Infrared thermal wave nondestructive testing system and method

Abstract

The present invention discloses an infrared thermal wave nondestructive testing system, comprising a testing table, a thermal excitation module, a thermal imaging device, a component to be tested, a computer, a time relay and a synchronous trigger; the thermal excitation module, the thermal imaging device and the component to be tested are respectively arranged on the testing table; the component to be tested is located on one side of the testing table, and the thermal excitation module and the thermal imaging device are located on the other side of the testing table; the thermal imaging device is facing the component to be tested; the excitation end of the thermal excitation module is aligned with the component to be tested; the computer is connected to the thermal excitation module via a time relay. The present invention also discloses an infrared thermal wave nondestructive testing method. The beneficial effects of the present invention are: through continuous excitation of the thermal excitation module, deeper defects can be effectively detected; and through the background subtraction-principal component analysis algorithm, the noise signal caused by the uneven surface of the component to be tested can be effectively eliminated, thereby improving the accuracy of quantitative detection of defect area.

Description

Technical Field

[0001] The present invention relates to the technical field of defect detection, and in particular to an infrared thermal wave nondestructive detection system and method. Background Art

[0002] Infrared thermal wave nondestructive testing (NDT) involves actively applying a controllable thermal excitation source to the component being tested, causing internal defects or damage to manifest as surface temperature differences. Thermal imaging equipment is then used to continuously record surface temperature changes, enabling nondestructive testing of internal defects. In recent years, with increasing demands for the safety and cost-effectiveness of in-service equipment, simply detecting the presence and nature of defects is no longer sufficient to meet industrial needs. NDT is shifting towards nondestructive evaluation, with quantitative detection of defect location, size, and other information becoming a research hotspot worldwide.

[0003] At present, due to the existence of interference factors such as uneven thermal excitation and environmental scattering, the infrared image information obtained by conventional infrared thermal wave non-destructive testing is easily obscured by interference information, affecting the accuracy of subsequent quantitative detection; and the conventional thermal excitation source has a single function, a limited thermal excitation range, and relatively divergent energy, which cannot adapt to the detection of different types of components and has poor detection accuracy. Summary of the Invention

[0004] The purpose of the present invention is to provide an infrared thermal wave nondestructive testing system and method to improve the detection accuracy in order to address the deficiencies of the existing technology.

[0005] The technical solution adopted by the present invention is: an infrared thermal wave nondestructive testing system, including a testing table, a thermal excitation module, a thermal imaging device, a component to be tested, a computer, a time relay and a synchronous trigger;

[0006] The detection table is arranged horizontally;

[0007] The thermal excitation module, thermal imaging equipment and the component to be tested are respectively arranged on the testing table;

[0008] The component to be tested is located on one side of the testing table, and the thermal excitation module and the thermal imaging device are located on the other side of the testing table; the thermal imaging device faces the component to be tested; and the excitation end of the thermal excitation module is aligned with the component to be tested;

[0009] The computer is connected to the input end of the time relay, and the output end of the time relay is connected to the thermal excitation module;

[0010] The computer is connected to the input end of the synchronization trigger, the output end of the synchronization trigger is connected to the input end of the thermal imaging device, and the output end of the thermal imaging device is connected to the computer.

[0011] According to the above solution, the thermal actuation module includes an inner lampshade, an outer lampshade, a wick, a front convex lens and a rear convex lens;

[0012] One end of the inner lampshade is a closed end, a wick is arranged inside the closed end, and a front convex lens is installed inside the inner lampshade; the other end of the inner lampshade is an open end, connected to the open end of the outer lampshade, and the other end of the outer lampshade is closed; a rear convex lens is installed inside the closed end of the outer lampshade; the centers of the wick, the focus of the front convex lens and the focus of the rear convex lens are located on the same axis.

[0013] According to the above solution, the open end of the inner lampshade and the open end of the outer lampshade are threadedly connected.

[0014] According to the above solution, a plurality of lifting rods are provided on the detection table, and the thermal excitation module, thermal imaging equipment and the component to be tested are respectively installed on the upper ends of the lifting rods; and the computer is connected to the lifting rods.

