An Infrared Thermography Detection Feature Extraction Method for Impact Damage of Composite Materials

Abstract

The present invention provides a method for extracting infrared thermal imaging detection features of composite material impact damage. The specific steps include: (1) detecting a test block through a thermal imaging detection device, collecting thermal imaging information diagrams of the test block, and forming a sequence of thermal imaging information diagrams; (2) selecting thermal imaging information diagrams with information on the impact damage area, and selecting two or more thermal imaging information diagrams during the process from the initial appearance to the gradual disappearance of the impact damage area to form a maximum thermal imaging information diagram sequence; (3) preprocessing the maximum thermal imaging information diagram to enhance the image contrast; (4) postprocessing the maximum thermal imaging information diagram; (5) obtaining the information on the defect area of the test block, that is, the impact damage area, damage length, damage width, and maximum damage diameter feature information of the defect area of the test block.

Description

Technical Field

[0001] The present invention relates to the technical field of thermal imaging detection, and particularly relates to a method for extracting infrared thermal imaging detection features of impact damage of composite materials. Background Art

[0002] With the improvement of the manufacturing process of carbon fiber reinforced epoxy matrix composites, the specific strength and specific modulus of carbon fiber composites have gradually increased, and their applications in aircraft structures have gradually transitioned from non-load-bearing structures and secondary load-bearing structures to load-bearing structures. In the 1960s, carbon fiber reinforced epoxy matrix composites were first applied to non-load-bearing structure parts such as fairings and cabin doors. By the end of the 20th century, carbon fiber composites were widely used in secondary load-bearing structure parts such as vertical tails, horizontal tails, and air intakes. At present, in some aircraft, the main load-bearing structure parts such as the fuselage and the central wing box also use carbon fiber composites instead of the original metal materials.

[0003] Due to the use of epoxy thermosetting resin as the matrix, carbon fiber composites have low toughness and high brittleness, which makes them prone to impact damage during manufacturing and service, resulting in a sharp decline in the remaining strength of carbon fiber composites and seriously threatening the normal flight of aircraft. At present, ultrasonic testing technology and infrared thermal imaging technology are generally used for non-destructive testing of carbon fiber composite impact damage defects. Ultrasonic testing requires a coupling agent, has a complex testing process, and low testing efficiency. Infrared thermal imaging detection is widely used in non-destructive testing projects of carbon fiber composites because of its simple testing process, large one-time imaging area, non-contact and pollution-free characteristics.

[0004] The infrared thermal imaging non-destructive testing technology for aircraft carbon fiber composites uses active detection. The object to be detected is actively heated by a thermal excitation source, and an infrared thermal imager is used to detect the temperature change on the surface of the workpiece to identify defect information. However, in the actual detection process, due to the influence of uneven heating, problems such as low signal-to-noise ratio, difficult recognition of defect contours, and inaccurate defect quantification often occur in the infrared thermal images of carbon fiber composite impact damage.

[0005] For example, a patent document with a Chinese application number of CN202110357421.X and a publication date of June 18, 2021 discloses a method for detecting the size of defects in a carbon fiber laminate by infrared thermal imaging, which includes the following steps: S1, making a simulated defect test block; S2, cleaning the workpiece to be detected; S3, adjusting and setting the infrared thermal imaging detection system; S4, selecting the optimal defect image for saving; S5, determining the number of contour pixels of the defect; S6, obtaining the relationship curve between the correction coefficient and the depth; S7, determining the number of contour pixels of the defect in the workpiece to be detected; S8, obtaining the defect depth of the workpiece to be detected based on the logarithmic second derivative peak method; S9, reading the correction coefficient according to the fitted relationship curve; S10, calculating the defect size of the workpiece to be detected. The invention uses the same base material as the actual product material to make the defect simulation test block, ensuring the consistency of the defect size quantification process; the invention reduces the influence of lateral heat diffusion on size quantification through the reference calibration of defects with different depths, greatly improving the accuracy of defect quantification.

