Detection method, system and equipment for different kinds of defects of carbon fiber skin and storage medium

Abstract

The invention relates to the technical field of composite material defect detection, and particularly discloses a method, a system and equipment for detecting different kinds of defects of a carbon fiber skin and a storage medium, and the method comprises the following steps: obtaining an infrared image sequence of a defect-containing carbon fiber skin; processing the infrared image sequence by using a pulse phase method to obtain a feature image; processing the feature image based on an improved Canny operator to obtain an edge image, and further obtaining a defect size parameter; and the defect depth is obtained through the BP neural network. The method can provide feature images accurately reflecting form information of different types of defects of the skin, obtains accurate sizes and depths of different types of defects, overcomes the defects of low efficiency, low precision, high cost and the like of a traditional detection method, facilitates subsequent accurate maintenance of the defect-containing skin, enables the skin to maintain good performance, and improves the detection accuracy of the skin. And a more accurate tool can be provided for design, evaluation and optimization of an aircraft structure.

Description

Technical Field

[0001] The present invention relates to the technical field of composite material defect detection, and in particular to a method, system, equipment and storage medium for detecting different types of defects in carbon fiber skins. Background Art

[0002] During long-term flight, key components such as the fuselage and wings will develop fatigue cracks due to repeated stress. These cracks may lead to structural failure, and in severe cases, damage to the fuselage or even disintegration in the air. Defects in the material itself, such as inclusions and unevenness, will affect its performance and durability and may cause breakage at critical moments. Different types of defects have different impacts on flight safety. Compared with previous technologies, this technology combines the pulse phase method, the improved Canny operator and the BP neural network to quantitatively detect different types of defects inside the carbon fiber skin, which has the advantages of shorter detection time and higher accuracy. Summary of the invention

[0003] In order to solve the problems existing in the prior art, the present invention provides a method, system, equipment and storage medium for detecting different types of defects in carbon fiber skins, which can efficiently and accurately detect the size and depth of different types of defects in carbon fiber skins, facilitate the subsequent precise repair of defective skins to maintain good performance, and solve the problems mentioned in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for detecting different types of defects in carbon fiber skins, comprising the following steps:

[0005] S1, obtaining an infrared image sequence of a defective carbon fiber skin;

[0006] S2, using a pulse phase method to process the infrared image sequence to obtain a characteristic image;

[0007] S3, processing the feature image based on the improved Canny operator to obtain the edge image, and then obtain the defect size parameter;

[0008] S4. Obtain the defect depth through BP neural network.

[0009] Preferably, in step S1, the defect types include cracks, pores, delamination, looseness, inclusions, ply errors, fiber buckling and misalignment;

[0010] The specific process of obtaining the infrared image sequence is as follows: an instantaneous high-energy heat pulse is loaded onto the defective carbon fiber skin, and part of the heat will be absorbed by the skin and transferred to its interior. Due to the defects inside the skin, the materials and structures of the defective and non-defective areas are different, which will change the heat transfer process and cause temperature differences on the corresponding skin surfaces. The temperature image of the skin surface is collected during the whole process to obtain an infrared image sequence.

[0011] Preferably, in step S2, the pulse phase method transforms the infrared image sequence into a phase image sequence using Fourier transform, and selects an image with the best defect presentation effect from the phase image sequence as the feature image.

[0012] In step S2, it specifically includes: the pulse phase method uses Fourier transform to transform the temperature change of each pixel of the infrared image sequence from the time domain to the frequency domain, and the formula is expressed as follows:

[0013]

[0014] Among them, F n is the transformation value of the nth frequency component, Δt is the sampling time interval, k is the infrared image number, N is the number of infrared images, T is the temperature, e is a natural constant, π is the circumference of a circle, i is an imaginary unit, Re n and Im n F n The real and imaginary parts of

[0015] Phase φ n Use the following formula to obtain:

[0016]

[0017] The image composed of the phase of each pixel at the same frequency is a phase map, and the phase maps corresponding to different frequencies constitute a phase map sequence; an image with the best defect presentation effect is selected from the phase map sequence as the feature image.