[0015] The present invention also adopts an infrared heat wave nondestructive testing method, which includes the following steps:

[0016] S1. Provide the infrared thermal wave nondestructive testing system as described above and install the component to be tested;

[0017] S2. Adjust the thermal excitation module so that the light spot covers the detection area;

[0018] S3. The computer controls and triggers the thermal excitation module to heat the surface of the component under test, and the thermal imaging device records an infrared image sequence of the component under test and transmits the infrared image sequence to the computer;

[0019] S4. The computer converts the obtained infrared image sequence into a three-dimensional array that can be decomposed into a time-varying signal and a non-time-varying signal; and performs non-time-varying signal elimination processing on the three-dimensional array to obtain a differential image sequence;

[0020] S5, converting the differential image sequence into a two-dimensional matrix, performing principal component analysis on the two-dimensional matrix, reconstructing the infrared image sequence, and selecting the best principal component;

[0021] S6. Perform edge segmentation on the infrared image corresponding to the optimal principal component obtained in step S5 to calculate the number of pixels in the defective area;

[0022] S7. Calculate the area of ​​each pixel and find the defect area based on the number of segmented pixels.

[0023] According to the above scheme, the specific method of converting the infrared image sequence into a three-dimensional array that can be decomposed into time-varying signals and non-time-varying signals in S4 is:

[0024] Each frame of infrared image is regarded as a two-dimensional matrix, the infrared image sequence is regarded as a three-dimensional array, and the infrared image sequence is decomposed into time-varying signals and non-time-varying signals.

[0025] According to the above scheme, the specific method of performing time-invariant signal subtraction processing on the three-dimensional array in S4 to obtain a differential image sequence is as follows:

[0026] The pixel intensity values ​​at corresponding positions in the infrared image sequence are averaged to generate an average image as a background model, and the time-invariant signal is extracted. The time-invariant signal is then eliminated by subtracting the background model from the original three-dimensional array to obtain a differential image sequence.

[0027] According to the above scheme, the S5 method is as follows: each frame of the differential image sequence is flattened row by row into a one-dimensional row vector to form a two-dimensional matrix, the eigenvalues ​​and eigenvectors of the covariance matrix are calculated, the eigenvectors corresponding to the first 4 to 6 largest eigenvalues ​​are selected, the infrared image sequence is reconstructed, and the principal component with the smallest background noise and the largest defect signal is selected as the optimal principal component for segmentation.

[0028] According to the above solution, the specific method for segmenting and calculating the number of pixels in the defective area in S6 is:

[0029] Select the pixel point corresponding to the center of the image defect of the best principal component as the initial segmentation area, and for each pixel on the boundary of the current segmentation area, , calculate the pixel and its 8-neighborhood unsegmented pixels If the attraction exceeds the set attraction threshold, pixel j is added to the segmented area. Repeat the operation until all pixels that meet the conditions are included, and the number of pixels in the defect area is obtained.

[0030] According to the above scheme, the specific method of S7 is: by using the laser ranging function of the infrared thermal imager, the actual size of the measured component and the corresponding number of pixels in the infrared image are compared to obtain the size of each pixel in the infrared image, and the measurement area of ​​each defect is calculated respectively.

[0031] The beneficial effects of the present invention are:

[0032] 1. The system of the present invention can avoid the problem of defect signals being masked by interference factors such as uneven thermal excitation and environmental scattering by designing a thermal excitation module, thereby improving detection accuracy.

[0033] 2. The present invention uses a convex lens group to collimate and change the beam size of the excitation source to adapt to the detection of different types of components, and adjusts the excitation duration of the thermal excitation module to adapt to different types of defect detection, greatly improving the detection accuracy and efficiency.

[0034] 3. The detection method of the present invention can effectively detect deeper defects through continuous excitation of the thermal excitation module; and through the background subtraction-principal component analysis algorithm, it can effectively eliminate the non-time-varying noise and time-varying noise caused by the uneven surface of the tested component (such as fiber fabric texture, surface pits and surface coating, etc.), effectively extract the defect signal, and improve the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a structural diagram of Example 1 of the present invention.

[0036] Figure 2 This is a control diagram of Example 1.

[0037] Figure 3 Schematic diagram of the internal structure of the thermal actuation module in Example 1.

