Computer Vision Development in Toronto

Manufacturing / Automotive Components

An Ontario-based automotive parts manufacturer (CAD $2.1B annual revenue, 4,500 employees, supplying stamped metal components, plastic moulded parts, and assembled modules to 8 OEMs including GM, Ford, Stellantis, Honda, and Toyota — operating 12 manufacturing plants across Ontario and Michigan) needed computer vision to automate quality inspection, reduce scrap, and meet the increasingly stringent PPM (Parts Per Million defects) targets demanded by OEM customers. (1) Quality inspection was manual: 280 quality inspectors visually examined parts on production lines across 12 plants — checking for surface defects (scratches, dents, porosity), dimensional accuracy (using gauges and CMMs for sample inspection), and assembly completeness (correct components, proper orientation, secure fastening). Manual inspection detected approximately 85% of defects — meaning 15% of defective parts reached customers. At 180 million parts per year, the 15% escape rate translated to 27 million potentially defective parts reaching OEM assembly plants (though most were caught by OEM incoming inspection). (2) OEM PPM pressure: OEM customers demanded progressively lower PPM targets — Toyota required under 5 PPM (less than 5 defective parts per million shipped), GM under 15 PPM, and Ford under 10 PPM. The manufacturer's current quality performance averaged 45 PPM across all customers — above every OEM's target and threatening Tier-1 supplier status. Two OEMs had issued quality warnings, and one had placed the manufacturer on conditional approval with 90 days to demonstrate improvement. (3) Inspection bottleneck: production line speeds had increased 25% over 3 years through equipment upgrades, but inspection capacity had not kept pace. Lines ran at 120 parts per minute for stamped components — human inspectors could not maintain consistent attention at this speed for 8-hour shifts. Quality data showed inspection escape rates increasing 40% in the final 2 hours of each shift — fatigue directly affected detection capability. (4) Scrap costs: when defects were detected, parts were scrapped at an average cost of CAD $4.20 per part (material and processing). However, late detection (after additional processing steps) increased scrap cost to CAD $18.50 per part. Early detection at the stamping press would catch defects before value-adding processes — but stamping inspection was the most difficult to automate due to speed, vibration, and varying lighting conditions. (5) Customer complaint cost: each OEM quality complaint required an 8D investigation report, containment actions (sorting existing inventory at the OEM plant and in-transit), and corrective action implementation. Average cost per complaint: CAD $85,000 (including sorting costs, OEM charges, and internal investigation). The manufacturer averaged 120 quality complaints per year — CAD $10.2M in annual complaint costs.

We developed and deployed a visual quality inspection system over 14 weeks. (1) In-line defect detection: high-speed camera systems installed on 35 production lines across 12 plants — 4-8 cameras per line capturing images of every part at production speed (120+ parts per minute). Lighting: custom LED illumination arrays designed for each part geometry — strobed lighting eliminating motion blur and creating consistent, repeatable illumination regardless of ambient conditions. Models: deep learning defect detection models trained on 4.2 million annotated images from the manufacturer's historical quality records. Defect types: surface defects (scratches, dents, tool marks, porosity, cracks — detected at sub-millimetre resolution), dimensional anomalies (warpage, incorrect trim, hole position deviation — detected through silhouette and feature measurement), and material defects (rust, contamination, coating defects — detected through colour and texture analysis). Real-time classification: each part classified as pass, fail, or suspect within 85ms — faster than the production cycle time, enabling 100% inspection without slowing production. (2) Stamping press inspection: purpose-built inspection stations at 15 stamping presses — the highest-speed, most challenging inspection environment. Die monitoring: continuous imaging of the stamping die surface between strokes — detecting die wear, buildup, and damage that caused progressive quality degradation. The system predicted die maintenance needs 200-500 strokes before defect onset — enabling proactive die change during scheduled breaks rather than reactive change after defect discovery. First-piece verification: automated first-piece inspection after die setup or changeover — comparing the first production parts against CAD specifications within 30 seconds (replacing 15-minute manual first-piece inspection). (3) Assembly verification: camera systems verifying assembled modules — component presence (correct parts installed), orientation (parts correctly positioned), and completeness (all fasteners, clips, and secondary operations completed). Poka-yoke integration: inspection system integrated with assembly station PLCs — stations locked and alerted operators when the vision system detected missing or incorrect components, preventing defective assemblies from advancing to the next operation. (4) Quality analytics: defect trend analysis — real-time monitoring of defect rates by type, part number, production line, shift, and time period. SPC integration: vision system measurements feeding automated SPC charts — detecting process drift and out-of-control conditions before they produced out-of-specification parts. Root cause correlation: defect patterns correlated with process parameters (press tonnage, cycle time, material lot, die age) — accelerating root cause investigation from days to hours. OEM reporting: automated generation of quality reports in each OEM's required format — PPAP documentation, quality performance dashboards, and 8D investigation data pre-populated from vision system records. (5) Edge architecture: all inspection processing performed at the edge (NVIDIA Jetson Orin modules at each inspection station) — no dependency on network connectivity or cloud for real-time inspection decisions. Central analytics: inspection results streamed to a central cloud platform (AWS ca-central-1) for cross-plant analytics, model updates, and OEM reporting. Model updates: new defect detection models trained on cloud and deployed to edge devices during maintenance windows — keeping detection capabilities current with new part introductions and evolving defect types.

Defect detection: escape rate from 15% to 2.8% — catching 97.2% of defects before shipment. PPM performance: from 45 PPM (above all OEM targets) to 8 PPM average across all customers — meeting Toyota's 5 PPM target (achieving 4.2 PPM), and well within GM and Ford targets. Both quality warnings rescinded within 6 months. Conditional approval restored to full approval. Scrap reduction: early detection (at stamping rather than after value-adding operations) reduced average scrap cost per defect from CAD $18.50 to CAD $4.20 — saving CAD $42M annually. Die monitoring: proactive die maintenance reduced die-related defects by 78% — and extended average die life by 15% through earlier intervention before catastrophic wear. Inspection efficiency: 280 human inspectors reduced to 85 (retained for final audit, complex assemblies, and vision system maintenance). Inspection consistency: zero detection degradation from shift fatigue — the vision system maintained identical accuracy at minute 1 and minute 480 of every shift. Customer complaints: from 120/year to 18/year — CAD $8.7M annual reduction in complaint costs. New business: quality improvement enabled 3 new OEM programmes that required sub-10 PPM capability as a qualification prerequisite — projected CAD $180M in additional annual revenue once in production. Annual value: CAD $42M (scrap reduction — early detection saving value-added processing costs), CAD $28M (operational efficiency — 195 FTE reduction in quality inspection at average CAD $58K fully loaded cost plus overhead), CAD $8.7M (complaint reduction — fewer OEM quality complaints and associated costs), CAD $6.5M (die life extension — proactive maintenance reducing die replacement and repair costs), CAD $4.8M (OEM relationship — maintained Tier-1 status and qualified for new programmes) = CAD $90M. Development cost: CAD $680,000. Payback: under 3 days.