Quantitative analysis method for internal defects of carbon fiber honeycomb interlayer based on artificial intelligence

Abstract

The invention belongs to the technical field of nondestructive testing and identification, and discloses a carbon fiber honeycomb interlayer internal defect quantitative analysis method based on artificial intelligence, which comprises the following steps: acquiring an infrared nondestructive testing image; performing defect image segmentation by using a YOLO neural network to obtain each pixel point value of the mask; setting a mask threshold value to control the generation size of the mask, and calibrating the mask threshold value to obtain a calibrated mask threshold value; acquiring an infrared nondestructive detection image of a workpiece to be detected under the same detection condition, inputting the infrared nondestructive detection image into the neural network, and applying the calibrated mask threshold value to obtain an accurate defect boundary; the accurate defect area is obtained by multiplying the number of the pixel points of the mask by the actual area of the pixel points, and quantitative analysis of the internal defects of the workpiece is realized. According to the carbon fiber honeycomb interlayer internal defect quantitative analysis method based on artificial intelligence, rapid and large-scale detection and defect quantitative analysis are achieved, and technical support is provided for quality control of composite materials.

Description

Technical Field

[0001] The invention relates to the technical field of nondestructive testing and identification, and in particular to an artificial intelligence-based quantitative analysis method for internal defects of a carbon fiber honeycomb sandwich. Background Art

[0002] As modern industry continues to increase its requirements for material performance, carbon fiber honeycomb sandwich composites are widely used in aerospace, high-speed trains, automobile manufacturing and other fields due to their light weight, high strength, excellent rigidity and impact resistance. These materials are usually composed of thin and strong carbon fiber panels and lightweight honeycomb core materials. This structure not only provides high strength and rigidity, but also reduces structural weight and improves energy efficiency. However, carbon fiber honeycomb sandwich composites are prone to internal defects such as delamination and holes during the manufacturing process. These defects seriously affect the mechanical properties and structural integrity of the material and increase safety risks. Traditional detection methods, such as ultrasonic testing and X-ray testing, have problems such as slow detection speed, high cost, complex operation or potential harm to the environment and personnel, and the accuracy and efficiency of quantitative analysis of defects need to be improved.

[0003] Infrared nondestructive testing technology uses the principle of infrared thermal imaging to infer internal defects by detecting thermal anomalies on the surface of materials. This method has the advantages of being fast and non-contact, and is particularly suitable for materials that are sensitive to the environment or difficult to detect by traditional methods. However, existing infrared nondestructive testing technology still faces challenges in defect detection and quantitative analysis of carbon fiber honeycomb sandwich composites, including how to accurately identify the boundaries of defects and how to achieve accurate measurement of defect size and area.

[0004] With the development of artificial intelligence technology, image recognition and analysis technology based on deep learning has provided new possibilities for defect detection of composite materials. As an efficient target detection algorithm, YOLO neural network has attracted attention in the field of industrial inspection due to its fast speed and high accuracy. Although defects in infrared images can be segmented using YOLO, the defect boundaries segmented by artificial intelligence are often not accurate enough because the boundaries of defects in infrared images are relatively fuzzy, which leads to problems such as the inability to accurately measure defect size and area. Summary of the invention

[0005] The purpose of the present invention is to provide an artificial intelligence-based quantitative analysis method for internal defects of carbon fiber honeycomb sandwiches, to solve the problem of inaccurate defect boundary recognition in infrared images by artificial intelligence technology, and to provide technical support for quality control of composite materials.

[0006] To achieve the above object, the present invention provides an artificial intelligence-based internal defect quantitative analysis method for carbon fiber honeycomb sandwich, comprising the following steps: Step S1, obtaining an infrared nondestructive testing image of a standard test block of a carbon fiber sandwich composite material with a known defect area; Step S2, using the YOLO neural network to perform defect image segmentation on the infrared nondestructive testing image obtained in step S1, and obtaining the value of each pixel point of the mask in the image segmentation result; Step S3, controlling the size of the mask generated by setting a mask threshold, calibrating the mask threshold using a defect of known area, and obtaining a calibrated mask threshold; Step S4, under the same infrared nondestructive testing conditions, obtaining an infrared nondestructive testing image of the workpiece to be tested, inputting it into the neural network and applying the calibrated mask threshold to obtain an accurate defect boundary; Step S5: Calculate the number of mask pixels and multiply it by the actual area of ​​the pixels to obtain an accurate defect area, thereby achieving quantitative analysis of internal defects of the carbon fiber honeycomb sandwich composite material.