[0006] For the detection method in this literature, after only performing infrared thermal imaging detection on the workpiece to be detected, the optimal defect image is selected according to the characteristic image for defect size detection. The principle of selecting the optimal image is to identify the defect area by distinguishing the temperature difference between the defect area and the intact area, and select the image captured at the moment when the defect is optimally displayed. In view of the current problems of low signal-to-noise ratio of the external thermal map of composite materials and inaccurate defect feature extraction, currently, mainly image processing methods for single-frame thermal maps such as contrast enhancement and filtering processing, threshold segmentation, region growing, and principal component analysis are used. All these studies subjectively identify the optimal thermal map in the thermal map sequence of composite material impact damage by the naked eye, and do not propose how to select an optimal thermal map with quantitative performance indicators in the entire thermal map sequence. In this way, it is easy to reduce the accuracy of defect information feature extraction of composite material impact damage. Summary of the Invention

[0007] The present invention provides a method for extracting features of infrared thermal imaging detection of composite material impact damage; through the method of the present invention, problems such as difficult recognition of defect contours and inaccurate defect quantification due to uneven heating of composite materials during the detection process can be solved, and the accuracy of defect information feature extraction of composite material impact damage can be effectively improved.

[0008] To achieve the above object, the technical solution of the present invention is: a method for extracting features of infrared thermal imaging detection of composite material impact damage, and the specific steps include:

[0009] (1) Detect the test block through a thermal imaging detection device, preset the acquisition frequency, collect the thermal imaging information map of the test block according to the preset acquisition frequency, and form a thermal imaging information map sequence.

[0010] (2)Select the thermal imaging information map with the information of the impact damage area, select more than two thermal imaging information maps during the process from the initial appearance to the gradual disappearance of the impact damage area to form the maximum thermal imaging information map sequence, then obtain the temperature difference of the maximum thermal imaging information map sequence and form the temperature difference time history curve, generate the maximum gradient value of the temperature difference time history curve according to the temperature difference time history curve, and the thermal map corresponding to the frame number with the maximum gradient value is the maximum thermal imaging information map.

[0011] (3)Preprocess the maximum thermal imaging information map to improve the image contrast.

[0012] (4)Postprocess the maximum thermal imaging information map.

[0013] (5)Obtain the information of the defect area of the test block, that is, the impact damage area, damage length, damage width, and maximum damage diameter characteristic information of the defect area of the test block.

[0014] In the above method, the impact damage condition of the composite material test block is detected by a thermal imaging detection device, and more than two thermal imaging information maps are selected to form the maximum thermal imaging information map sequence. According to the temperature difference-time history curve of the impact damage area on the test block, and the frame number corresponding to the maximum gradient value is determined according to the temperature difference time history curve to determine the maximum thermal imaging information map. Then, after image processing of the maximum thermal imaging information map, the corresponding damage parameters are extracted. Since the maximum thermal imaging information map is determined according to the maximum gradient value formed by the temperature difference time history curve, the maximum thermal imaging information map is the actually damaged area. Then, by processing the image and extracting the corresponding parameters, a quantitative index can be proposed for how to extract the single-needle optimal thermal map of the composite material impact damage characteristics, which can effectively improve the accuracy of extracting the information characteristics of the composite material impact damage defects.

[0015] Further, step (1) specifically includes irradiating the test block with a thermal excitation source emitted by a halogen lamp for 3S - 5S, with a collection frequency of 5 - 15Hz, and a total of 900 - 1100 frames of thermal imaging information maps are collected.

[0016] Further, step (2) specifically includes selecting the number of frames of more than two thermal imaging information maps that can completely display the process from the appearance to the gradual disappearance of the impact damage area. The above method can make the selected number of frames of the thermal imaging information map completely display the entire process of the impact damage area, making the extraction method more accurate.

[0017] Further, step (3) specifically includes:

[0018] (3.1)Perform gray-scale transformation on the maximum thermal imaging information map.

[0019] (3.2) The maximum thermal imaging information map is preprocessed by histogram equalization to enhance the image contrast.

[0020] (3.3) The image after histogram equalization is processed by high-pass filtering.

[0021] The above method filters the image after gray-scale change, enabling the thermal imaging information map to be less affected by noise and ensuring the accuracy of the extracted feature values.

[0022] Further, step (4.2) specifically includes: processing the edge of the defect area by passivation masking.

[0023] The specific steps of the passivation masking include:

[0024] (4.21) Blur the image of the defect area to obtain a blurred area image.

[0025] (4.22) Subtract the blurred area image from the original area image.

[0026] (4.23) Take the sum of the original area image and the blurred area image as a template.

[0027] (4.24) Add the template to the original area image; obtain the area image after passivation masking.

[0028] The above method processes the edge of the defect area by passivation masking, further improving the accuracy of the image.

[0029] Further, step (1) specifically includes irradiating the test block with a thermal excitation source emitted by a halogen lamp for 4S, with a collection frequency of 10Hz, and a total of 1000 frames of thermal imaging information maps are collected.

