Defect detection method and system for prepreg laying layer and computer storage medium

Abstract

The invention provides a defect detection method and system for a prepreg laying layer and a computer storage medium, and the method comprises the steps: carrying out the frequency conversion thermal excitation of a to-be-detected prepreg laying layer, collecting an infrared phase-locked diagram of the to-be-detected prepreg laying layer, and converting the infrared phase-locked diagram into an infrared phase-locked gray-scale map; carrying out boundary identification on the infrared phase-locked grey-scale map, and then carrying out image segmentation on the infrared phase-locked grey-scale map by taking a grey-scale threshold value as a constraint condition to obtain an infrared phase-locked grey-scale map with divided grey-scale regions; and performing shape recognition on the gray region greater than the gray threshold to obtain a region containing a singular point and / or an edge curve, and further obtaining the defect type of the to-be-detected prepreg paving layer and the total area of the defect region. According to the method, the linearly increased variable-frequency infrared excitation source is adopted to excite the prepreg paving layer, so that defect information of different depths can be excited, infrared signals radiated by near-surface and deeper defects can be detected by an external detector, and the problem that deep defects are difficult to detect by an infrared phase locking method in the prior art is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of the production and layup of composite prepregs, and particularly relates to a method, a system and a computer storage medium for detecting defects in prepreg layup. Background Art

[0002] Fiber reinforced composite prepreg refers to reinforced fibers (such as carbon fiber, glass fiber, aramid fiber, etc.) that have been impregnated with a resin matrix, and the resin has been partially cured (B-stage) for subsequent final forming. Prepregs are usually used to manufacture high-performance composite parts and are widely used in industries such as aerospace, automotive, and sports goods. During the production process, different resin systems and reinforced fibers will have different production process parameters, such as temperature, pressure, time, etc., which need to be adjusted according to specific circumstances. If the adjustment is improper, defects such as bubbles, delamination, and stratification will occur. The quality of the prepreg directly affects the structure and performance of the final composite product. Therefore, strict inspection of the prepreg is required before production and use to identify possible defects.

[0003] The common prepreg defects and their identification methods are as follows: First, visual inspection, observing whether there are obvious damages, cracks, stains or foreign objects by the naked eye; and touching the surface of the prepreg by hand to check whether there are rough or uneven places. Second, thickness measurement, using an ultrasonic thickness gauge or a contact thickness gauge to measure the thickness of the prepreg to confirm whether it meets the design requirements. Third, resin content determination, determining the resin content by weighing the mass of the prepreg and calculating its density to check whether it is within the specified range. Fourth, infrared spectroscopy, using infrared spectroscopy to analyze the resin content, which can provide more accurate data. Fifth, vacuum suction method, evaluating the void situation in the prepreg by vacuum suction. Sixth, focused X-ray CT scanning or X-ray imaging, using X-ray computed tomography technology to observe the internal pore situation; or detecting whether there are defects in the internal structure of the prepreg by X-ray imaging technology. Seventh, density method, calculating the fiber volume fraction by measuring the density and combining with the theoretical value. Eighth, microscopic observation, observing the fiber distribution using an optical microscope or an electron microscope. Ninth, mechanical property testing, conducting a tensile test on the prepreg to evaluate its strength and elastic modulus; testing the rigidity and toughness of the prepreg by a bending test. Tenth, thermal analysis method and thermogravimetric analysis method, using differential scanning calorimetry (DSC) to analyze the thermal properties of the prepreg, such as the starting temperature and heat release of the curing reaction; using thermogravimetric analysis (TGA) to evaluate the thermal stability of the prepreg, especially the performance change at high temperatures. Eleventh, non-destructive testing (NDE), using ultrasonic flaw detection technology to detect defects such as delamination and cracks in the prepreg.

[0004] Among the above analysis methods, the relatively important non-destructive detection methods include ultrasonic flaw detection, X-ray imaging, and infrared image detection. Among them, due to the scattering interference in the near-surface area of ultrasonic waves, when detecting composite near-surface defects with ultrasonic waves, there are interference and clutter, and the near-surface image is blurred. Moreover, the products made of prepreg are often relatively thin, so the ultrasonic flaw detection method has certain limitations and is not suitable for defect detection of prepreg products. Although the X-ray imaging method can accurately present the defect characteristics and distribution characteristics inside the material, due to its certain radiation and the extremely high investment cost of the entire equipment, it is not conducive to the layout of the production line. The application of infrared detection technology in prepreg defect detection is an evolving field, and many research institutions and enterprises are exploring and developing new technologies and methods. Among them, Lock-in Thermography (LIT) is an advanced non-destructive testing technology, especially suitable for detecting subtle defects inside materials. This method uses a periodic heating signal to excite the material and captures the change in the surface temperature of the material over time through an infrared camera. By analyzing these temperature fluctuations, the location, size, and nature of the defects inside the material can be identified. Lock-in infrared detection is based on the thermo-wave principle, that is, the material is periodically heated by an external heat source (infrared light source or laser), and then the infrared camera records the temperature change on the surface of the material. Defects inside the material (such as cracks, delaminations, pores, etc.) will affect the propagation of thermo-waves, resulting in different local temperature responses from the normal area. Therefore, through Fourier transform or other signal processing techniques, the frequency components related to defects can be extracted from the temperature data.

