A method and system for grading defects in X-ray weld images

Abstract

The invention discloses a method and system for grading defects in an X-ray weld image. The grading method includes: acquiring an image of an X-ray weld; determining the type and position of a defect in the X-ray weld image according to the X-ray weld image and size; determining the defect level of the X-ray weld image according to the type, position and size of the defect. The grading method in the present invention realizes the full automation of X-ray weld image defect detection and grading, avoids missed judgments and misjudgments caused by manual evaluation, and improves the grading accuracy.

Description

technical field [0001] The invention relates to the field of defect grading, in particular to a method and system for grading defects in X-ray weld images. Background technique [0002] In the field of welding defect detection, X-ray inspection is one of the important methods commonly used in conventional nondestructive testing. At present, in actual production, the main method is to manually analyze the weld image, and determine whether there are defects and the type, location, size, etc. of defects based on experience, so as to evaluate the welding quality and give corresponding ratings. The manual evaluation method is affected by human factors such as personal technical level, experience, fatigue, emotion and external conditions, which is inefficient, unreliable, and poor in consistency, and the strong light of the film reading is easy to burn people's eyes. Modern machinery manufacturing has higher and higher requirements for precision, and manual operation is increasin...

AI Technical Summary

Problems solved by technology

These two methods only use deep learning methods to classify and identify defect candidate areas. The generation of defect candidate areas still uses traditional image processing methods. This method of candidate area generation is sensitive to interference factors and has low robustness.

At the same time, these methods do not involve the final rating method

Images