[0030] The above method can obtain the maximum number of frames of the formed thermal imaging information map by adopting this parameter.

[0031] Further, step (2) specifically includes: if there is no thermal imaging information map with the information of the impact damage area within the preset collection frequency, then increase the collection frequency.

[0032] The above method increases the collection frequency to further expand the number of collected frames and ensure the acquisition readiness when there is no information of the impact damage area.

[0033] Further, step (2) specifically also includes: if there is still no thermal imaging information map with the information of the impact damage area within the increased collection frequency, then determine that the test block has not been impacted; if an impact damage area appears, select more than two thermal imaging information maps during the process from the start to the gradual disappearance of the impact damage area and form a maximum thermal imaging information map sequence. Description of the Drawings

[0034] Figure 1 It is the thermal image of Embodiment 1 of the present invention that can completely display the process of the impact damage area from appearance to gradual disappearance.

[0035] Figure 2 It is the schematic diagram of the region of interest of Embodiment 1 of the present invention.

[0036] Figure 3 It is the schematic diagram of the temperature difference - time (frame number) history curve of the test block of the present invention.

[0037] Figure 4 It is the schematic diagram of the gradient value of the rising edge of the temperature difference - time history curve of the test block of Embodiment 1 of the present invention.

[0038] Figure 5 It is the maximum thermal imaging information diagram of Embodiment 1 of the present invention.

[0039] Figure 6 It is the best thermal image after histogram equalization processing in Embodiment 1 of the present invention.

[0040] Figure 7 It is the best thermal image after high - pass filtering processing in Embodiment 1 of the present invention.

[0041] Figure 8 It is the best thermal image after passivation masking method processing in Embodiment 1 of the present invention.

[0042] Figure 9 It is the thermal image of Embodiment 2 of the present invention that can completely display the process of the impact damage area from appearance to gradual disappearance.

[0043] Figure 10 It is the schematic diagram of the region of interest of Embodiment 2 of the present invention.

[0044] Figure 11 It is the schematic diagram of the gradient value of the rising edge of the temperature difference - time history curve of the test block of Embodiment 2 of the present invention.

[0045] Figure 12 It is the best thermal image of Embodiment 2 of the present invention.

[0046] Figure 13 It is the best thermal image after histogram equalization processing in Embodiment 2 of the present invention.

[0047] Figure 14 It is the best thermal image after high - pass filtering processing in Embodiment 2 of the present invention.

[0048] Figure 15 It is the best thermal image after passivation masking method processing in Embodiment 2 of the present invention. Detailed implementation manners

[0049] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0050] Embodiment 1.

[0051] As Figures 1 - 8 shown, a method for extracting infrared thermal imaging detection features of impact damage of a composite material, the specific steps include:

[0052] (1) Detect the test block through a thermal imaging detection device, preset the acquisition time and acquisition frequency, and collect the thermal imaging information map of the test block within the irradiation time according to the preset acquisition frequency, and form a sequence of thermal imaging information maps. Specifically, the preset acquisition irradiation time is 3S - 5S, and the preset acquisition frequency is 5 - 15HZ.

[0053] (2) Select the thermal imaging information map with the information volume of the impact damage area, select two or more thermal imaging information maps during the process from the start to the gradual disappearance of the impact damage area to form the maximum thermal imaging information map sequence, then obtain the temperature difference of the maximum thermal imaging information map sequence and form a temperature difference time history curve, generate the maximum gradient value of the temperature difference time history curve according to the temperature difference time history curve, and the thermal map corresponding to the frame number corresponding to the maximum gradient value is the maximum thermal imaging information map.

[0054] (3) Preprocess the maximum thermal imaging information map to improve the image contrast.

[0055] (4) Post-process the maximum thermal imaging information map.

[0056] (5) Obtain the information of the defect area of the test block, that is, the impact damage area, damage length, damage width, and maximum damage diameter characteristic information of the defect area of the test block.

[0057] Detect the impact damage condition of the composite material test block through a thermal imaging detection device, select two or more thermal imaging information maps to form the maximum thermal imaging information map sequence, according to the temperature difference - time history curve of the impact damage area on the test block, and determine the frame number corresponding to the maximum gradient value according to the temperature difference time history curve to determine the maximum thermal imaging information map, then perform image processing on the maximum thermal imaging information map and extract the corresponding damage parameters. Since the maximum thermal imaging information map is determined according to the maximum gradient value formed by the temperature difference time history curve, the maximum thermal imaging information map is the actually damaged area, and then by processing the image and extracting the corresponding parameters, a quantitative index can be proposed for how to extract the single-needle optimal thermal map of the impact damage characteristics of the composite material, which can effectively improve the accuracy of extracting the defect information characteristics of the impact damage of the composite material.

