Defect identification method

Abstract

A defect identification method comprises the following steps: establishing a defect database; building a defect target recognition training model and training; building a defect identification model and testing, extracting the clustering gray scale and morphological characteristics of the target detection range, and judging the defect type and the qualification standard of the target detection range. According to the invention, based on the defect or damage image, a deep learning method is adopted, a defect or damage database is constructed, a target identification model is established and generated, defects and damage are positioned, and the defect or damage identification model is improved in accuracy through hyper-parameter optimization.

Description

technical field [0001] The invention relates to the field of product defect identification, in particular to a defect identification method. Background technique [0002] In industrial production, the field of defect detection and identification includes visual inspection of surface and internal defects such as undercut, weld bead, collapse, porosity, slag inclusion, and incomplete penetration. Conventional non-destructive testing methods include radiographic testing, eddy current testing, penetrant testing, ultrasonic testing and magnetic particle testing. Industrial imaging detection technology can be applied to oil pipeline defects and welding defects. [0003] Although the current industrial imaging defect detection technology can obtain accurate defect images, the timeliness is poor, the film cannot be reused, and the cost is high. Digital real-time imaging has also been applied in industrial defect detection scenarios. It mainly includes manual and computer-aided di...

AI Technical Summary

Problems solved by technology

The test results of these methods are very dependent on the experience level of professional testers, and some defects have no clear criteria, which are greatly affected by the subjective consciousness of testers.

The actual manual identification process requires continuous observation of X-ray images for a long time, and high-intensity work will cause visual fatigue of inspectors, resulting in missed and false detections of defects

Although the computer-aided method of identification reduces the workload of personnel, the manual interactive operation method still requires a lot of manual intervention, and it is often aimed at specific types of defects, and the overall identification accuracy is still insufficient

Images