[0005] Although the infrared lock-in method has many advantages in the detection of prepreg defects, it also has certain limitations. First, depth limitation. The infrared lock-in method is more suitable for detecting surface or near-surface defects (based on the heat diffusion equation). For deep defects, the attenuation of thermal waves may lead to signal weakening, making it difficult to detect fine defects deep inside. Second, heating penetration depth limitation. The penetration ability of the heating source is limited. For thick parts or materials with high thermal conductivity, higher energy input or longer detection time may be required, which may affect the detection efficiency. Third, material property influence. The prepreg itself is a composite structure, and its thermophysical properties are different from those of traditional metal materials. Thermophysical properties such as the thermal conductivity and heat capacity of the material will affect the detection results. Fourth, absorption coefficient influence. The absorption coefficient of the material surface has a great influence on the heating effect, and different surface treatments or coatings may affect the detection results. Fifth, environmental interference. Background temperature fluctuations. Changes in the ambient temperature will affect the detection results, so detection needs to be carried out in a stable environment. Sixth, complex data processing. Lock-in infrared detection requires complex signal processing such as Fourier transform to extract useful information, which increases the time and difficulty of data processing. Seventh, result interpretation. Even after signal processing, sometimes expert knowledge is needed to correctly interpret the detection results, especially in the case of complex defects.

[0006] Patent CN114577814A proposes a detection scheme based on the lock-in infrared method, but this method has the defect that it is difficult to give the accurate defect spatial position when detecting defects in ICs with multi-layer metal structures. The heat transfer coefficient of the composite prepreg does not belong to metal materials, resulting in its internal defects being insensitive to depth information, so the lock-in infrared method cannot be directly used for defect detection. Patent CN103149217A provides a method for detecting surface and subsurface defects of optical components. Affected by the transmission ability of the optical device to infrared light waves, there are certain differences in its application. For fiber-reinforced prepregs, this algorithm cannot be directly used to process the defect detection inside the prepreg. Patent CN117092163A discloses a double-pulse non-destructive detection method for sandwich structure composites, which uses sinusoidally varying thermal waves to excite the material to be detected, and still cannot solve the above seven problems.

[0007] In view of this, developing a defect detection method that can excite defect information at different depths of prepreg plies has become an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0008] The purpose of the present invention is to solve the problems existing in the above-mentioned prior art, and provides a defect detection method, system and computer storage medium for prepreg plies. Starting from the control parameters of the infrared lock-in method, corresponding data processing methods are carried out according to the characteristics of prepreg defect detection to realize the extraction and identification of defect information.

[0009] One of the objects of the present invention is to provide a method for detecting defects in a prepreg laminate, including

[0010] S1: Apply variable-frequency thermal excitation to the prepreg laminate to be detected and collect the infrared lock-in thermography of the prepreg laminate to be detected;

[0011] S2: Convert the infrared lock-in thermography into an infrared lock-in grayscale image, perform boundary recognition on the infrared lock-in grayscale image to obtain an infrared lock-in grayscale image with boundary lines;

[0012] S3: Perform image segmentation on the infrared lock-in grayscale image with boundary lines with a gray threshold as a constraint condition to obtain an infrared lock-in grayscale image that divides the gray region;

[0013] S4: Perform shape recognition on the gray region greater than the gray threshold to obtain a region containing singular points and / or edge curves; based on the singular points and / or the edge curves, obtain the defect type and the total area of the defect region of the prepreg laminate to be detected.

[0014] In a preferred embodiment of the present invention, in step S1,

[0015] The frequency of the infrared excitation heat source is as shown in formula (1)

[0016] f = f1 + f2T formula (1)

[0017] In formula (1), f is the frequency of the infrared excitation heat source, in Hz; f1 is the initial frequency of the infrared excitation heat source, in Hz; f2 is the time coefficient of the infrared excitation heat source, in Hz / s; T is the excitation time of the variable-frequency thermal excitation, in s.

[0018] In a preferred embodiment of the present invention, in step S1,

[0019] The value range of f1 is 0.1 to 1000 Hz, preferably 0.1 to 100 Hz; and / or,

[0020] The value range of f2 is 0.1 to 100 Hz / s, preferably 0.1 to 20 Hz / s; and / or,

[0021] The value range of T is 0 to 30 s, preferably 0 to 10 s.

[0022] In a preferred embodiment of the present invention, in step S2,

[0023] The infrared lock-in thermography is converted into an infrared lock-in grayscale image by using the weighted average method; and / or,

[0024] The boundary of the infrared lock-in grayscale image is recognized by using the wavelet analysis method.

[0025] In a preferred embodiment of the present invention, in step S4,

[0026] The edge detection algorithm is used to perform shape recognition on the gray area greater than the gray threshold.

[0027] In a preferred embodiment of the present invention, in step S4,

[0028] If the area greater than the gray threshold is a continuously distributed non-zero value and the size is within a preset range, it is recognized as a bubble defect, and the area of the bubble defect is obtained based on the area of the singular point; and / or,

[0029] If the area greater than the gray threshold forms a connected region of gray features, it is determined whether the edge curve contains the boundary line of the prepreg ply to be detected; if the edge curve does not contain the boundary line of the prepreg ply to be detected, it is recognized as a local debonding defect, and the area of the connected region is the area of the local debonding defect; and / or, if the edge curve contains the boundary line of the prepreg ply to be detected, it is recognized as a delamination defect, and the area of the connected region is the area of the local debonding defect.

