Visual Inspection

Visual inspection systems check parts, assemblies, or finished vehicles from photos or video to find defects and confirm that installed components match what the line expects.

Business impact

• Inspection time saved per aircraft — Boeing reports more than 17 hours saved per aircraft.
• Individual inspections performed — Ford reports 150 million individual inspections.
• Quality issues flagged — Ford reports 400,000 quality issues flagged.
• Specifications verified simultaneously — Renault reports more than 40 specifications verified at once.
• Inspections per vehicle — Stellantis reports 100 inspections per vehicle.
• Inspections per day — Stellantis reports more than 50,000 inspections per day.

Data requirements

• Images (image) — Used to inspect parts, assemblies, paint, welds, and fasteners.
• Video (video) — Used in the Airbus proof-of-concept and in line-side inspection setups.
• Inspector photographs (image) — Used in Boeing’s part-validation workflow.

AI methods and techniques

• Computer vision — Used to analyze photos or video of assemblies, parts, or vehicles for defects and configuration checks.
• Deep learning — Used in the Airbus proof-of-concept to recognize completed assembly steps.
• Machine learning — Used in Ford, Renault, and Volkswagen-related descriptions to support defect or configuration detection.
• Optical character recognition — Used by Boeing to extract part information from inspector photographs.

Industries

Real-world evidence

11 documented case studies on record.

Companies using this: Airbus, Boeing, Ford, GE Aerospace, GM, Northrop Grumman, Renault, Stellantis, Volkswagen, U.S. National Oceanic and Atmospheric Administration (NOAA), Procter & Gamble.