[0058] Step (1) specifically includes irradiating the test block with a thermal excitation source emitted by a halogen lamp for 3S - 5S, with a collection frequency of 5 - 15Hz, and a total of 900 - 1100 frames of thermal imaging information graphs are collected. In this embodiment, the size of the test block A is 150*100mm, the thickness is 3.89mm, the impact capacity on the test block is 72.19J, and the thermal imaging detection device uses four halogen lamps with a power of 500W as the thermal excitation source to irradiate the test block. The halogen lamp emits a thermal excitation source to irradiate the test block for 4S, the collection frequency is 10Hz, and a total of 1000 frames of thermal imaging information graphs (hereinafter referred to as thermal graphs) are collected.

[0059] Step (2) specifically includes selecting the number of frames of thermal imaging information graphs that can completely display the process of the impact damage area from appearance to gradual disappearance. In this embodiment, the operator selects the 39th to 186th frames from 1000 frames of thermal imaging information graphs that can completely display the process of the impact damage area from appearance to gradual disappearance by observing the state change of the thermal graph. As Figure 1 shown, the 39th is the thermal graph of the damage area in the 39th frame, and its damage area is small. By the 186th frame, its area increases to a large size, and between the 39th frame and the 186th frame, the damage area is in an increasing state.

[0060] If there is no thermal imaging information graph with the information volume of the impact damage area within the preset collection frequency, the collection frequency is increased. Specifically, the collection frequency is increased to 10HZ.

[0061] If there is still no thermal imaging information graph with the information volume of the impact damage area within the increased collection frequency, it is determined that the test block has not been impacted.

[0062] If an impact damage area appears, two or more thermal imaging information graphs of the process from the beginning to the gradual disappearance of the impact damage area are selected and a maximum thermal imaging information graph sequence is formed.

[0063] In each frame of the thermal graph that can completely display the process of the impact damage area from appearance to gradual disappearance, a plurality of regions of interest are picked up on the impact damage area and the non-impact damage area. The picking process is that the operator randomly selects a preset number of regions of interest. In this embodiment, the regions of interest are multiple regions within the impact damage area and the non-impact damage area. The operator picks 8 regions of interest in the impact damage area and 4 regions of interest in the non-impact damage area. Among them, the 8 regions of interest in the impact damage area are arranged along a similar circumferential direction. As Figure 2 shown.

[0064] Taking the temperature difference obtained from the average temperatures of the regions of interest in the impact-damaged area and the non-impact-damaged area as the ordinate, that is, the difference between the average temperature of the impact-damaged area and the average temperature of the non-impact-damaged area as the ordinate, and taking the number of frames as the abscissa, plot the temperature difference-time (number of frames) history curve of the test block, as Figure 3 shown. In this embodiment, the test block is Figure 3 the line segment represented by test block A in

[0065] From Figure 3 it can be seen that the test block reaches the maximum temperature difference at the 110th frame, and through a partial thermal imaging information diagram of the test block such as Figure 1 it is known that according to the trend of Figure 1 , the impact damage area shown by test block A at the 110th frame is in a decreasing state, not as large as the impact damage area shown in the thermal images before the 110th frame, indicating that the thermal image that can best display the impact damage surface of the test block, that is, the optimal thermal image in the thermal image sequence, is before the maximum temperature difference. Therefore, the maximum thermal imaging information diagram that can extract the impact damage characteristics of the test block should be in the rising edge of the temperature difference-time history curve. Then, differentiate the temperature difference-time history curve corresponding to the thermal imaging information diagram that can clearly show the impact damage characteristics of the test block in the rising edge, such as the 39th - 110th frames of the test block, so as to obtain the maximum gradient value of the rising edge of the temperature difference-time history curve. The differential solution formula is:

[0066] gradu = ∂u / ∂x

[0067] In the formula, grad is the gradient value of the rising edge of the temperature difference-time history curve, ∂u represents the temperature difference value corresponding to the temperature between adjacent frames, and ∂x represents the number of frames of the interval of the thermal imaging information diagram. For example, the gradient value of the 3rd frame is the difference between the temperature of the 3rd frame and the temperature of the 2nd frame divided by the difference between the 3rd frame and the 2nd frame number, and it is preset that the temperature value corresponding to the 0th frame is 0. As Figure 4 shown.