[0038] Figure 4 Schematic diagram of defect arrangement of the tested component in Example 2 Figure 1 .

[0039] Figure 5 Schematic diagram of defect arrangement of the tested component in Example 2 Figure 2 .

[0040] Figure 6 Schematic diagram of the original infrared image sequence in Example 2.

[0041] Figure 7 Schematic diagram of the differential image sequence in Example 2.

[0042] Figure 8 Schematic diagram of the infrared image sequence reconstructed in Example 2.

[0043] Figure 9 This is a schematic diagram of the defect segmentation result in Example 2.

[0044] Among them: 1 - detection table, 2 - thermal excitation module, 3 - thermal imaging equipment, 4 - lifting rod, 5 - fixture, 6 - component under test, 6.1 - large tow carbon fiber plate, 6.2 - polytetrafluoroethylene film, 7 - wick, 8 - inner lampshade, 9 - front convex lens, 10 - outer lampshade, 11 - rear convex lens, 12 - synchronization trigger, 13 - computer, 14 - time relay. DETAILED DESCRIPTION

[0045] In order to better understand the present invention, the present invention is further described below with reference to the accompanying drawings and specific embodiments.

[0046] like Figure 1 and Figure 2An infrared thermal wave nondestructive testing system shown includes a testing table 1, a thermal excitation module 2, a thermal imaging device 3, a component to be tested 6, a computer 13, a time relay 14 and a synchronous trigger 12;

[0047] The detection table 1 is arranged horizontally;

[0048] The thermal excitation module 2, thermal imaging device 3 and tested component 6 are respectively arranged on the testing table 1;

[0049] The component under test 6 is located on one side of the detection table 1, and the thermal excitation module 2 and the thermal imaging device 3 are located on the other side of the detection table 1; the thermal imaging device 3 faces the component under test 6; the excitation end of the thermal excitation module 2 is aligned with the component under test 6;

[0050] The computer 13 is connected to the input end of the time relay 14, and the output end of the time relay 14 is connected to the two thermal excitation modules 2;

[0051] The computer 13 is connected to the input end of the synchronization trigger 12 , the output end of the synchronization trigger 12 is connected to the input end of the thermal imaging device 3 , and the output end of the thermal imaging device 3 is connected to the computer 13 .

[0052] Preferably, the test surface 1 is equipped with multiple lifting rods 4 that can be rotated and raised. The thermal actuation module 2, thermal imaging device 3, and component under test 6 are each mounted on the upper ends of the lifting rods 4, which are fixed to the test surface 1. The computer 13 is connected to the lifting rods 4 to control their elevation and rotation. There are four lifting rods 4, and the height adjustment range of the lifting rods 4 is 0.3 to 1 meter. The component under test 6 is mounted on the upper ends of the lifting rods 4 via a U-shaped fixture 5.

[0053] In the present invention, a thermal excitation module 2 is used to apply thermal excitation to the surface of the component under test 6. When the heat flow encounters a defective area, the heat diffusion is blocked, resulting in local accumulation of heat, forming a temperature gradient on the surface of the component under test 6. The temperature distribution difference on the surface of the component under test 6 is then captured by the thermal imaging device 3, and the thermal image signal is sent to the computer 13; the frame rate of the thermal imaging device 3 is 30Hz.

[0054] In the present invention, the computer 13 controls the thermal actuation module 2 via the time relay 14 to operate for a duration ranging from 0 to 999 seconds, which is 6 seconds in this embodiment.

[0055] Preferably, if Figure 3 As shown, the thermal actuation module 2 includes an inner lampshade 8, an outer lampshade 10, a wick 7, a front convex lens 9 and a rear convex lens 11;

[0056] One end of the inner lampshade 8 is a closed end, a wick 7 is arranged inside the closed end, and a front convex lens 9 is installed in the inner lampshade 8; the other end of the inner lampshade 8 is an open end, which is connected to the open end of the outer lampshade 10 (the open end of the inner lampshade 8 and the open end of the outer lampshade 10 are threadedly connected), and a rear convex lens 11 is installed at the other end of the outer lampshade 10; the wick 7 generates a diffuse light beam, which is refracted and focused to a point by the front convex lens 9, and the outer lampshade 10 is moved to control the position of the rear convex lens 11, so that the light beam is refracted by the rear convex lens 11 to generate a light spot with controllable size and uniform brightness on the surface of the measured component 6.