[0030] In a preferred embodiment of the present invention, the defect detection method for the prepreg ply further includes the following steps:

[0031] S5: Obtain the total area of the prepreg ply to be detected based on the infrared lock-in gray scale image containing the boundary line; based on the total area of the defect area in the prepreg ply to be detected and the total area of the prepreg ply to be detected, obtain the product evaluation result.

[0032] In a preferred embodiment of the present invention, the defect detection method for the prepreg ply further includes the following steps:

[0033] S6: Adjust the infrared lock-in gray scale image containing the boundary line based on the region containing the singular point and / or the edge curve obtained in step S4; the adjustment of the infrared lock-in gray scale image containing the boundary line includes at least one of increasing the contrast, enhancing the gray value, and reducing the gray threshold.

[0034] The second object of the present invention is to provide a defect detection system for a prepreg laminate, including a variable-frequency thermal excitation and acquisition module, a grayscale conversion and boundary recognition module, a grayscale region segmentation module, a shape recognition module, and a defect type judgment and calculation module that are communicatively connected in sequence; the variable-frequency thermal excitation and acquisition module is used to perform variable-frequency thermal excitation on the prepreg laminate to be detected and collect the infrared lock-in thermography of the prepreg laminate to be detected; the grayscale conversion and boundary recognition module is used to convert the infrared lock-in thermography into an infrared lock-in grayscale image, perform boundary recognition on the infrared lock-in grayscale image, and obtain an infrared lock-in grayscale image with boundary lines; the grayscale region segmentation module is used to perform image segmentation on the infrared lock-in grayscale image with boundary lines with a grayscale threshold as a constraint condition to obtain an infrared lock-in grayscale image with divided grayscale regions; the shape recognition module is used to perform shape recognition on the grayscale regions greater than the grayscale threshold to obtain regions containing singular points and / or edge curves; the defect type judgment and calculation module is used to obtain the defect type and the total area of the defect region of the prepreg laminate to be detected based on the singular points and / or the edge curves.

[0035] The third object of the present invention is to provide a computer-readable storage medium, which stores at least one computer-executable program. When the at least one program is executed by a computer, the computer is caused to execute the steps in the defect detection method for a prepreg laminate as described in the first object of the present invention.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0037] 1. In the defect detection method for a prepreg laminate of the present invention, before collecting the infrared lock-in thermography of the prepreg laminate, the prepreg laminate is pre-excited by a variable-frequency infrared excitation source, and the frequency of the variable-frequency infrared excitation source increases linearly, so that the process of the prepreg laminate being thermally excited is different. Therefore, under this thermal wave excitation, the infrared signals radiated by defects on the near surface and relatively deep layers can both be detected by an external detector (that is, defect information at different depths can be excited), making the detection depth of this defect detection method better, solving the problem in the prior art that the infrared lock-in method is more suitable for detecting surface or near-surface defects and difficult to detect deep defects, and finally the detection result obtained by this defect detection method is more accurate.

[0038] 2. In the defect detection method for a prepreg laminate of the present invention, the grayscale change method is adopted to segment the defect image, and the method of image recognition is used to analyze the regions with obvious defects; for regions where the grayscale change is not restricted, a wavelet algorithm for regional grayscale integration is performed, thereby improving the defect recognition ability. Description of the Drawings

[0039] Figure 1 It is a schematic diagram of the defect detection method for a prepreg laminate of the present invention;

[0040] Figure 2 Schematic diagram of the defect detection method for the prepreg laminate of the present invention;

[0041] Figure 3 Frequency schematic diagram of the signal source after linear modulation of the present invention;

[0042] Figure 4 Schematic diagram for identifying the shape of bubble defects;

[0043] Figure 5 Schematic diagram for identifying the shape of debonding defects;

[0044] Figure 6 Schematic diagram for identifying the shape of delamination defects. Detailed implementation manners

[0045] The present invention will be further described in detail below with reference to the accompanying drawings:

[0046] Example 1

[0047] As Figure 1 and Figure 2 shown, the present invention provides a defect detection method for a prepreg laminate, including the following steps

[0048] Step S1: Perform variable-frequency thermal excitation on the prepreg laminate to be detected and collect the infrared lock-in thermogram of the prepreg laminate to be detected.

[0049] The present invention uses an infrared thermal imager equipped with a variable-frequency infrared excitation source to collect the infrared lock-in thermogram of the prepreg laminate to be detected. A variable-frequency infrared excitation source is configured around the infrared detection camera to achieve variable-frequency control of infrared light waves. First, an infrared detection camera is arranged on the prepreg production line, and the infrared detection camera and the prepreg are vertically distributed. Specifically, the infrared detection camera is located above the prepreg and is connected to the guide rail, and its movement on the guide rail can be controlled. The fixed frame of the infrared detection camera needs to have a certain rigidity to avoid significant changes in the images collected during the movement of the infrared detection camera. Preferably, the infrared camera is an electrically cooled infrared camera; and / or, the infrared camera uses a near-focus lens; and / or, the focal length range of the infrared camera is 0.2 to 2 meters.