[0007] Preferably, in step S1, an infrared nondestructive testing device is used to test a standard test block with a known defect area, and an infrared nondestructive testing image is obtained; When acquiring infrared nondestructive testing images, the infrared equipment should be used to test the standard test block to ensure that the acquired workpiece and defect images are not distorted or deformed, and to avoid changes in the shape and proportion of the workpiece and defect images.

[0008] Preferably, the infrared nondestructive testing system includes an infrared thermal excitation module, an infrared thermal image acquisition module, and a control and post-processing module.

[0009] Preferably, in step S2, the infrared nondestructive testing image obtained in step S1 is segmented using a YOLO neural network to obtain each pixel value of the mask in the image segmentation result. The specific process is as follows: Step S21, first, a data set of infrared non-destructive defects is prepared, and the defects in the data set are framed and marked with polygons, and the polygons are larger than the defect boundaries; Step S22, then, input the data set into the YOLO image segmentation neural network, perform training and obtain the trained neural network; Step S23, finally, use the trained neural network to perform image segmentation on the infrared nondestructive testing image of the standard test block obtained in step S1, and the image segmentation result will include a mask of the defect; The essence of the mask is a matrix, which contains the value of each mask pixel.

[0010] Preferably, in step S3, the size of the mask generated is controlled by setting a mask threshold, and the mask threshold is calibrated using a defect of known area to obtain a calibrated mask threshold. The specific process is as follows: Step S31, based on the matrix of the defect mask obtained in step S2, a mask threshold is set, and when the value in the matrix is ​​less than the set defect threshold, mask pixels less than the threshold are removed to control the size of the final mask generated; Step S32, calibrating the mask threshold using a defect of known area to obtain a calibrated mask threshold; After setting a certain mask threshold, count the number of mask pixels and multiply it by the actual area of ​​the pixels to get the defect area under the threshold; after traversing the mask thresholds within the range, get the defect area under each threshold, and the threshold with the smallest error with the actual defect area is the calibrated mask threshold.

[0011] Preferably, in step S4, under the same infrared nondestructive testing conditions, an infrared nondestructive testing image of the workpiece to be tested is obtained, input into a neural network and a calibrated mask threshold is applied to obtain an accurate defect boundary. The specific process is as follows: Step S41, under the same infrared nondestructive testing conditions as step S1, obtaining an infrared nondestructive testing image of the workpiece to be tested; Among them, the detection parameters include detection distance and angle, ambient temperature, acquisition frequency, exposure time, focal length, field of view, filtering parameters, and imaging parameters; Step S42: input the acquired infrared nondestructive testing image into the neural network for prediction, and apply the calibrated mask threshold to obtain an accurate defect segmentation mask if defects exist.

[0012] Preferably, in step S5, the defect mask is obtained according to step S4, the number of mask pixels is calculated, and the number is multiplied by the actual area of ​​the pixels to obtain an accurate defect area, thereby achieving quantitative analysis of internal defects of carbon fiber honeycomb sandwich composite materials.

[0013] Therefore, the present invention adopts the above-mentioned artificial intelligence-based internal defect quantitative analysis method of carbon fiber honeycomb sandwich, and the beneficial effects are as follows: (1) The present invention utilizes infrared nondestructive testing technology combined with artificial intelligence algorithms to perform defect image segmentation on infrared images through the YOLO neural network. After using the mask threshold calibration method, it can more accurately perform quantitative analysis of defects. (2) The present invention utilizes artificial intelligence methods and only needs to set calibrated thresholds. Under the same detection conditions, it can achieve rapid, large-scale detection and defect quantitative analysis tasks, providing technical support for the quality control of composite materials.