[0057] In the present invention, by controlling the distance between the light spot and the rear convex lens 11, the object distance can be adjusted, thereby controlling whether the light beam is focused, collimated, or diverged after passing through the rear convex lens 11, and further controlling the size of the light spot so that the light spot can cover the detection area (the component to be detected 6 is located within the detection area); when the light beam is in a collimated state, the brightness distribution is uniform.

[0058] In the present invention, the centers of the wick 7, the focus of the front convex lens 9 and the focus of the rear convex lens 11 are located on the same axis.

[0059] In the present invention, the rated power of the lamp core 7 is 1000W; the focal length of the front convex lens 9 is 10 cm, and the focal length of the rear convex lens 11 is 20 cm; the inner lampshade 8 and the outer lampshade 10 are threadedly connected.

[0060] In the present invention, the axial distance between the wick 7 and the front convex lens 9 is greater than the focal length of the front convex lens 9; by adjusting the distance between the inner lampshade 8 and the outer lampshade 10, the axial distance between the rear convex lens 11 and the front convex lens 9 can be changed, so that the axial distance between the two is greater than the image distance of the front convex lens 9.

[0061] In the present invention, the wick 7 is 20 cm away from the front convex lens 9. According to the convex lens imaging formula: 1 / f1=1 / v1+1 / u1 (f1 is the focal length of the front convex lens 9, u1 is the object distance, and v1 is the image distance), it can be calculated that the emitted light beam is refracted by the front convex lens 9 and converges 20 cm to the right. Similarly, when the axial distance between the rear convex lens 11 and the light beam convergence point is equal to the focal length, the image distance is infinite, and the light beam is collimated into a parallel light beam by the rear convex lens 11; when the axial distance between the rear convex lens 11 and the light beam convergence point is greater than the focal length, the light beam is focused by the rear convex lens 11; when the axial distance between the rear convex lens 11 and the light beam convergence point is less than the focal length, the light beam is diverged by the rear convex lens 11. Therefore, the range and intensity of the excitation light beam can be controlled by adjusting the position of the rear convex lens 11. In this embodiment, the axial distance between the rear convex lens 11 and the light beam convergence point is equal to the focal length of the rear convex lens 11, 20 cm, so that the light beam is collimated into a parallel light beam (this is the aforementioned collimated state), as shown in FIG. Figure 3 shown.

[0062] The present invention also designs an infrared heat wave nondestructive testing method for detecting the defect area of ​​the tested component 6, which includes the following steps:

[0063] S1. Using the infrared thermal wave nondestructive testing system as described above, and installing the component to be tested 6;

[0064] S2, adjust the thermal excitation module 2 so that the light spot covers the detection area;

[0065] S3, the computer 13 controls the triggering of the thermal excitation module 2 to heat the surface of the component under test 6, and the thermal imaging device 3 records a sequence of n infrared images of the component under test 6 (n is not less than 500) and transmits the sequence of infrared images to the computer 13;

[0066] S4. The computer 13 converts the obtained n-frame infrared sequence image into a three-dimensional array that can be decomposed into a time-varying signal and a non-time-varying signal; and performs a non-time-varying signal elimination process on the three-dimensional array to obtain a differential image sequence;

[0067] S5, converting the differential image sequence into a two-dimensional matrix, performing principal component analysis on the two-dimensional matrix, reconstructing the infrared image sequence, and selecting the best principal component;

[0068] S6. Perform edge segmentation on the infrared image corresponding to the optimal principal component obtained in step S5 to calculate the number of pixels in the defective area;

[0069] S7. Calculate the area of ​​each pixel and find the defect area based on the number of pixels segmented in step S6.

[0070] In S1, the appropriate number of thermal actuation modules 2 is selected according to the type of the component to be tested 6, and each component is connected to the detection table 1 through the lifting rod 4. The component to be tested 6 is clamped on the fixture 5 so that the thermal actuation module 2, the thermal imaging device 3 and the component to be tested 6 are at the same horizontal height. The computer 13, the synchronization trigger 12, the time relay 14, the thermal imaging device 3 and the thermal actuation module 2 are connected by wire.