[0068] The test block reaches the maximum gradient value at the 60th frame, that is, the thermal imaging information diagram of the test block at the 60th frame is the maximum thermal imaging information diagram; as Figure 5 shown.

[0069] In one embodiment, step (3) specifically includes:

[0070] Step (3.1) specifically includes first performing gray-scale transformation on the maximum thermal imaging information diagram to improve the image quality of the maximum thermal imaging information diagram and make the display effect of the image clearer.

[0071] Step (3.2) specifically includes using the histogram equalization processing method to perform image preprocessing on the maximum thermal imaging information diagram of the test block to enhance the contrast. As Figure 6 shown.

[0072] In this embodiment, the histogram equalization processing method is a common method for adjusting contrast using an image histogram in the field of image processing. The specific adjustment method is a prior art and will not be described again here.

[0073] Step (3.3) specifically includes applying high-pass filtering to the image after histogram equalization to eliminate small noise signals and enhance the contrast between the defective area and the non-defective area.

[0074] In this embodiment, the high-pass filtering method is a common noise reduction method used in image processing in the field of image processing. It is a prior art and will not be described here. Figure 7 shown.

[0075] In one embodiment, step (4) specifically includes:

[0076] Step (4.2) specifically includes: processing the edge of the defect area by passivation masking.

[0077] The passivation masking specific steps include:

[0078] (4.21) The defect area image is blurred to obtain a blurred area image.

[0079] (4.22) Subtract the blurred region image from the original region image.

[0080] (4.23) The original area image and the blurred area image are added together to form a template.

[0081] Let f1(x,y) represent the blurred region image, the formula is:

[0082] g mask (x,y)=f(x,y)-f1(x,y)

[0083] Where f(x,y) represents the grayscale value of the original region image at the coordinate (x,y), g mask (x, y) is the template image function at the coordinate (x, y).

[0084] (4.24) Add the template to the original region image to obtain the region image after passivation masking. Figure 8 shown.

[0085] g(x,y)=f(x,y)+g mask (x,y)

[0086] Where g(x,y) represents the image of the area after unsharp masking at the coordinate (x,y).

[0087] In one embodiment, step (5) specifically includes measuring the impact damage area A1, damage length A2, damage width A3, and maximum damage diameter A4 characteristic information of the defect area from the area image after passivation masking treatment.

[0088] To further verify the effectiveness of this method, measurements are carried out. The measurement method can be achieved by ultrasonic C-scan technology or infrared thermal imaging (adopted in this embodiment). By using ultrasonic C-scan technology to detect the impact damage area, damage length, damage width, and maximum damage diameter characteristic information of the test block, and then comparing it with the method of this embodiment and calculating the error, the results are shown in Table 1 below:

[0089]

[0090] Table 1

[0091] It can be seen from the table that the overall error is small and controllable. Therefore, the method of this embodiment is feasible.

[0092] Embodiment 2.

[0093] As Figure 3 、 Figures 9 - 15 shown, a method for extracting infrared thermal imaging detection features of composite material impact damage specifically includes the following steps:

[0094] (1) Detect the test block through a thermal imaging detection device, preset the acquisition time and acquisition frequency, and collect the thermal imaging information map of the test block within the irradiation time according to the preset acquisition frequency, and form a sequence of thermal imaging information maps. Specifically, the preset acquisition irradiation time is 3S - 5S, and the preset acquisition frequency is 5 - 15HZ.

[0095] (2) Select the thermal imaging information map with impact damage area information, select two or more thermal imaging information maps during the process from the start to the gradual disappearance of the impact damage area to form the maximum thermal imaging information map sequence, then obtain the temperature difference of the maximum thermal imaging information map sequence and form a temperature difference time history curve, generate the maximum gradient value of the temperature difference time history curve, and the thermal map corresponding to the frame number of the maximum gradient value is the maximum thermal imaging information map.

[0096] (3) Preprocess the maximum thermal imaging information map to enhance the image contrast.

[0097] (4) Post-process the maximum thermal imaging information map.

[0098] (5) Obtain the information of the defect area of the test block, that is, the impact damage area, damage length, damage width, and maximum damage diameter characteristic information of the defect area of the test block.

