📋 Case Study

3D Vision-Guided Bin-Picking for Aerospace Fastener Kits

Highly reflective titanium fasteners (M3–M8), nested geometry, dense packing, and strict traceability requirements

🏗️ Project Overview

Automated kitting cell for Boeing 787 wing assembly line

🎯 Challenge

Highly reflective titanium fasteners (M3–M8), nested geometry, dense packing, and strict traceability requirements

🔧 Design Approach

Structured light scanner with multi-angle capture; point cloud denoising using bilateral filtering + RANSAC segmentation; gripper pose optimization with grasp quality metric (Q-value) ≥ 0.82

📐 Design Diagram

3D Vision-Guided Bin-Picking for Aerospace Fastener Kits Challenges: • Reflective Ti (M3–M8) • Nested geometry • Traceability required SL Structured Light (multi-angle) Point Cloud Denoising & Segmentation (Bilateral + RANSAC) ≥18.7 pts/mm² Grasp Optimization Q ≥ 0.82 P = 98.1% Bin with Titanium Fasteners Gripper Scanner Processing Grasp Challenge

AI-generated project design illustration

📐 Key Calculations

Point Cloud Density Requirement

≥ 12 pts/mm² for sub-mm grasp precision
Result: 18.7 pts/mm²
Enables accurate centroid and normal vector estimation

Grasp Success Probability

P = exp(−β × (1−Q))
Result: 98.1%
Validated against 12,000 pick attempts

📊 Results

99.2% first-attempt success rate; 100% serialization traceability via vision-verified QR code scan pre-grasp; 3.2× faster than manual kitting

💡 Lessons Learned

  • Multi-angle structured light eliminated shadowing on concave fastener heads
  • Bilateral filtering preserved sharp edges while removing laser speckle noise
  • Q-value thresholding prevented low-confidence grasps that caused jamming

Key Takeaways

  • 1Multi-angle structured light eliminated shadowing on concave fastener heads
  • 2Bilateral filtering preserved sharp edges while removing laser speckle noise
  • 3Q-value thresholding prevented low-confidence grasps that caused jamming