[0071] In S2, the connection position of the outer lampshade 10 and the inner lampshade 8 is adjusted, and the distance between the front convex lens 9 and the rear convex lens 11 is adjusted to obtain a light spot of appropriate size; S3 sets an appropriate thermal excitation duration through the computer 13.

[0072] In S4, the specific method of converting the infrared image sequence into a three-dimensional array that can be decomposed into time-varying signals and non-time-varying signals is as follows:

[0073] The vertical direction of the infrared image is the y-axis, and the horizontal direction is the x-axis. Each frame of the infrared image has x0 pixels in the x direction and y0 pixels in the y direction. Each pixel has a temperature data, so each frame of the infrared image can be regarded as a two-dimensional matrix of x0×y0; the infrared image sequence of n frames with a dimension of x0×y0 can be regarded as a three-dimensional array ,Since the time-invariant signal changes little or basically remains stable in each frame, the ,infrared image sequence can be decomposed into time-varying and time-invariant signals, as shown in ,Equation (1):

[0074] (1);

[0075] ;

[0076] ;

[0077] ;

[0078] Where: For location The pixel at The intensity of the frame, is a time-invariant signal, is a time-varying signal; in the formula, M is the total number of frame sequences; A is x0, and B is y0.

[0079] In S4, the three-dimensional array is processed to subtract the time-invariant signal to obtain a differential image sequence. The specific method is as follows:

[0080] Average the pixel intensity values ​​at corresponding positions in the n-frame infrared image sequence to generate an average image As a background model, separate the time-invariant signal , as shown in formula (2); then the original n-frame infrared image sequence is eliminated by subtracting the background model to obtain the differential image sequence , as shown in formula (3) and Figure 7 As shown:

[0081]

[0082] (2);

[0083]

[0084] (3);

[0085] Where: is the total number of frame sequences; It is the three-dimensional array corresponding to the difference image sequence.

[0086] In S5, the differential image sequence is converted into a two-dimensional matrix, and principal component analysis is performed on the two-dimensional matrix. The eigenvectors corresponding to the first 4 to 6 largest eigenvalues ​​are selected (usually the first 10 are calculated, and then the number of selected is determined based on the effect of observing the specific principal component image, generally the first 4 to 6 are selected; existing methods can be used for processing). The infrared image sequence is reconstructed, and the principal component with the strongest defect signal and the weakest noise signal is retained. The specific method is as follows:

[0087] Each frame of the above differential image sequence is flattened row by row into a one-dimensional row vector containing x0×y0 pixels, and the n row vectors are organized into a two-dimensional matrix with n rows (x0×y0) columns. The eigenvalues ​​and eigenvectors of the covariance matrix are calculated. The larger eigenvalues ​​represent the main direction of the data, and the eigenvectors represent the direction of data change. Therefore, the eigenvectors corresponding to the first 4 to 6 largest eigenvalues ​​are selected to reconstruct the infrared image sequence, that is, the infrared images corresponding to each principal component. The principal component with the smallest background noise and the largest defect signal is selected as the best principal component for segmentation. The best principal component can be selected by comparing infrared images according to different research objects. In addition, the common means in the industry, such as selecting the eigenvectors corresponding to the 5 largest eigenvalues ​​in Example 1, that is, the infrared images reconstructed from the first to fifth principal components, are as follows. Figure 8 As shown in the figure; after comparative analysis of the pictures, the first principal component mainly contains the light intensity distribution information, the second principal component mainly contains the background noise, the defect signals in the fourth and fifth principal components are not obvious, and the background noise of the third principal component is the smallest, and the defect signals are the largest and most obvious. Therefore, the third principal component is selected as the optimal principal component for subsequent quantitative segmentation.