[0099] In the above method, the impact damage condition of the composite material specimen is detected by a thermal imaging detection device, and more than two thermal imaging information graphs are selected to form a maximum thermal imaging information graph sequence. According to the temperature difference-time history curve of the impact damage area on the specimen, and based on the temperature difference time history curve, the frame number corresponding to the maximum gradient value is determined to be the maximum thermal imaging information graph. Then, after image processing of the maximum thermal imaging information graph, the corresponding damage parameters are extracted. Since the maximum thermal imaging information graph is determined based on the temperature difference time history curve formed to obtain the maximum gradient value, the maximum thermal imaging information graph is the area actually damaged. Then, by processing the image and extracting the corresponding parameters, a quantitative index for the single-needle optimal thermal map for extracting the impact damage characteristics of the composite material can be ensured, which can effectively improve the accuracy of extracting the information characteristics of the impact damage defects of the composite material.

[0100] Step (1) specifically includes irradiating the specimen with a thermal excitation source emitted by a halogen lamp for 3S - 5S, with a collection frequency of 5 - 15Hz, and a total of 900 - 1100 frames of thermal imaging information graphs are collected. In this embodiment, the specimen used in this embodiment is different from that in Embodiment 1. The size of the specimen B is 150 * 10mm, the thickness is 3.11mm, and the impact capacity on the specimen is 18.29J. The thermal imaging detection device uses four halogen lamps with a power of 500W as the thermal excitation source to irradiate the specimen. The halogen lamp emits a thermal excitation source to irradiate the specimen for 4S, the collection frequency is 10Hz, and a total of 1000 frames of thermal imaging information graphs (hereinafter referred to as thermal maps) are collected.

[0101] Step (2) specifically includes selecting the number of frames of the thermal imaging information graph that can completely display the process of the impact damage area from appearance to gradual disappearance. In this embodiment, the operator observes the state change of the thermal map and selects the thermal imaging information graphs from the 1000 frames of thermal imaging information graphs that can completely display the process of the impact damage area from appearance to gradual disappearance, from the 22nd frame to the 108th frame. As Figure 9 shown, the 22th is the thermal imaging information graph of the damage area in the 22nd frame, and its damage area is small. By the 108th frame, its area increases to a large size, and between the 22nd frame and the 108th frame, the damage area is in an increasing state.

[0102] [[ID=ID=11]]If there is no thermal imaging information graph with the information volume of the impact damage area within the preset collection frequency, the collection frequency is increased. Specifically, the collection frequency is increased to 10HZ. [[ID=ID=12]] [[ID=ID=13]]

[0103] [[ID=ID=14]]If there is still no thermal imaging information graph with the information volume of the impact damage area within the increased collection frequency, it is determined that the specimen has not been impacted. [[ID=ID=15]] [[ID=ID=16]]

[0104] If an impact damage area appears, select more than two thermal imaging information diagrams during the process from the start to the gradual disappearance of the impact damage area and form a maximum thermal imaging information diagram sequence.

[0105] In each thermal image that can completely display the process of the impact damage area from appearance to gradual disappearance, pick up multiple regions of interest on the impact damage area and the non-impact damage area. The picking process is that the operator randomly selects a preset number of regions of interest. In this embodiment, the operator picks 8 regions of interest in the impact damage area and 4 regions of interest in the non-impact damage area, as Figure 10 shown.

[0106] Use the temperature difference obtained from the average temperatures of the regions of interest in the impact damage area and the non-impact damage area as the ordinate, that is, the difference between the average temperature of the impact damage area and the average temperature of the non-impact damage area as the ordinate, and use the number of frames as the abscissa to plot the temperature difference-time (number of frames) history curve of the test block, as Figure 3 shown. In this embodiment, the test block is Figure 3 the line segment represented by test block B in

[0107] From Figure 3 it can be seen that test block B reaches the maximum temperature difference at the 110th frame. And from some thermal images of the test block such as Figure 9 it can be known that according to the trend of Figure 9 , the impact damage area shown by test block B at the 110th frame is in a decreasing state, and is not as large as the impact damage area shown in the thermal images before the 110th frame. This shows that the thermal image that can best display the impact damage surface of the test block, that is, the maximum thermal imaging information diagram in the thermal image sequence, is before the maximum temperature difference. Therefore, the optimal thermal image for extracting the impact damage characteristics of the test block should be in the rising edge of the temperature difference-time history curve. Differentiate the temperature difference-time history curve corresponding to the thermal imaging information diagrams from the 22nd frame to the 108th frame of test block B that can clearly show the impact damage characteristics, so as to obtain the gradient value of the rising edge of the temperature difference-time history curve. The differential solution formula is:

[0108] gradu = ∂u / ∂x

[0109] In the formula, grad is the gradient value of the rising edge of the temperature difference-time history curve, ∂u represents the temperature difference value corresponding to the temperature between adjacent frames, and ∂x represents the number of frames between the thermal imaging information diagrams. For example, for the gradient value of the 3rd frame, it is the difference between the temperature of the 3rd frame and the temperature of the 2nd frame divided by the difference between the 3rd frame and the 2nd frame. It is preset that the temperature value corresponding to the 0th frame is 0. As Figure 11 shown.

[0110] In another embodiment, after obtaining the maximum temperature difference, in combination with Figure 9From the shown thermal imaging information diagram, it can be seen that at the 80th frame, the loss area of the thermal imaging information diagram has started to become smaller. Therefore, the differential of the temperature difference time history curve corresponding to the thermal image in the 22nd - 80th frames of the test block B, where the impact damage characteristics can be clearly seen, is selected, and then the gradient value of the rising edge of the temperature difference time history curve is obtained. In this way, by further combining the thermal imaging information diagram, the calculation structure becomes more accurate. The differential solution formula is:

[0111] gradu = ∂u / ∂x

[0112] In the formula, grad is the gradient value of the rising edge of the temperature difference time history curve, ∂u represents the temperature difference value corresponding to the temperature between adjacent frames, and ∂x represents the number of frames of the interval of the thermal imaging information diagram. For example, for the gradient value of the 3rd frame, it is the difference between the temperature of the 3rd frame and the temperature of the 2nd frame divided by the difference between the 3rd frame and the 2nd frame. It is preset that the temperature value corresponding to the 0th frame is 0. As Figure 11 shown.

[0113] The test block reaches the maximum gradient value at the 52nd frame, that is, the thermal image of the test block at the 52nd frame is the best thermal image; as Figure 12 shown.

[0114] In one embodiment, step (3) specifically includes:

[0115] Step (3.1) specifically includes first performing gray-scale transformation on the maximum thermal imaging information diagram to improve the image quality of the maximum thermal imaging information diagram and make the display effect of the image clearer.

[0116] Step (3.2) specifically includes using the histogram equalization processing method to perform image preprocessing on the maximum thermal imaging information diagram of the test block to enhance the contrast. As Figure 13 shown.

[0117] In this embodiment, the histogram equalization processing method is a common method in the field of image processing for adjusting the contrast using the image histogram. Its specific adjustment method is prior art and will not be elaborated here.

[0118] Step (3.3) specifically includes performing high-pass filtering on the image after histogram equalization to eliminate small noise signals and enhance the contrast between the defect area and the non-defect area.

[0119] In this embodiment, the high-pass filtering processing method is a common noise reduction method for image processing in the field of image processing. Specifically, it is prior art and will not be elaborated here. The image after high-pass filtering processing is as Figure 14 shown.

[0120] In one embodiment, step (4) specifically includes:

[0121] Step (4.2) specifically includes: processing the edge of the defect area through passivation masking.

[0122] The specific steps of the passivation masking include:

[0123] (4.21) Blurring the defect area image to obtain a blurred area image.

[0124] (4.22) Subtracting the blurred area image from the original area image.

[0125] (4.23) Taking the sum of the original area image and the blurred area image as a template.

[0126] Let f1(x, y) represent the blurred image, and the formula is:

[0127] g mask (x, y) = f(x, y) - f1(x, y)

[0128] In the formula, f(x, y) represents the gray value of the original area image at the position of coordinates (x, y), and g mask (x, y) is the template image function at the position of coordinates (x, y).

[0129] (4.24) Adding the template to the original area image to obtain the passivation masked area image as Figure 15 shown.

[0130] g(x, y) = f(x, y) + g mask (x, y)

[0131] In the formula, g(x, y) represents the passivation masked area image at the position of coordinates (x, y).

[0132] In an embodiment, step (5) specifically includes measuring the impact damage area B1, damage length B2, damage width B3, and maximum damage diameter B4 characteristic information of the defect area from the passivation masked image.