[0018] Preferably, in step S3, the improved Canny operator specifically includes the following aspects:

[0019] a. The improved Canny operator uses median filtering instead of Gaussian filtering;

[0020] b. Improved the calculation method of gradient amplitude;

[0021] c. Use the maximum inter-class variance method to achieve adaptive threshold segmentation.

[0022] Preferably, in a, a median filter is used instead of a Gaussian filter, and the median filter sets the gray value of each pixel in the feature image to the median of the gray values ​​of all pixels in a certain neighborhood window of the point, thereby effectively removing noise while retaining edge and detail information;

[0023] In b, the calculation method of the gradient amplitude is improved. On the basis of the traditional Canny operator only calculating the derivatives of the pixel points in the 0° and 90° directions, the derivatives in the 45° and 135° directions are added. The calculation formula is as follows:

[0024] G 45° (x,y)=f(x+1,y+1)-f(x-1,y-1)

[0025] G 135° (x,y)=f(x-1,y+1)-f(x+1,y-1)

[0026] Among them, G 45° Represents the derivative of the gray value of the pixel in the 45° direction, G 135° represents the derivative of the gray value of the pixel in the 135° direction, x represents the horizontal coordinate of the pixel, y represents the vertical coordinate of the pixel, and f represents the gray value of the pixel;

[0027] The gradient magnitude and direction are calculated as follows:

[0028]

[0029] Where M represents the gradient amplitude and θ represents the gradient direction;

[0030] The calculation method of gradient amplitude is improved to obtain more directional information, improve edge recognition accuracy, and reduce edge loss and misjudgment;

[0031] In the above c, the maximum inter-class variance method is used to realize adaptive threshold segmentation. The maximum inter-class variance method divides the image into two parts, background and target, by calculating the grayscale histogram of the image. The threshold when the inter-class variance of these two parts is the largest is the optimal threshold. The calculation formula of the inter-class variance is as follows:

[0032] g=ω 0 ω 1 (μ 0 -μ 1 ) 2

[0033] Among them, g is the between-class variance, ω 0 is the ratio of target pixels to total pixels, ω 1 is the ratio of background pixels to total pixels, μ 0 is the average gray value of the target pixel, μ 1is the average gray value of background pixels.

[0034] Preferably, the BP neural network includes an input layer, a hidden layer and an output layer; the input layer includes the temperature difference peak value, temperature difference peak time, major diameter, circumference and area of ​​the defect; the output layer is the defect depth; the temperature difference peak value is the maximum value of the temperature difference between the defective area and the non-defective area of ​​the skin after pulse excitation; the temperature difference peak time is the time when the temperature difference peak occurs; the formula of the output process in the BP neural network is expressed as follows:

[0035]

[0036] where x j is the output value of the hidden layer or output layer, m is the number of neurons in the input layer or hidden layer, ω ij is the weight from the i-th neuron to the j-th neuron, x i is the input value of the hidden layer or output layer, b j is the threshold of the jth neuron.

[0037] On the other hand, in order to achieve the above-mentioned purpose, the present invention also provides the following technical solution: a system for detecting different types of defects of carbon fiber skins: comprising the following modules:

[0038] Infrared image sequence acquisition module: obtain infrared image sequence of defective carbon fiber skin;

[0039] Feature image acquisition module: using pulse phase method to process the infrared image sequence to obtain feature images;

[0040] Defect size acquisition module: Based on the improved Canny operator, the feature image is processed to obtain the edge image, and then the defect size parameters are obtained;

[0041] Defect depth acquisition module: obtains defect depth through BP neural network.

[0042] On the other hand, to achieve the above-mentioned purpose, the present invention also provides the following technical solution: an electronic device, the electronic device comprising: a processor; and a memory for storing one or more programs;

[0043] When the one or more programs are executed by the processor, the processor executes the method for detecting different types of defects of carbon fiber skin.