[0088] The specific method for segmenting and calculating the number of pixels in the defective area in S6 is:

[0089] Select the pixel point corresponding to the center of the infrared image defect of the best principal component as the initial segmentation area, and for each pixel on the boundary of the current segmentation area, , calculate the pixel and its 8-neighborhood unsegmented pixels Gravity ,gravitational Calculate by formulas such as (4) to (6). If gravity If the value exceeds the set gravity threshold T (the threshold is the value when the gravity F(i,j) gradient is the maximum, obtained by multiple segmentation calculations, which is the existing technology), then pixel j is added to the segmentation area, and the above operation is repeated until all pixels that meet the conditions are included, and the number of pixels in the defect area is obtained. The segmentation result is as follows Figure 9 shown.

[0090] (4);

[0091] (5);

[0092] (6);

[0093] Where, is the fuzzy similarity function, Represents pixels and The gray value difference between is the fuzzy constant, which is used to control the sensitivity of the similarity function to grayscale difference and is set to 0.18; is the distance function, ( x i ，y i )、( x j ，y j ) are pixels and The location coordinates of Pixels and The attraction between is the gravitational constant, set to 0.2, which is used to adjust the gravitational strength.

[0094] The specific method for calculating the area of ​​each pixel in S7 and determining the defect area based on the above-mentioned number of pixels is as follows: by using the laser ranging function of the infrared thermal imager, the actual size of the component under test 6 is compared with the number of pixels corresponding to it in the infrared image, the size of each pixel in the infrared image can be obtained, and the measurement area and detection error rate of each defect can be calculated respectively.

[0095] In the present invention, the detection error rate = |defect measurement area - defect actual area| / actual area. The detection error rate is used to verify the segmentation effect of this method. The lower the error rate, the more accurate the detection result.

[0096] Example 1

[0097] like Figures 1-3 The infrared thermal wave nondestructive testing system shown in the figure includes two sets of thermal actuation modules 2, each connected to the output end of a time relay 14. The two sets of thermal actuation modules 2 are symmetrically arranged on either side of a thermal imaging device 3, with the actuation ends of the two thermal actuation modules 2 aligned with the component under test 6. The relevant parameters of the infrared thermal wave nondestructive testing system are designed as follows: the thermal imaging device 3 has a frame rate of 30Hz and a resolution of 640×480; the height adjustment range of the lifting rod 4 is 0.3 to 1m; the rated power of the wick 7 is 1000W; the focal length of the front convex lens 9 is 10cm, and the focal length of the rear convex lens 11 is 20cm.

[0098] Example 2

[0099] In this embodiment, the tested component 6 is a square carbon fiber fabric composite laminate with a side length of 250 mm and a thickness of 4 mm, which is composed of multiple layers of large tow carbon fiber plates 6.4 (a total of 14 layers, Figure 5 Only three layers are shown), such as Figure 4 As shown. A polytetrafluoroethylene film 6.2 is inserted between the layers of the component to be tested 6 to preset the delamination defect, as shown. Figure 5 As shown, there are six defects from 1 to 6, each of which is circular with a diameter of Figure 5 As shown in the figure, the preset depth of defects No. 1, 2, and 3 is 1.14 mm (the distance of the defect on the heated surface of the test piece), and the preset depth of defects No. 4, 5, and 6 is 2.28 mm.

[0100] The method for detecting defects of the component 6 in this embodiment using the device described in Example 1 is as follows:

[0101] The device described in Example 1 is used and installed, and the position of the thermal excitation module 2 is adjusted so that a uniform light spot with a diameter of 300 mm and a size sufficient to cover the detection area can be obtained on the component under test 6;

[0102] The computer 13 controls the triggering of the thermal excitation module 2 to heat the surface of the component 6 under test for 6 seconds, and controls the thermal imaging device 3 to record from 1 second before heating and continuously record for 40 seconds, obtaining a total of 1200 frames of infrared image sequence. The results are as follows: Figure 6 As shown; the vertical direction of the image is the y-axis, the horizontal direction is the x-axis, and each infrared image The axis has 640 pixels, The axis has 480 pixels.

[0103] Convert the infrared image sequence into a three-dimensional array that can be decomposed into time-varying signals and non-time-varying signals. The specific method is:

[0104] The 1200 frame dimensions are The infrared image sequence can be regarded as a three-dimensional array ,Since the time-invariant signal changes little or basically remains stable in each frame, the ,infrared image sequence can be decomposed into time-varying and time-invariant signals, as shown in ,Equation (1).