[0050] In a preferred implementation manner of the present invention, the frequency range of the infrared excitation heat source is configured according to formula (1) to be applicable to the internal resin and fibers of prepregs with different heat conduction capabilities.

[0051] f = f1 + f2T Formula (1)

[0052] In Equation (1), f is the frequency of the infrared excitation heat source, with the unit of Hz; f1 is the initial frequency of the infrared excitation heat source, with the unit of Hz; f2 is the time coefficient of the infrared excitation heat source, with the unit of Hz / s; T is the excitation time of the variable-frequency thermal excitation, with the unit of s. Preferably, the value range of f1 is 0.1 - 1000 Hz, preferably 0.1 - 100 Hz; and / or, the value range of f2 is 0.1 - 100 Hz / s, preferably 0.1 - 20 Hz / s; and / or, the value range of T is 0 - 30 s, preferably 0 - 10 s, most preferably 3 - 10 s; and / or, the frequency f of the infrared excitation heat source ranges between 10 - 1000 Hz. It should be noted that for a relatively thick prepreg layer, the time coefficient f2 of the infrared excitation heat source can be increased.

[0053] In the stage of collecting the infrared lock-in thermography of the prepreg layer to be detected, the frequency of the infrared excitation source increases linearly, so as to resolve more abundant information. When using a linearly modulated infrared excitation heat source to replace the infrared excitation heat source with a fixed frequency of ω to perform thermal excitation on the prepreg layer to be detected, the frequency of the signal source after linear modulation increases linearly, as Figure 3 shown. Assuming that the thermal excitation duration of the signal source after linear modulation is r and the signal broadening of the signal source after linear modulation is B, the signal of the thermal wave excitation on the surface of the prepreg layer to be detected is as shown in Equation (2):

[0054]

[0055] In Equation (2), Q(x = 0, t) is the surface temperature of the prepreg layer to be detected (at the position of x = 0 and at the moment of t), with the unit of °C; Q0 is the initial surface temperature of the prepreg layer to be detected, with the unit of °C; j is the imaginary unit (j 2 = -1); f is the frequency of the infrared excitation heat source, with the unit of Hz; t is the time, with the unit of seconds; B is the signal broadening of the signal source after linear modulation, with the unit of seconds; r is the thermal excitation duration of the signal source after linear modulation, with the unit of seconds.

[0056] Taking the excitation thermal wave of Equation (2) as the signal source instead of the thermal wave with a fixed frequency of ω, the thermal distribution function on the semi-infinite plate is as shown in Equations (3) - (5):

[0057]

[0058]

[0059]

[0060] In Equations (3) - (5), T(x, t) is the distribution function of temperature with respect to distance and time (at the position of x and at the moment of t); T0 is the ambient temperature, with the unit of °C; a is the thermal diffusivity, with the unit of m2 / s; k is the phase angle of temperature fluctuation, with the unit of rad.

[0061] Therefore, in the excited thermal wave excitation state of Equation (2), the thermal diffusion length of the prepreg ply to be detected is as shown in Equation (6):

[0062]

[0063] In Equation (6), μ fm is the thermal diffusion length, with the unit of m. It should be noted that B / r in Equation (6) is actually the modal modulation factor, which follows the linear modulation characteristic and is an important parameter for the linear change of the control frequency with time. After establishing the correlation coefficient between the thermal diffusion length and the frequency modal adjustment, it helps the detection system to detect the defects deep in the material. If the parameter B / r is zero, the above equation degenerates to the steady-state diffusion form with a fixed frequency of ω.

[0064] When the prepreg ply to be detected is a homogeneous defect-free specimen, its radiation thermal wavelength under the aforementioned modulated modal thermal excitation is λ, and the expression of the radiation thermal wavelength is as shown in Equation (7):

[0065]

[0066] It can be seen from Equation (7) that with the linear increase of the frequency of the infrared excitation source, the deeper defects will also have a more significant phase difference, thus being easily captured by the highly sensitive infrared camera.

[0067] Step S2: Convert the infrared lock-in diagram into an infrared lock-in grayscale diagram, and perform wavelet analysis on the infrared lock-in grayscale diagram to obtain an infrared lock-in grayscale diagram with boundary lines.

[0068] This step is the preliminary processing of the infrared lock-in diagram, specifically including the grayscale change processing and boundary recognition processing of the infrared lock-in diagram, so as to provide boundary conditions for subsequent processing. Exemplarily, the conversion of the grayscale diagram can be achieved by the weighted average method, that is, taking the weighted sum of the three-channel color components as the grayscale value, that is

[0069] Grayscale = 0.299×R + 0.587×G + 0.114×B Equation (8)

[0070] Using the method of Equation (8) to convert the infrared lock-in diagram, an infrared lock-in grayscale diagram is obtained. It should be noted that the grayscale value of the pixel points in the infrared lock-in grayscale diagram is related to the temperature of the pixel points in the infrared lock-in diagram collected in Step S1, the temperature is related to the frequency of the variable-frequency infrared excitation source, and the frequency of the variable-frequency infrared excitation source is related to the detection depth. Therefore, the change of the frequency of the variable-frequency infrared excitation source will cause the change of the temperature, and the change speed of the temperature reflects the superimposed signals of different detection depths. After processing, the detection signals of different depths can be obtained.