[0044] On the other hand, to achieve the above-mentioned purpose, the present invention also provides the following technical solution: a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for detecting different types of defects in carbon fiber skins.

[0045] The beneficial effects of the present invention are as follows: the method of the present invention can provide characteristic images that accurately reflect the morphological information of different types of defects in the skin, obtain the accurate size and depth of different types of defects, overcome the shortcomings of traditional detection methods such as low efficiency, low accuracy, and high cost, and facilitate subsequent precise repair of defective skins to maintain good performance, and can also provide more accurate tools for the design, evaluation, and optimization of aircraft structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 Schematic diagram of the steps of the method for detecting different types of defects of carbon fiber skin in this embodiment;

[0047] Figure 2 This is a schematic diagram of a defective skin in this embodiment;

[0048] Figure 3 Schematic diagram of the detection system module for different types of defects of carbon fiber skin in this embodiment;

[0049] Figure 4 This is a schematic diagram of the structure of an electronic device in an embodiment of the present invention;

[0050] In the figure, 110 - infrared image sequence acquisition module; 120 - feature image acquisition module; 130 - defect size acquisition module; 140 - defect depth acquisition module; 210 - processor; 220 - storage. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] See also Figure 1 The embodiment of the present invention takes the carbon fiber skin of Xirui SR-20 as an example, and provides a technical solution: a method for detecting different types of defects of the carbon fiber skin of Xirui SR-20, which comprises the following steps:

[0053] Step S1: Obtain an infrared image sequence of the defective carbon fiber skin of the Xirui SR-20.

[0054] Furthermore, the schematic diagram of the defective skin is as follows Figure 2As shown, the largest rectangular block is the carbon fiber skin, and the other geometric bodies are defects. The defect types are delamination defects and crack defects. The length and width of the delamination defect are between 4-35mm and the height is 0.2mm. The length of the crack defect is between 18-50mm, the width is between 0.5-2mm, the height is 0.2mm, and the defect depth is between 0.3-1.5mm. The upper surface of the skin is heated by a flash lamp, and the infrared thermal imager collects the temperature image of the skin surface from before heating. The acquisition time is 200s to obtain an infrared image sequence.

[0055] Step S2: Processing the infrared image sequence using a pulse phase method to obtain a characteristic image.

[0056] In step S2, the pulse phase method uses Fourier transform to transform the temperature change of each pixel in the infrared image sequence from the time domain to the frequency domain:

[0057]

[0058] Among them, F n is the transformation value of the nth frequency component, Δt is the sampling time interval, k is the infrared image number, N is the number of infrared images, T is the temperature, e is a natural constant, π is the circumference of a circle, i is an imaginary unit, Re n and Im n F n The real and imaginary parts of . Phase φ n It can be obtained by the following formula:

[0059]

[0060] The image composed of the phase of each pixel at the same frequency is a phase map, and the phase maps corresponding to different frequencies constitute a phase map sequence. The defects of the phase map can be detected at a greater depth, are less affected by uneven heating, and have a higher signal-to-noise ratio and contrast than the original infrared image. From the phase map sequence, an image with the best defect presentation effect is selected as the feature image.

[0061] Step S3: Use the improved Canny operator to process the feature image to obtain the edge image, and then obtain the defect size.

[0062] Furthermore, the improved Canny operator first uses median filtering instead of Gaussian filtering. Since Gaussian filtering is a weighted average processing of the image, it will cause some detail information in the image to be blurred, especially when processing high-frequency information, it may make the edge and texture of the image blurred, which is not conducive to obtaining accurate edge information. The median filter sets the gray value of each pixel in the feature image to the median of the gray values ​​of all pixels in a certain neighborhood window of the point, which can maintain the distribution characteristics of the signal, effectively remove noise, and well protect the edge and detail information, so that the image still maintains a high degree of clarity after processing.

[0063] Secondly, the calculation method of the gradient amplitude is improved. On the basis of the traditional Canny operator which only calculates the derivatives of the pixel points in the 0° and 90° directions, the derivatives in the 45° and 135° directions are added. The calculation formula is as follows.