[0105] (1);

[0106] ;

[0107] ;

[0108] ;

[0109] Where: For location The pixel at The intensity of the frame, is a time-invariant signal, is a time-varying signal; in the formula, M is the total number of frame sequences, which is equal to 1200; A is 640, and B is 480.

[0110] The three-dimensional array is processed to subtract the time-invariant signal to obtain a differential image sequence. The specific method is as follows:

[0111] Average the pixel intensity values ​​at corresponding positions in the 1200-frame infrared image sequence to generate an average image As a background model, separate the time-invariant signal , as shown in formula (2); then the 1200-frame infrared image sequence is eliminated by subtracting the background model to obtain the differential image sequence , as shown in formula (3) and Figure 7 shown.

[0112]

[0113] (2);

[0114]

[0115] (3);

[0116] Where: is the total number of frame sequences, equal to 1200; It is the three-dimensional array corresponding to the difference image sequence.

[0117] Flatten each frame of the above differential image sequence row by row into a one-dimensional row vector containing 307,200 pixels, and form 1,200 row vectors into a two-dimensional matrix with 1,200 rows and 307,200 columns. Calculate the eigenvalues ​​and eigenvectors of the covariance matrix. The larger eigenvalues ​​represent the main direction of the data, and the eigenvectors represent the direction of data change. Select the eigenvectors corresponding to the first five largest eigenvalues. The first to fifth principal components are the infrared images reconstructed by the first five eigenvectors, such as Figure 8 As shown in the figure, after comparative image analysis, the first principal component mainly contains light intensity distribution information, the second principal component mainly contains background noise, and the 2.28 mm deep defect signals in the fourth and fifth principal components are not obvious. The background noise in the third principal component is the smallest and the defect signals are the most obvious. Therefore, the third principal component is selected as the optimal principal component for subsequent quantitative segmentation.

[0118] Perform edge segmentation on the best principal component obtained above to calculate the number of pixels in the defect area. The specific method is:

[0119] Select the six defect center pixels in the third principal component as the initial segmentation area, and for each pixel at the boundary of the current segmentation area , calculate the pixel and its 8-neighborhood unsegmented pixels Gravity ,gravitational Calculate using formulas such as (4) to (6). If If the gravitational threshold T is exceeded, pixel j is added to the segmented area and the above operation is repeated until all pixels that meet the conditions are included. The number of pixels in the defective area is obtained and the segmentation result is as follows: Figure 9 shown.

[0120] (4);

[0121] (5);

[0122] (6);

[0123] Where, is the fuzzy similarity function, Represents pixels and The gray value difference between is a fuzzy constant used to control the sensitivity of the similarity function to grayscale difference, which is set to 0.18; is the distance function, ( x i ， y i )、( x j ，y j ) are pixels and The location coordinates of It's a pixel and The attraction between is a gravitational constant, set to 0.2, which is used to adjust the strength of gravity.

[0124] The area of ​​each pixel was calculated, and the defect area was determined based on the number of pixels. The specific method was: using the infrared thermal imager's laser ranging function, comparing the actual size of the specimen with the number of pixels corresponding to it in the infrared image, it was determined that each pixel in the infrared image was 0.39 mm. The measured area and detection error rate of each defect were calculated, and the maximum detection error rate was 13.34% and the minimum was 7.44%.

[0125] The contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0126] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent replacements for some of the technical features therein. However, any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An infrared thermal wave nondestructive testing system, characterized in that: It includes a testing table, a thermal excitation module, a thermal imaging device, a component to be tested, a computer, a time relay and a synchronous trigger; The detection table is arranged horizontally; The thermal excitation module, thermal imaging equipment and the component to be tested are respectively arranged on the testing table; The component to be tested is located on one side of the testing table, and the thermal excitation module and the thermal imaging device are located on the other side of the testing table; the thermal imaging device faces the component to be tested; and the excitation end of the thermal excitation module is aligned with the component to be tested; The computer is connected to the input end of the time relay, and the output end of the time relay is connected to the thermal excitation module; The computer is connected to the input end of the synchronization trigger, the output end of the synchronization trigger is connected to the input end of the thermal imaging device, and the output end of the thermal imaging device is connected to the computer; The thermal actuation module includes an inner lampshade, an outer lampshade, a wick, a front convex lens and a rear convex lens; One end of the inner lampshade is a closed end, a wick is arranged inside the closed end, and a front convex lens is installed inside the inner lampshade; the other end of the inner lampshade is an open end, connected to the open end of the outer lampshade, and the other end of the outer lampshade is closed; a rear convex lens is installed inside the closed end of the outer lampshade; the centers of the wick, the focal point of the front convex lens, and the focal point of the rear convex lens are located on the same axis; The open end of the inner lampshade is threadedly connected to the open end of the outer lampshade.

