Research on aircraft skin rivet detection technology based on the normal vector-density clustering algorithm

Area of Science:

• Aerospace Engineering
• Materials Science
• Computational Geometry

Background:

• Riveting quality significantly impacts aircraft aerodynamic performance and fatigue life.
• Accurate analysis of rivet head point clouds is essential for quality assessment.
• Existing methods may lack precision in extracting and analyzing rivet head data.

Purpose of the Study:

• To propose and validate a novel rivet flushness detection method.
• To accurately extract rivet head point cloud data for quality analysis.
• To develop a system for intuitive visualization of rivet flushness detection outcomes.

Main Methods:

• Point cloud data sampling based on normal vectors.
• Density clustering algorithm for rivet head point cloud extraction.
• Random Sample Consensus (RANSAC) algorithm for rivet head contour fitting and parameter extraction.
• Introduction of a specific quality detection metric for rivet head flushness.

Main Results:

• The proposed method demonstrates high accuracy and minimal errors when compared to theoretical values.
• Successful application in analyzing rivet flushness on aircraft skin and theoretical models.
• Development of a visualization system for intuitive interpretation of results.

Conclusions:

• The developed method provides a robust and accurate approach for rivet flushness detection.
• The visualization system enhances the practical application and understanding of rivet quality analysis.
• This method holds significant engineering value for quality control in aircraft manufacturing and maintenance.