[0064] G 45° (x,y)=f(x+1,y+1)-f(x-1,y-1)

[0065] G 135° (x,y)=f(x-1,y+1)-f(x+1,y-1)

[0066] The gradient magnitude and direction are calculated as follows.

[0067]

[0068] Obtaining information from more directions can improve edge recognition accuracy, retain more edge information, ensure edge continuity, and reduce edge loss and misjudgment.

[0069] Finally, the maximum inter-class variance method is used to achieve adaptive threshold segmentation. The traditional Canny operator requires manual experience to select a threshold to binarize the image, which is not suitable for finding the most suitable threshold for image segmentation. The maximum inter-class variance method selects the threshold adaptively based on the characteristics of the image pixel value distribution to avoid the situation where the manual selection of the threshold is not effective. The maximum inter-class variance method divides the image into two parts, the background and the target, by calculating the grayscale histogram of the image. The threshold when the inter-class variance of these two parts is the largest is the optimal threshold. The calculation formula of the inter-class variance is as follows.

[0070] g=ω 0 ω 1 (μ 0 -μ 1 ) 2

[0071] Among them, g is the between-class variance, ω 0 is the ratio of target pixels to total pixels, ω 1 is the ratio of background pixels to total pixels, μ 0 is the average gray value of the target pixel, μ 1 is the average gray value of background pixels. The edge image is obtained by processing the feature image with the improved Canny operator, and the size parameters of the defect, such as the long diameter, perimeter, and area, can be calculated based on the edge image.

[0072] Step S4: Obtain the defect depth using BP neural network.

[0073] Furthermore, the temperature difference peak value, temperature difference peak time, long diameter, perimeter, and area of ​​the defect are used as the input of the BP neural network, and the defect depth is used as the output of the network. A sufficient number of the above parameters of different types of defects are collected to form a data set to train the network. The mean square error between the network output value and the actual value is used as the objective function, and the gradient descent method is used to calculate the minimum value of the objective function. When the objective function value meets the requirements, the training is stopped, and the network can output the depth of different types of defects more accurately. The relevant formulas of the BP neural network are as follows.

[0074]

[0075] where x j is the output value of the hidden layer or output layer, m is the number of neurons in the input layer or hidden layer, ω ij is the weight from the i-th neuron to the j-th neuron, x i is the input value of the hidden layer or output layer, b j is the threshold of the jth neuron.

[0076]

[0077] Where E is the mean square error, n is the number of data sets, t i is the actual depth of the i-th group, y i Output depth of the network for group i.

[0078]

[0079] Where Δω is the change in weight at each iteration, and η is the learning rate.

[0080]

[0081] Where Δb is the change in the threshold at each iteration.

[0082] An embodiment of the present invention provides a method for detecting different types of defects in the carbon fiber skin of a Xirui SR-20. The method can provide a characteristic image that accurately reflects the morphological information of different types of defects in the skin, and obtain the accurate size and depth of different types of defects. It overcomes the shortcomings of traditional detection methods such as low efficiency, low accuracy, and high cost, and is helpful for subsequent precise repair of defective skins to maintain good performance. It can also provide more accurate tools for the design, evaluation, and optimization of aircraft structures.

[0083] Based on the same inventive concept as the above method embodiment, the present application embodiment also provides a system for detecting different types of defects in carbon fiber skins, which can implement the functions provided by the above method embodiment, such as Figure 3 As shown, the system includes the following modules:

[0084] Infrared image sequence acquisition module 110: obtains an infrared image sequence of a defective carbon fiber skin;

[0085] Feature image acquisition module 120: processes the infrared image sequence using a pulse phase method to obtain a feature image;

[0086] Defect size acquisition module 130: Processing the feature image based on the improved Canny operator to obtain the edge image, and then obtain the defect size parameter;

[0087] Defect depth acquisition module 140: obtains defect depth through BP neural network.

