Automatic Defects Recognition of Lap Joint of Unequal Thickness Based on X-Ray Image Processing

Area of Science:

• Materials Science
• Non-Destructive Testing
• Image Processing

Background:

• Automatic defect recognition in X-ray radiographs of lap joints with unequal thickness is challenging due to grayscale variations and non-vertical weld seams.
• Variations in workpiece wall thickness create inconsistent gray levels, complicating image analysis.
• Workpiece shape and fixturing lead to angled weld seams, hindering automated defect localization.

Purpose of the Study:

• To develop an effective method for automatic defect detection in X-ray images of lap joints with unequal thickness.
• To address challenges posed by grayscale variations and non-vertical weld seams in defect recognition.
• To enhance the accuracy and reliability of automated inspection systems for welded components.

Main Methods:

• Implemented an X-ray image correction technique utilizing invariant moments.
• Introduced a novel background removal method based on image processing to mitigate grayscale variations.
• Employed a combination of image noise suppression, image segmentation, and mathematical morphology for automatic defect detection.

Main Results:

• The proposed method successfully identified gas pores in automatically welded lap joints of unequal thickness.
• The image correction and background removal techniques effectively reduced difficulties associated with grayscale variations.
• The integrated defect detection approach demonstrated suitability for automated inspection.

Conclusions:

• The developed method provides an effective solution for automatic defect recognition in challenging X-ray inspection scenarios.
• This approach is suitable for automated detection of gas pores in lap joints with varying material thickness.
• The study contributes to advancing automated non-destructive testing techniques in manufacturing.