Non-destructive Testing Of Composite Materials Based On Infrared Thermography

Carbon fiber Reinforced Polymer(CFRP)has been widely used in aerospace,wind power generation and automobile manufacturing as a kind of superior performance material with light weight and high strength.However,in the case of cyclic stress and external impact,laminated composite structures may cause defects such as internal delamination or debonding,which are not easily detected from the surface.In order to ensure the internal quality and safety of the workpiece,Optical Pulsed Thermography(OPT)detection system is used to detect internal defects.At the same time,in order to eliminate the problem of uneven illumination and low-resolution caused by the system,common thermal feature extraction algorithms are adopted and compared.In order to improve the defect detection rate,an automatic seeded region growing method(ASRG)based on thermographic signal reconstruction(TSR)is proposed.This method can significantly improve the contrast between defects and sound regions and achieve accurate detection of defects.At the same time,this method can automatically select key parameters such as input images,seed points and thresholds.In order to objectively and quantitatively evaluate the detection performance of different algorithms,the paper used quantitative measures of F-score indicator to evaluate the processing results.The main content of this article is as follows:(1)The non-destructive testing technology and material properties of composite materials were studied.Based on the basic principles and theories of optical pulsed thermography technology,the excitation mode,response characteristics and excitation source are discussed and analyzed.Analyze the selection of key components in the selfconstructed OPT system.For the carbon fiber reinforced polymer specimen used in the experiment,a variety of defect-containing structures are designed to ensure the consistency of the test pieces and the actual workpieces.The temperature field distribution changes caused by the model will be analyzed.From the thermal image sequence acquired with the infrared camera,the temperature of surface change of the specimen containing the defect can be get.In order to enhance the contrast of the defect and reduce the noise interference,the defect feature extraction algorithm is performed,such as principal component analysis(PCA),independent component analysis(ICA),thermographic signal reconstruction(TSR)and pulsed phase thermography(PPT).Qualitative and quantitative analysis of the results and providing the objective basis for defect segmentation and quantification.(2)In order to realize automatic segmentation of defects,a hybrid of thermographic signal reconstruction(TSR)and automatic seeded region growing(ASRG)algorithm is proposed,which is conducive to the non-destructive testing and physical performance evaluation of composite materials.A frame with the largest kurtosis value is selected as the segmented image in the feature image after thermographic signal reconstruction.Then the seed point and threshold are determined according to the defect characteristic and the infrared heat map feature.Finally,the binary image is obtained and implemented.Based on the segmented image to achieve defect size quantification.The paper analyzes and compares the applicability of multiple types of thermal image feature extraction and segmentation methods in infrared non-destructive testing.Through quantitative analysis,the performance of the automatic segmentation method proposed in this paper is better.