[0088] Based on the same inventive concept as the above method embodiment, the present application embodiment also provides an electronic device, such as Figure 4 As shown, the device includes: a processor 210; and a memory 220, for storing one or more programs;

[0089] When the one or more programs are executed by the processor 210, the processor is enabled to execute the method for detecting different types of defects of the carbon fiber skin.

[0090] The method for detecting different types of defects of carbon fiber skins includes the following:

[0091] Obtain infrared image sequences of defective carbon fiber skins;

[0092] The infrared image sequence is processed by pulse phase method to obtain a characteristic image;

[0093] The edge image is obtained by processing the feature image based on the improved Canny operator, and then the defect size parameter is obtained;

[0094] The defect depth is obtained through BP neural network.

[0095] Based on the same inventive concept as the above method embodiment, the embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by the processor 210, the method for detecting different types of defects in the carbon fiber skin is implemented.

[0096] The method for detecting different types of defects of carbon fiber skins includes the following:

[0097] Obtain infrared image sequences of defective carbon fiber skins;

[0098] The infrared image sequence is processed by pulse phase method to obtain a characteristic image;

[0099] The edge image is obtained by processing the feature image based on the improved Canny operator, and then the defect size parameter is obtained;

[0100] The defect depth is obtained through BP neural network.

[0101] In several embodiments provided in the embodiments of the present invention, it should be understood that the disclosed apparatus and method can also be implemented in other ways. The apparatus and method embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the apparatus, method and computer program product according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0102] In addition, the functional modules in the various embodiments of the present invention may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0103] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention is essentially or partly contributed to the prior art or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, an electronic device, or a network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk. It should be noted that, in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also includes elements inherent to such a process, method, article or device. Without more constraints, an element defined by the phrase "comprising a..." does not exclude the existence of other identical elements in the process, method, article or apparatus comprising the element.

[0104] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "said" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0105] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0106] The word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0107] The "first\second" mentioned in the embodiments is only to distinguish similar objects, and does not represent a specific order for the objects. It is understandable that the "first\second" can be interchanged with the specific order or sequence where permitted. It should be understood that the objects distinguished by "first\second" can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0108] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting different types of defects in carbon fiber skins, characterized in that: The steps include: S1, obtaining an infrared image sequence of a defective carbon fiber skin; S2, using a pulse phase method to process the infrared image sequence to obtain a characteristic image; S3, processing the feature image based on the improved Canny operator to obtain the edge image, and then obtain the defect size parameter; S4. Obtain the defect depth through BP neural network.

2. The method for detecting different types of defects in carbon fiber skin according to claim 1, characterized in that: In step S1, the defect types include cracks, pores, delamination, looseness, inclusions, ply errors, fiber buckling, and misalignment; The specific process of obtaining the infrared image sequence is as follows: an instantaneous high-energy heat pulse is applied to the defective carbon fiber skin, and part of the heat will be absorbed by the skin and transferred to the inside. Due to the defects inside the skin, the materials and structures of the defective and non-defective areas are different, which will change the heat transfer process and cause temperature differences on the corresponding skin surfaces. The temperature image of the skin surface is collected during the whole process to obtain an infrared image sequence.

3. The method for detecting different types of defects in carbon fiber skin according to claim 1, characterized in that: In step S2, the pulse phase method transforms the infrared image sequence into a phase image sequence using Fourier transform, and selects an image with the best defect presentation effect from the phase image sequence as a feature image. In step S2, it specifically includes: the pulse phase method uses Fourier transform to transform the temperature change of each pixel of the infrared image sequence from the time domain to the frequency domain, and the formula is expressed as follows: Among them, F n is the transformation value of the nth frequency component, Δt is the sampling time interval, k is the infrared image number, N is the number of infrared images, T is the temperature, e is a natural constant, π is the circumference of a circle, i is an imaginary unit, Re n and Im n F n The real and imaginary parts of Phase Use the following formula to obtain: The image composed of the phase of each pixel at the same frequency is a phase map, and the phase maps corresponding to different frequencies constitute a phase map sequence; an image with the best defect presentation effect is selected from the phase map sequence as the feature image.