[0133] To further verify the effect of this method, it is carried out through measurement. The measurement method can be realized by ultrasonic C-scan technology or infrared thermal imaging (adopted in this embodiment). By using ultrasonic C-scan technology to detect the impact damage area, damage length, damage width, and maximum damage diameter characteristic information of the test block, and then comparing it with the method of this embodiment and calculating the error, the results are shown in Table 2 below:

[0134]

[0135] Table 2

[0136] As can be seen from the table, the overall error is small and controllable. Therefore, the method of this embodiment is feasible.

Claims

1. A method for extracting infrared thermal imaging detection features of composite material impact damage, characterized in that: The specific steps include: (1) Detect the test block with a thermal imaging detection device, preset the acquisition frequency, collect the thermal imaging information maps of the test block according to the preset acquisition frequency, and form a sequence of thermal imaging information maps; (2) Select the thermal imaging information maps with the information of the impact damage area, select more than two thermal imaging information maps in the process from the start to the gradual disappearance of the impact damage area to form the maximum thermal imaging information map sequence, then obtain the temperature difference of the maximum thermal imaging information map sequence and form a temperature difference time history curve, generate the maximum gradient value of the temperature difference time history curve according to the temperature difference time history curve, and the thermal map corresponding to the frame number of the maximum gradient value is the maximum thermal imaging information map; pick up multiple regions of interest on the impact damage area and the non-impact damage area in each frame of the thermal map that can completely display the process from the appearance to the gradual disappearance of the impact damage area. The picking process is that the operator randomly selects a preset number of regions of interest; use the average temperature of the regions of interest in the impact damage area and the non-impact damage area to obtain the temperature difference as the ordinate, that is, the difference between the average temperature of the impact damage area and the average temperature of the non-impact damage area as the ordinate, and use the frame number as the abscissa to draw the temperature difference time history curve of the test block; (3) Preprocess the maximum thermal imaging information map to enhance the image contrast; (4) Post-process the maximum thermal imaging information map; (5) Obtain the information of the defect area of the test block, that is, the impact damage area, damage length, damage width, and maximum damage diameter characteristic information of the defect area of the test block.

2. A method for extracting features of infrared thermal imaging detection of impact damage of a composite material according to claim 1, characterized in that: Step (1) specifically includes irradiating the test block with a thermal excitation source emitted by a halogen lamp for 3S - 5S, the acquisition frequency is 5 - 15Hz, and a total of 900 - 1100 frames of thermal imaging information maps are collected.

3. A method for extracting features of infrared thermal imaging detection of impact damage of a composite material according to claim 2, characterized in that: Step (2) specifically includes selecting the number of frames of more than two thermal imaging information maps that can completely display the process from the appearance to the gradual disappearance of the impact damage area.

4. The method for extracting the infrared thermal imaging detection characteristics of the impact damage of the composite material according to claim 1, characterized in that: Step (3) specifically includes: (3.1) Perform gray-scale transformation on the maximum thermal imaging information map; (3.2) Use the histogram equalization processing method to preprocess the maximum thermal imaging information map to enhance the image contrast; (3.3) Perform high-pass filtering on the image after histogram equalization.

5. A method for extracting features of infrared thermal imaging detection of impact damage of a composite material according to claim 1, characterized in that: Step (4) specifically includes: processing the edge of the defect area through passivation masking; The specific steps of the passivation masking include: (4.21) Blur the defect area image to obtain the blurred area image; (4.22) Subtract the blurred area image from the original area image; (4.23) Take the sum of the original area image and the blurred area image as the template; (4.24) Add the template to the original area image; obtain the area image after passivation masking.

6. A method for extracting infrared thermal imaging detection features of impact damage of a composite material according to claim 2, characterized in that: Step (1) specifically includes irradiating the test block with a thermal excitation source emitted by a halogen lamp for 4S, the acquisition frequency is 10Hz, and a total of 1000 frames of thermal imaging information maps are collected.

7. A method for extracting infrared thermal imaging detection features of impact damage of a composite material according to claim 6, characterized in that: Step (2) specifically includes that if there is no thermal imaging information map with the information of the impact damage area within the preset acquisition frequency, then increase the acquisition frequency.

8. A method for extracting infrared thermal imaging detection features of impact damage of a composite material according to claim 1, characterized in that: Step (2) specifically further includes that if there is still no thermographic information map of the impact damage area information within the increased preset acquisition frequency, it is determined that the test block has not been impacted and damaged; if an impact damage area appears, more than two thermographic information maps during the process from the start to the gradual disappearance of the impact damage area are selected and a maximum thermographic information map sequence is formed.

Images