[0071] The wavelet analysis method is to perform wavelet decomposition on the infrared lock-in grayscale image and extract the high-frequency coefficients, and calculate the gradient of the high-frequency coefficients, so as to obtain the change rate of the grayscale values of adjacent pixel points and the fluctuation range of the change rate in a specific direction, and accordingly determine the boundary of the infrared image (the boundary usually corresponds to the area with a relatively high change rate of grayscale values). The following is an exemplary description of the "area with a relatively high change rate of grayscale values". The grayscale within the detection range is normalized, that is, the maximum range Gray of grayscale fluctuation is obtained. For the case where the local change rate reaches more than 50% Gray, especially the area greater than 70%, it is determined as the boundary area. It should be noted that performing wavelet analysis on the image for boundary recognition belongs to a conventional method in the art and will not be elaborated here.

[0072] Step S3: Perform image segmentation on the infrared lock-in grayscale image containing boundary lines with the grayscale threshold as the constraint condition to obtain an infrared lock-in grayscale image that divides the grayscale regions (i.e., a binary image); wherein, the infrared lock-in grayscale image of the grayscale regions includes a grayscale region greater than the grayscale threshold and a grayscale region not greater than the grayscale threshold.

[0073] This step is a further processing of the infrared lock-in image. First, determine the grayscale threshold according to the characteristics of the overall grayscale value of the image, and then use this grayscale threshold as the constraint condition to divide the grayscale regions, and finally obtain an infrared lock-in grayscale image that divides the grayscale regions. Those skilled in the art know that for prepreg laminates containing defects of different depths, the closer the defect is to the surface of the prepreg laminate, the more obvious the grayscale value of its edge during detection, while for defects relatively deep in the prepreg laminate, the grayscale value of its edge is not significant during detection. When determining the grayscale threshold, a prepreg laminate with a thickness of about 3 - 5 mm can be made, and defects of different depths (hollow rings or other inserts / films) are embedded during the manufacturing process. Use the method of the present invention to find defects with a preset fixed shape (preset shapes such as rings, film shapes), extract the edge grayscale values of these defects in the infrared lock-in grayscale image, form the edge grayscale value range of the defects (including all depths), and extract the grayscale value with a 75 - 85% confidence level in this range as the grayscale threshold. This is the so-called "determine the grayscale threshold according to the characteristics of the overall grayscale value of the image".

[0074] For the area greater than the grayscale threshold, it is considered as the area with defects. For the area not greater than the grayscale threshold, generally its temperature change is not significant, indicating that there is an actual physical connection inside and it is not a defect area; however, the area not greater than the grayscale threshold may also be system noise. If the product quality requirements for the prepreg laminate are relatively high, the area not greater than the grayscale threshold can be further analyzed. Since the present invention performs variable-frequency thermal excitation on the prepreg laminate to be detected, image differencing with different time lengths can be performed. If the area not greater than the grayscale threshold exists in different differenced images at a fixed position, it is presumed that the area not greater than the grayscale threshold is a tiny defect that has not been recognized, further improving the accuracy of the defect detection method.

[0075] Step S4: Perform shape recognition on the grayscale area greater than the grayscale threshold to obtain an area containing singular points and / or edge curves; based on the singular points and / or edge curves, obtain the defect type and the total area of the defect area of the prepreg laminate to be detected. It should be noted that "performing shape recognition on the grayscale area greater than the grayscale threshold" mainly uses a conventional edge detection algorithm to obtain a continuous boundary, such as the Canny algorithm in Patent CN117092163A. This shape recognition method will not be elaborated here. If some areas are discontinuous, interpolation calculation can be performed on the area greater than the grayscale threshold to obtain a continuous boundary. Preferably, perform area integration on the coordinate sequence of the edge curve to obtain the area of the region within the edge curve. Specifically in this embodiment, the area of the region within the edge curve also represents the number of pixel points within the edge curve.

[0076] The following separately explains the "singular point" and the "edge curve". A singular point refers to a closed small area, not a point, but its structural size is below a certain range, and the grayscale values of all pixel points within the closed small area are higher than the grayscale threshold. At least 70% of the pixel points within the continuous boundary of the "edge curve" have grayscale values higher than the grayscale threshold; the grayscale values of the pixel points outside the edge curve are not higher than the grayscale threshold, and the grayscale values of the pixel points on the edge curve are higher than the grayscale threshold. This step includes three cases. The first case is that when performing shape recognition on the grayscale area greater than the grayscale threshold, only an area containing singular points is obtained. At this time, based on the singular points, the total area of the defect area in the prepreg laminate to be detected can be obtained. The second case is that when performing shape recognition on the grayscale area greater than the grayscale threshold, only an area containing edge curves is obtained. At this time, performing area integration on the coordinate sequence of the edge curve can obtain the total area of the defect area in the prepreg laminate to be detected. The third case is that when performing shape recognition on the grayscale area greater than the grayscale threshold, an area containing both singular points and / or edge curves is obtained. At this time, performing area integration on the coordinate sequence of the edge curve and combining it with the area of the singular points can obtain the total area of the defect area in the prepreg laminate to be detected.

