Connector Bent Pins Defect Case Study
A global electronics manufacturer needed to automate inspections of multi-pin header connectors for bent pins and dust. Manual inspection was slow and led to a 2% miss rate. We built a segmenter recipe on the OV80i, pairing it with a telecentric lens to eliminate perspective distortion.
“Getting to this performance level & accuracy was impressive. The team had good lens knowledge & getting multiple classes of defects with logic was impressive.”
— Head of Quality
The Challenge
Manual inspection of the header connectors was slow, costly, and unreliable, leading to a consistent 2% miss rate for subtle pin bends. This was challenging due to: perspective distortion with standard lenses making edge pins appear at different angles, 3D variation where pins can be bent in any direction (X, Y, or Z axis), and multiple defects including dust particles that required individual manual cleaning.
Our Solution
We implemented a classifier algorithm on the OV80i using a telecentric lens with 0.3x magnification to reduce perspective error. With just 2 training images (1 good connector, 1 bad connector) and customized logic in Node-Red to calculate pin position, we achieved 100% accuracy across all test samples. The ability to pair AI with custom optics and logic made for a unique solution.
Hardware Setup
- • Camera: OV80i at 110mm working distance
- • Lens: 0.3x telecentric lens
- • Light: Standard ring light
Key Results
- • 100% accuracy on 25 test samples
- • Zero misses, zero overkills
- • ±5 pixel tolerance (~50 μm)
Technical Implementation Details
Our engineering team deployed a comprehensive solution including hardware setup, camera configuration, alignment procedures, and AI model training...
Performance Results
Detailed testing results, accuracy metrics, ROI calculations, and deployment timeline...
Want to Read the Full Case Study?
Access our complete library of case studies, technical details, and implementation guides.
Access Full Case StudyFree download with email registration