Live Production Validation ยท May 2026

AI-Powered Sheet Metal Inspection โ€” Field Results

Two real-world deployments. Two automotive clients. Zero compromise on accuracy.

100% JBM Accuracy
98.2% OEM Combined
1,073 Samples Tested
2 Automotive Clients

Case Study 01: JBM Auto โ€” Body-in-White Inspection

AI-Powered Metal Sheet Inspection System | Hole, Bolt & Perimeter Detection

๐Ÿ“… May 12, 2026
๐Ÿญ Live Conveyor Inline
๐Ÿ‘ค Manager: Mr. Balamohan
๐Ÿ“„ PO: EIDP 284/12.05.2026
๐Ÿ‘๏ธ

Manual Inspection Bottleneck

High-volume production required consistent hole and bolt counting โ€” impossible to maintain with human operators.

๐Ÿ“

Zero Defect Tolerance

Even one missing hole or bolt reaching customer becomes costly โ€” required 100% detection capability.

๐Ÿ“Š

No Data Traceability

Manual inspection left no audit trail โ€” impossible to track defects or perform root cause analysis.

โฑ๏ธ

Throughput Constraints

Production speed outpaced inspection capacity โ€” needed automated system for continuous operation.

๐Ÿ”ต Hole Count Detection

AI model detects all circular & oval holes and validates against configured thresholds per sheet grade. Count mismatch โ†’ instant NG flag.

๐Ÿ”ฉ Bolt Count Detection

Detects and counts bolt positions per sheet. Grade-wise spec validation. Mismatch triggers instant NG classification.

๐Ÿ“ Perimeter Defect Detection

Side camera monitors the complete sheet perimeter for edge defects, cracks, and deformations in real time.

100%
JBM Live Production Accuracy
Metric Target Achieved Status
Detection Accuracy โ‰ฅ99% 100% โœ“ Exceeded
Hole Detection โ‰ฅ95% 100% โœ“ Exceeded
Bolt Detection โ‰ฅ95% 100% โœ“ Exceeded
System Uptime โ‰ฅ98% 99.7% โœ“ Exceeded
System Commissioned
Functionally Compliant
Accepted for Validation
4/4 KPIs Met
Client Signed Off

Case Study 02: Automotive OEM โ€” Door Panel Inspection

Automated Metal Door Defect Detection System | LHS & RHS Door Panels

๐Ÿ“… May 2026
๐Ÿญ FG Conveyor ยท Live Production
๐Ÿ“ LHS + RHS Door SKUs
๐Ÿค– Multi-Camera AI Vision
๐Ÿ•ณ๏ธ

Hole Detection

Missing holes detected on LHS & RHS panels by multiple camera views to achieve 100% coverage.

ใ€ฐ๏ธ

Wrinkle Detection

Surface wrinkles identified via AI model across all camera views with consistent thresholds.

๐Ÿ“Ž

Folding Defects

Edge fold anomalies captured primarily by Side camera in live production environment.

โšก

Crack Detection

Surface and edge cracks detected in RHS panels with high confidence scoring and classification.

๐Ÿ“ท Machine Vision Camera

Model IV-GE501GC-T1-C
Type 5.0MP GigE
Sensor Sony IMX264
Resolution 2448 ร— 2048
Qty 4 Units

๐Ÿ”ญ Machine Vision Lenses

6mm Model IV-FA0615A
Focal Length 6mm ยท F1.8
6mm Qty 3 Units
16mm Model IV-FA1610A
16mm Qty 1 Unit

๐Ÿ’พ Processing Unit

AI Engine NVIDIA Jetson
Framework ONNX/TensorRT
Inference Time <100ms
Latency Sub-second
Integration REST API
Defect Category RHS Accuracy LHS Accuracy Combined
Hole Detection 98.5% 99.1% 98.8%
Wrinkle Detection 97.2% 96.8% 97.0%
Folding Defects 95.6% 96.4% 96.0%
Crack Detection 96.4% 94.8% 95.6%
Overall Accuracy 97.0% 96.8% 98.2%
98.2% Combined Accuracy
100% Good Parts Detected
โ‰ฅ95% on All Categories
1,000 Samples Validated
Production-Ready System

Ready to Deploy AI Inspection at Your Facility?

These systems are proven in production. Customizable for your specific sheet metal, door, or component specifications.

Get in Touch โ†’
Scroll to Top

Request a Demo