A Bevel Gear Quality Inspection System Based on Multi-Camera Vision Technology.

Area of Science:

• Manufacturing Engineering
• Computer Vision
• Metrology

Background:

• Manual inspection of automotive bevel gears is labor-intensive, slow, and prone to inaccuracies.
• Existing automated methods lack the comprehensive defect detection and precise dimensional measurement capabilities required for modern production.

Purpose of the Study:

• To develop an automated quality inspection system for automotive bevel gears using multi-camera vision technology.
• To enable simultaneous detection of surface defects and measurement of gear dimensions.
• To enhance the efficiency, speed, and accuracy of bevel gear quality control.

Main Methods:

• Development of a synthetic bevel gear quality inspection system integrating multi-camera vision.
• Implementation of three novel algorithms: Neighborhood Average Difference (NAD), Circle Approximation Method (CAM), and Fast Rotation-Position (FRP).
• Systematic testing for various defect types (e.g., cracks, scratches, dents) and dimensional precision.

Main Results:

• The system successfully detects surface defects such as knock damage, cracks, scratches, and dents.
• The smallest detectable defect size is 0.4 mm × 0.4 mm.
• Dimensional measurement precision is achieved at approximately 40-50 μm.
• Inspection cycle time is under 1.3 seconds per unit.

Conclusions:

• The developed multi-camera vision system provides a highly precise and rapid solution for automotive bevel gear quality inspection.
• The system meets the stringent requirements for real-time online inspection in high-volume production environments.
• This automated approach offers significant improvements in cost-efficiency and accuracy compared to traditional manual inspection methods.