Method and system for nondestructive testing of aviation components

Through the active infrared thermal imaging system and the method of segmenting into predicted micro-areas, combined with local statistical models and incremental learning algorithms, the automation and efficiency problems of non-destructive testing of heterogeneous aviation components are solved, and efficient testing of aviation components with complex geometric shapes is achieved.

CN115280358BActive Publication Date: 2025-09-26SAFRAN SA
6 Cites 0 Cited by

Patent Information

Application Number
CN202180019665.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-03-13
Filing Date
2021-03-09
Publication Date
2025-09-26
Estimated Expiration
2041-03-09

AI Technical Summary

Technical Problem

Existing technologies for non-destructive testing of heterogeneous and complex geometrically shaped aviation components suffer from time-consuming and labor-intensive testing methods, difficulty in automated interpretation, and the fact that existing methods are only applicable to homogeneous materials and cannot effectively address the detection difficulties caused by heterogeneity and complex geometric shapes.

Method used

An active infrared thermal imaging system is used to acquire digital images of aviation components. The images are segmented into prediction micro-areas and compared with local statistical models using statistical prediction algorithms to form abnormal micro-maps. The local models are dynamically updated using incremental learning algorithms to achieve automated detection.

Benefits of technology

It improves the automation and accuracy of detection, reduces detection time, can effectively handle heterogeneous materials and complex geometries, and enhances the robustness and efficiency of the method.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

A method for performing non-destructive testing on an aviation component comprises the steps of obtaining a plurality of digital images of a unit area of ​​the aviation component, estimating a characteristic image (IMC) representing the unit area, each pixel of the characteristic image (IMC) comprising a characteristic vector (VC), segmenting the characteristic image (IMC) into a plurality of predicted micro-zones (MZPs), and comparing the characteristic vector (VC) of each pixel of each predicted micro-zone (MZP) with a pre-estimated local statistical model (MZP(ZU)-MOD) of the predicted micro-zone (MZP), wherein the local model (MZP(ZU)-MOD) of the predicted micro-zone (MZP)) is obtained by a learning algorithm from the characteristic vectors of the pixels of a learning micro-zone of an annotated characteristic image, wherein the learning micro-zone comprises the predicted micro-zone.
Need to check novelty before this filing date? Find Prior Art