[0077] The present invention relates to three typical defects, namely bubbles, local debonding, and delamination. In regions containing singular points and / or edge curves, different defect types have different mathematical characteristics. Exemplarily, as Figure 4 shown, for suspected bubble defects, during the process of shape recognition in the region greater than the gray threshold, the "region greater than the gray threshold" is a continuously distributed non-zero value (the "non-zero value" refers to the gray value), and at the same time, the structural size of the continuously distributed non-zero value is below a certain range (such as below 1 square centimeter), then it can be recognized as a bubble defect. Therefore, if the "region greater than the gray threshold" is a continuously distributed non-zero value and the size of the "region greater than the gray threshold" is within the preset range, the bubble defect type can be output. If the pixels of the infrared lock-in image are relatively high, the area of the region can be basically locked through the pixel singular region of the recognized bubble defect. Since the pixels of the infrared lock-in image in this embodiment are relatively high, therefore, the area of the pixel singular region of the recognized bubble defect is the area of the bubble defect. It has been mentioned above that "a singular point refers to a closed small region", because the area integral can be performed on the coordinate sequence of this closed small region to obtain the area of the bubble defect, which is the Figure 2 solution of the singular point in

[0078] As Figure 5 shown, for suspected local debonding defects, during the process of shape recognition in the region greater than the gray threshold, the connected region that just constitutes the gray feature, that is, the region composed of pixel points with the same gray feature and adjacent positions, is not interrupted by pixel points with other different gray features. As Figure 6 shown, for suspected local delamination defects, during the process of shape recognition in the region greater than the gray threshold, it includes a first edge curve and a second edge curve, and the first edge curve and the second edge curve together constitute a connected region of gray features; wherein, the first edge curve belongs to a part of the boundary line in step S2. If the "region greater than the gray threshold" constitutes a connected region of gray features, it is further determined whether the "edge curve" contains the boundary line of the prepreg ply to be detected; if the edge curve does not contain the boundary line of the prepreg ply to be detected, it is recognized as a local debonding defect; if the edge curve contains the boundary line of the prepreg ply to be detected, it is recognized as a delamination defect. Whether the edge curve contains the boundary line of the prepreg ply to be detected can be specifically determined by the coordinates of the edge curve and the coordinates of the boundary line.

[0079] In a preferred embodiment of the present invention, the area integral is performed on the coordinate sequence of the edge curve according to formulas (9) and (10)

[0080]

[0081] p(x,y) = xy

[0082] Q(x, y) = x 2 + y 2 Equation (10)

[0083] In Equation (2), L is the edge curve of the closed region D; x and y are the coordinates of the pixel points on the edge curve.

[0084] If the points satisfying the positions of the pixel points on the edge curve L form a grayscale image, it is a local debonding defect. In other words, the coordinates of the corresponding pixel points are the position (x, y) values of the L curve, which are the positions where the debonding defect exists. By calculating according to Equation (9) and Equation (10), the area of the local debonding defect can be obtained.

[0085] In a preferred embodiment of the present invention, the edge curve includes a first edge curve and a second edge curve, and area integration is performed on the coordinate sequences of the first edge curve and the second edge curve according to Equation (11) and Equation (12)

[0086]

[0087] p(x, y) = xy

[0088] Q(x, y) = x + y Equation (12)

[0089] In Equation (11), L1 is the first edge curve of the closed region D (which is a part of the boundary line in step S2); L2 is the second edge curve of the closed region D; x and y are the coordinates of the pixel points on the first edge curve and the second edge curve.

[0090] It should be noted that Equation (11) is for verification and solution, and the integral solution of area and length is not carried out in this loop. Therefore, for the convenience of solving partial derivatives, the function equation is set as shown in Equation (12).

[0091] As can be seen from the above, defect recognition is carried out based on the characteristics of the "region greater than the grayscale threshold". If the "region greater than the grayscale threshold" is a continuously distributed non-zero value and the size of the "region greater than the grayscale threshold" is within a preset range, it is recognized as a bubble defect, and the area of the bubble defect is obtained based on the area of the pixel singular region. If the "region greater than the grayscale threshold" constitutes a connected region of grayscale features, it is further determined whether the "edge curve" contains the boundary line of the prepreg ply to be detected; if the edge curve does not contain the boundary line of the prepreg ply to be detected, it is recognized as a local debonding defect, and the area of the local debonding defect is obtained by calling Equation (9) and Equation (10); if the edge curve contains the boundary line of the prepreg ply to be detected, it is recognized as a delamination defect, and the area of the delamination defect is obtained by calling Equation (11) and Equation (12).

[0092] In a preferred embodiment of the present invention, the method for detecting defects in the prepreg ply further includes the following steps:

[0093] Step S5: Obtain the total area of the prepreg ply to be detected based on the infrared lock-in gray-scale image with boundary lines; based on the total area of the defect area in the prepreg ply to be detected and the total area of the prepreg ply to be detected, obtain the product evaluation result. It should be noted that the product area also represents the total number of pixel points within this area. In the aforementioned step S2, the infrared lock-in gray-scale image with boundary lines has been obtained. Therefore, the total number of pixel points within this boundary line is the product area.