2. The infrared thermal wave nondestructive testing system according to claim 1, characterized in that: A plurality of lifting rods are arranged on the detection table, and the thermal excitation module, the thermal imaging device and the component to be detected are respectively installed on the upper ends of the lifting rods; and the computer is connected to the lifting rods.

3. An infrared thermal wave nondestructive testing method, characterized in that: The method comprises the following steps: S1. Using the infrared thermal wave nondestructive testing system according to claim 2, and installing the component to be tested; S2. Adjust the thermal excitation module so that the light spot covers the detection area; S3. The computer controls and triggers the thermal excitation module to heat the surface of the component under test, and the thermal imaging device records an infrared image sequence of the component under test and transmits the infrared image sequence to the computer; S4. The computer converts the obtained infrared sequence image into a three-dimensional array decomposed into a time-varying signal and a non-time-varying signal; and performs non-time-varying signal elimination processing on the three-dimensional array to obtain a differential image sequence; S5, converting the differential image sequence into a two-dimensional matrix, performing principal component analysis on the two-dimensional matrix, reconstructing the infrared image sequence, and selecting the best principal component; S6. Perform edge segmentation on the infrared image corresponding to the optimal principal component obtained in step S5 to calculate the number of pixels in the defective area; S7. Calculate the area of ​​each pixel and find the defect area based on the number of segmented pixels.

4. The infrared thermal wave nondestructive testing method according to claim 3, characterized in that: The specific method for converting infrared image sequences in S4 is: Each frame of infrared image is regarded as a two-dimensional matrix, the infrared image sequence is regarded as a three-dimensional array, and the infrared image sequence is decomposed into time-varying signals and non-time-varying signals.

5. The infrared thermal wave nondestructive testing method according to claim 4, characterized in that: The specific method for performing time-invariant signal subtraction on a three-dimensional array in S4 to obtain a differential image sequence is as follows: The pixel intensity values ​​at corresponding positions in an n-frame infrared image sequence are averaged to generate an average image as a background model, and the time-invariant signal is extracted. The time-invariant signal is then eliminated by subtracting the background model from the original three-dimensional array to obtain a differential image sequence.

6. The infrared thermal wave nondestructive testing method according to claim 5, characterized in that: The S5 method is as follows: each frame of the differential image sequence is flattened row by row into a one-dimensional row vector to form a two-dimensional matrix, the eigenvalues ​​and eigenvectors of the covariance matrix are calculated, the eigenvectors corresponding to the first 4 to 6 largest eigenvalues ​​are selected, the infrared image sequence is reconstructed, and the principal component with the smallest background noise and the largest defect signal is selected as the optimal principal component for segmentation.

7. The infrared thermal wave nondestructive testing method according to claim 6, characterized in that: The specific method for segmenting and calculating the number of pixels in the defective area in S6 is: Select the pixel point corresponding to the center of the image defect of the best principal component as the initial segmentation area, and for each pixel on the boundary of the current segmentation area, , calculate the pixel and its 8-neighborhood unsegmented pixels If the attraction exceeds the set attraction threshold, pixel j is added to the segmented area. Repeat the operation until all pixels that meet the conditions are included, and the number of pixels in the defect area is obtained.

8. The infrared thermal wave nondestructive testing method according to claim 7, characterized in that: The specific method of S7 is: by using the laser ranging function of the infrared thermal imager, the actual size of the measured component and the corresponding number of pixels in the infrared image are compared to obtain the size of each pixel in the infrared image, and the measurement area of ​​each defect is calculated respectively.

Images