Research On Quantification Of Internal Defects Based On Pulse Thermal Imaging

Non-destructive testing is a technolohy detecting the internal structure,physical properties and state of the tested object without destroying the object.Non-destructive testing technology is of great significance in many aspects such as ensuring product quality and equipment safe service.Among all active infrared thermal imaging technologies,pulse thermal imaging technology is an active infrared thermal imaging detection technology based on transient heating.And pulse thermal imaging is widely used in the detection and quantification of internal defects in equipment and products due to its advantages of fast detection speed,low detection environment requirements,and easy quantitative calculation.At present,most of the related researches on the quantification of internal defects in pulsed thermal imaging only study different quantification methods based on the defect center,do not effectively use the temperature information of the remaining positions of the defect surface,and do not fully consider the results of defect detection on the quantification accuracy.Impact.Based on the defect detection and quantification technology of pulse thermal imaging,thesis proposes a thermal image sequence feature extraction method based on curve similarity.Quantification of defects.The main research contents of thesis are as follows:(1)Research on deep quantization methods.Based on the longitudinal heat transfer model of pulse thermal imaging,the mathematical relationship between the temperature response curve at the defect and the depth of the internal defect is quantitatively given,and the difference in depth quantification accuracy of different quantification methods is explained by the fuzzy effect of the heat conduction process.At the same time,the temperature information of each position on the defect surface is used as the input of the quantification method to improve the precision and stability of the depth quantification of internal defects.The comparison results based on the depth quantization method adopt a relatively optimal quantization method to quantify the internal defects.(2)Feature extraction and quantification method based on temperature response similarity.The feature extraction method based on the similarity of temperature response is used to process the thermal image sequence.This method can effectively eliminate the redundant information in the thermal image sequence,remove the noise,and enhance the deep defect information.Based on this method,the defects of the standard sample and the tested sample are segmented,and the average surface temperature of the internal defects is deeply quantified by using the segmented images.At the same time,the temperature response similarity value is based on the pulse thermal imaging mechanism,which is similar to the traditional quantitative calculation method,and can be used as a quantitative parameter to quantify the buried depth of internal defects.(3)Experimental verification.The different quantification methods are compared and demonstrated through simulation data and thermal imaging experimental data.Comparing the quantification accuracy of defect surface and defect center,the relative error of depth quantization can be reduced by 1.5%-2.0% in the simulation model;the relative error of depth quantization can be reduced by 3.0%-4.0% in the thermal imaging experiment.Comparing the similarity of temperature response with the traditional depth quantization method,the relative error of depth quantization is reduced by 1.0% in the simulation model,and the relative error is within 5%;in the thermal imaging experiment,the relative error of depth quantization is reduced by 1.4%,and the relative error is stable at about 9%.