[0094] In a preferred embodiment of the present invention, the method for detecting defects in the prepreg ply further includes the following steps:

[0095] S6: Based on the region containing singular points and / or edge curves obtained in step S4, adjust the infrared lock-in gray-scale image with boundary lines. After the adjustment is completed, perform step S3 and step S4 again. Preferably, the adjustment method includes at least one of increasing the contrast, enhancing the gray-scale value, and reducing the gray-scale threshold. Among them, increasing the contrast refers to adjusting the difference degree between regions with different gray-scale values in the "infrared lock-in gray-scale image with boundary lines" to make the difference in gray-scale values more obvious; enhancing the gray-scale value refers to enhancing the gray-scale value of each pixel point in the "infrared lock-in gray-scale image with boundary lines". Specifically, for the region greater than the gray-scale threshold (i.e., the region considered to have defects), increase the contrast and / or enhance the gray-scale value and / or reduce the gray-scale threshold, and re-segment the infrared lock-in gray-scale image with boundary lines to obtain an infrared lock-in gray-scale image with divided gray-scale regions. This can further improve the accuracy of identifying typical defect types. Exemplarily, the histogram equalization method can be used to increase the contrast of the region greater than the gray-scale threshold, and the gray-scale transformation can be used to enhance the gray-scale value of the region greater than the gray-scale threshold; among them, the two methods are conventional methods for increasing the contrast and enhancing the gray-scale value, which will not be elaborated here.

[0096] In practical applications, in some cases, when segmenting the infrared lock-in gray-scale image with boundary lines with the gray-scale threshold as the constraint condition, the gray-scale values of some regions in the normal adhesion state are greater than the gray-scale threshold. This may be due to slightly different conditions of the glue coating process and slightly different thicknesses of the glue coating layer. In other cases, when segmenting the infrared lock-in gray-scale image with boundary lines with the gray-scale threshold as the constraint condition, the change in the resulting gray-scale value is not significant. However, this does not mean that there are no defects. There may actually still be weak defects. In other words, in this case, step S3 involves the situation of missed detection, and the missed detection is caused by inappropriate threshold setting. Therefore, the methods of increasing the contrast and reducing the gray-scale threshold can be adopted.

[0097] Exemplarily, before performing steps S4 and S5, contrast enhancement can be first performed on the "region not greater than the gray threshold", and after image segmentation is completed, steps S4 and S5 are then performed. If, after contrast enhancement, it is still found in step S4 that there are connected disconnected regions (that is, pixel points belonging to the same object are split into two or more parts due to image segmentation, and these parts are spatially discontinuous), at this time, the gray threshold is lowered, and the infrared lock-in gray-scale image with boundary lines is segmented based on the lowered gray threshold as a constraint condition. After image segmentation is completed, step S4 is performed again.

[0098] Embodiment 2

[0099] As Figure 2 shown, the present invention provides a defect detection for a prepreg layup, including an acquisition module, a gray-scale conversion module, a boundary division module, a gray-scale region segmentation module, a shape recognition module, a calculation module, and a judgment module. The following will elaborate on these four modules in detail.

[0100] A variable-frequency thermal excitation and acquisition module, configured to perform variable-frequency thermal excitation on the prepreg layup to be detected and collect the infrared lock-in image of the prepreg layup to be detected. The specific method is as shown in step S1 in Embodiment 1, and will not be elaborated here.

[0101] A gray-scale conversion and boundary recognition module, communicatively connected to the variable-frequency thermal excitation and acquisition module, configured to convert the infrared lock-in image into an infrared lock-in gray-scale image, perform boundary recognition on the infrared lock-in gray-scale image, and obtain an infrared lock-in gray-scale image with boundary lines. The specific method is as shown in step S2 in Embodiment 1, and will not be elaborated here.

[0102] A gray-scale region segmentation module, communicatively connected to the gray-scale conversion and boundary recognition module, configured to perform image segmentation on the infrared lock-in gray-scale image with boundary lines based on a gray threshold as a constraint condition, and obtain an infrared lock-in gray-scale image with divided gray-scale regions. The specific method is as shown in step S3 in Embodiment 1, and will not be elaborated here.

[0103] A shape recognition module, communicatively connected to the gray-scale region segmentation module, configured to perform shape recognition on the gray-scale regions greater than the gray threshold, and obtain a region containing singular points and / or edge curves. The specific method is as shown in step S4 in Embodiment 1, and will not be elaborated here.

[0104] A defect type judgment and calculation module, communicatively connected to the gray-scale region segmentation module, configured to obtain the defect type and the total area of the defect region of the prepreg layup to be detected based on the singular points and / or the edge curves. The specific method is as shown in step S4 in Embodiment 1, and will not be elaborated here.

[0105] In the description of the present invention, it should be noted that, unless otherwise clearly defined and limited, the terms "connected" and "coupled" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0106] In the description of the present invention, unless otherwise specified, the orientation or positional relationship indicated by the terms "upper", "lower", "left", "right", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention.