4. The method for detecting different types of defects in carbon fiber skin according to claim 1, characterized in that: In step S3, the improved Canny operator specifically includes the following aspects: a. The improved Canny operator uses median filtering instead of Gaussian filtering; b. Improved the calculation method of gradient amplitude; c. Use the maximum inter-class variance method to achieve adaptive threshold segmentation.

5. The method for detecting different types of defects in carbon fiber skins according to claim 4 is characterized in that: In a, a median filter is used instead of a Gaussian filter. The median filter sets the gray value of each pixel in the feature image to the median of the gray values ​​of all pixels in a certain neighborhood window of the point, thereby effectively removing noise while retaining edge and detail information. In b, the calculation method of the gradient amplitude is improved. On the basis of the traditional Canny operator only calculating the derivatives of the pixel points in the 0° and 90° directions, the derivatives in the 45° and 135° directions are added. The calculation formula is as follows: G 45° (x,y)=f(x+1,y+1)-f(x-1,y-1) G 135° (x,y)=f(x-1,y+1)-f(x+1,y-1) Among them, G 45° Represents the derivative of the gray value of the pixel in the 45° direction, G 135° represents the derivative of the gray value of the pixel in the 135° direction, x represents the horizontal coordinate of the pixel, y represents the vertical coordinate of the pixel, and f represents the gray value of the pixel; The gradient magnitude and direction are calculated as follows: Where M represents the gradient amplitude and θ represents the gradient direction; The calculation method of gradient amplitude is improved to obtain more directional information, improve edge recognition accuracy, and reduce edge loss and misjudgment; In the above c, the maximum inter-class variance method is used to realize adaptive threshold segmentation. The maximum inter-class variance method divides the image into two parts, background and target, by calculating the grayscale histogram of the image. The threshold when the inter-class variance of these two parts is the largest is the optimal threshold. The calculation formula of the inter-class variance is as follows: g=ω0ω1(μ0-μ1) 2 Among them, g is the between-class variance, ω0 is the ratio of target pixels to the total number of pixels, ω1 is the ratio of background pixels to the total number of pixels, μ0 is the average gray value of target pixels, and μ1 is the average gray value of background pixels.

6. The method for detecting different types of defects in carbon fiber skin according to claim 1, characterized in that: The BP neural network includes an input layer, a hidden layer and an output layer; the input layer includes the temperature difference peak value, temperature difference peak time, major diameter, circumference and area of ​​the defect; the output layer is the defect depth; the temperature difference peak value is the maximum value of the temperature difference between the defect area and the non-defect area of ​​the skin after pulse excitation; the temperature difference peak time is the time when the temperature difference peak occurs; the formula of the output process in the BP neural network is expressed as follows: where x j is the output value of the hidden layer or output layer, m is the number of neurons in the input layer or hidden layer, ω ij is the weight from the i-th neuron to the j-th neuron, x i is the input value of the hidden layer or output layer, b j is the threshold of the jth neuron.

7. A system for detecting different types of defects in carbon fiber skins according to any one of claims 1 to 6, characterized in that: Includes the following modules: Infrared image sequence acquisition module (110): acquires an infrared image sequence of a defective carbon fiber skin; A characteristic image acquisition module (120) processes the infrared image sequence using a pulse phase method to obtain a characteristic image; Defect size acquisition module (130): processing the feature image based on the improved Canny operator to obtain the edge image, and then obtain the defect size parameter; Defect depth acquisition module (140): obtains defect depth through BP neural network.

8. An electronic device, characterized in that: The electronic device comprises: a processor (210); and a memory (220) for storing one or more programs; When the one or more programs are executed by the processor (210), the processor executes the method for detecting different types of defects of carbon fiber skin as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by the processor (210), the method for detecting different types of defects of carbon fiber skin as described in any one of claims 1 to 6 is implemented.

Images