route as already deployed at scale at customer sites
- Scenarios with no samples / no imaging conditions
Pilot approach
Deployed route: pick one assembly workstation for a scenario-acceptance pilot; motion recognition route: run a scenario-acceptance evaluation first. Pilot scope, timeline and acceptance criteria are confirmed by both parties after investigation.
FAQ
Is AI-SOP a fully mature, complete product?
The product family is delivered and expanding — HkVisionPro deployed, OrpaonVision close to commercial (acceptance per scenario), OrpaonSOP and AI-SOP-main progressing; let's discuss by route.
Can you promise detection accuracy?
Accuracy must be confirmed in scenario acceptance against specific workstations, samples and acceptance criteria — no promises without basis.
How is this different from ordinary visual inspection?
It focuses on "work action compliance" — turning SOP execution into inspectable, evidence-grade data.
Get the 15-minute demo and pilot checklist
Every product comes with a 15-minute demo; after the demo we provide a unified pilot boundary checklist.
Book a Product Demo