[0107] The above technical solutions are only one implementation manner of the present invention. For those skilled in the art, based on the disclosed principles of the present invention, various types of improvements or deformations can be easily made, not limited to the technical solutions described in the above specific embodiments of the present invention. Therefore, the foregoing description is only preferred and does not have a limiting meaning.

Claims

1. A method for detecting defects in a prepreg layup, characterized in that, including S1: Apply variable-frequency thermal excitation to the prepreg ply to be detected and collect the infrared lock-in thermography of the prepreg ply to be detected; S2: Convert the infrared lock-in thermography into an infrared lock-in grayscale image, perform boundary recognition on the infrared lock-in grayscale image to obtain an infrared lock-in grayscale image with boundary lines; S3: Perform image segmentation on the infrared lock-in grayscale image with boundary lines with a grayscale threshold as a constraint condition to obtain an infrared lock-in grayscale image that divides grayscale regions; S4: Perform shape recognition on the grayscale regions greater than the grayscale threshold to obtain regions containing singular points and / or edge curves; based on the singular points and / or the edge curves, obtain the defect type and the total area of the defect region of the prepreg ply to be detected.

2. The defect detection method for the prepreg ply according to claim 1, wherein, In step S1, the frequency of the infrared excitation heat source is as shown in formula (1) f = f1 + f2T formula (1) In formula (1), f is the frequency of the infrared excitation heat source, with the unit of Hz; f1 is the initial frequency of the infrared excitation heat source, with the unit of Hz; f2 is the time coefficient of the infrared excitation heat source, with the unit of Hz / s; T is the excitation time of the variable-frequency thermal excitation, with the unit of s.

3. The defect detection method for the prepreg ply according to claim 2, wherein In step S1, the value range of f1 is 0.1 - 1000 Hz, preferably 0.1 - 100 Hz; and / or, the value range of f2 is 0.1 - 100 Hz / s, preferably 0.1 - 20 Hz / s; and / or, the value range of T is 0 - 30 s, preferably 0 - 10 s.

4. The defect detection method for the prepreg ply according to claim 1, characterized in that, In step S2, use the weighted average method to convert the infrared lock-in thermography into an infrared lock-in grayscale image; and / or, use the wavelet analysis method to perform boundary recognition on the infrared lock-in grayscale image.

5. The defect detection method for the prepreg ply according to claim 1, characterized in that, In step S4, use the edge detection algorithm to perform shape recognition on the grayscale regions greater than the grayscale threshold.

6. The defect detection method for a prepreg ply according to claim 1, wherein, In step S4, if the regions greater than the grayscale threshold are continuously distributed non-zero values and the size is within the preset range, it is recognized as a bubble defect, and the area of the bubble defect is obtained based on the area of the singular point; and / or, if the regions greater than the grayscale threshold form a connected region of grayscale features, then determine whether the edge curve contains the boundary lines of the prepreg ply to be detected; if the edge curve does not contain the boundary lines of the prepreg ply to be detected, it is recognized as a local debonding defect, and the area of the connected region is the area of the local debonding defect; and / or, if the edge curve contains the boundary lines of the prepreg ply to be detected, it is recognized as a delamination defect, and the area of the connected region is the area of the local debonding defect.

7. The method for detecting defects in a prepreg ply according to claim 1, wherein, The defect detection method for the prepreg ply further includes the following steps: S5: Obtain the total area of the prepreg ply to be detected based on the infrared lock-in grayscale image with boundary lines; based on the total area of the defect region in the prepreg ply to be detected and the total area of the prepreg ply to be detected, obtain the product evaluation result.

8. The method for detecting defects in a prepreg ply according to claim 1, characterized in that The defect detection method for the prepreg ply further includes the following steps: S6: Adjust the infrared lock-in grayscale image with boundary lines based on the region containing singular points and / or edge curves obtained in step S4; the adjustment of the infrared lock-in grayscale image with boundary lines includes at least one of increasing the contrast, enhancing the grayscale value, and reducing the grayscale threshold.

9. A defect detection system for a prepreg laminate, characterized in that, It includes a variable-frequency thermal excitation and acquisition module, a grayscale conversion and boundary recognition module, a grayscale region segmentation module, a shape recognition module, and a defect type judgment and calculation module that are communicatively connected in sequence; the variable-frequency thermal excitation and acquisition module is used to perform variable-frequency thermal excitation on the pre-preg layer to be detected and collect the infrared lock-in thermography of the pre-preg layer to be detected; the grayscale conversion and boundary recognition module is used to convert the infrared lock-in thermography into an infrared lock-in grayscale image, perform boundary recognition on the infrared lock-in grayscale image, and obtain an infrared lock-in grayscale image with boundary lines; the grayscale region segmentation module is used to perform image segmentation on the infrared lock-in grayscale image with boundary lines with a grayscale threshold as a constraint condition to obtain an infrared lock-in grayscale image with divided grayscale regions; the shape recognition module is used to perform shape recognition on the grayscale regions greater than the grayscale threshold to obtain regions containing singular points and / or edge curves; The defect type judgment and calculation module is used to obtain the defect type and the total area of the defect region of the pre-preg layer to be detected based on the singular points and / or the edge curves.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer-executable program, and when the at least one program is executed by the computer, the computer executes the steps in the defect detection method of the pre-preg layer according to any one of claims 1 to 8.

Images