[0014] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1This is a flow chart of the internal defect quantitative analysis method of the carbon fiber honeycomb sandwich based on artificial intelligence of the present invention; Figure 2 is the mask threshold calibration curve of the present invention; Figure 3 For the present invention The original image of the infrared nondestructive detection of defects and the masks under different mask thresholds; among them, (a) is the original image; (b) is the mask with a mask threshold of 0.1; (c) is the mask with a mask threshold of 0.66; (d) is the mask with a mask threshold of 0.9. DETAILED DESCRIPTION

[0016] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0017] like Figure 1 As shown, the internal defect quantitative analysis method of the carbon fiber honeycomb sandwich based on artificial intelligence of the present invention comprises the following steps: Step S1, obtaining an infrared nondestructive testing image of a standard test block of a carbon fiber sandwich composite material with a known defect area.

[0018] Use infrared nondestructive testing equipment to test standard test blocks with known defect areas and obtain infrared nondestructive testing images. When obtaining infrared nondestructive testing images, the infrared equipment should be directly facing the standard test blocks to ensure that the obtained workpiece and defect images are not distorted or deformed as much as possible, and to avoid changes in the shape and proportion of the workpiece and defect images.

[0019] Step S2: Use the YOLO neural network to perform defect image segmentation on the infrared nondestructive testing image obtained in step S1, and obtain the value of each pixel point of the mask in the image segmentation result.

[0020] Step S21, first, create a data set of infrared lossless defects, and the defects in the data set are framed and marked with polygons, and the polygons should be slightly larger than the defect boundaries.

[0021] Step S22: Then, the data set is input into the YOLO image segmentation neural network for training and obtaining the trained neural network.

[0022] Step S23: Finally, the trained neural network is used to segment the infrared nondestructive testing image of the standard test block obtained in step S1, and the image segmentation result will contain the defect mask. The essence of the mask is a matrix, which contains the value of each mask pixel.

[0023] Step S3, controlling the size of the mask generated by setting a mask threshold, and calibrating the mask threshold using a defect of known area to obtain a calibrated mask threshold.

[0024] Step S31 : setting a mask threshold to remove mask pixels smaller than the threshold, thereby controlling the size of the final mask.

[0025] Based on the matrix of the defect mask obtained in step S2, when the value in the matrix is ​​less than the set defect threshold, the point is removed and the generated defect mask becomes smaller. Among them, when making the defect data set in step S2, the polygon of the defect should be slightly larger than the defect boundary. The purpose is to make the defect mask recognized by the neural network model slightly larger, so as to avoid the recognized mask being smaller than the actual size of the defect, and at the same time, it also allows the defect mask to have room for reduction, ensuring the feasibility of correcting the defect area.

[0026] Step S32: calibrate the mask threshold using defects of known area to obtain a calibrated mask threshold.

[0027] After setting a certain mask threshold, count the number of mask pixels and multiply it by the actual area of ​​the pixels to get the defect area under the threshold. After traversing the mask thresholds within a certain range, the defect area under each threshold is obtained. The threshold with the smallest error with the actual defect area is the calibrated mask threshold.

[0028] Step S4: Under the same infrared nondestructive testing conditions, an infrared nondestructive testing image of the workpiece to be tested is obtained, input into the neural network and the calibrated mask threshold is applied to obtain an accurate defect boundary.

[0029] Step S41, under the same infrared nondestructive testing conditions as step S1, an infrared nondestructive testing image of the workpiece to be tested is obtained.

[0030] Among them, the detection parameters include detection distance and angle, ambient temperature, acquisition frequency, exposure time, focal length, field of view, filtering parameters, imaging parameters, etc.

[0031] Step S42: input the acquired infrared nondestructive testing image into the neural network for prediction, and apply the calibrated mask threshold. If defects exist, an accurate defect segmentation mask can be obtained.

[0032] Step S5: Calculate the number of mask pixels and multiply it by the actual area of ​​the pixels to obtain an accurate defect area, thereby achieving quantitative analysis of internal defects of the carbon fiber honeycomb sandwich composite material.

[0033] According to step S4, the defect mask is obtained, the number of pixels in the defect mask is calculated, and the number is multiplied by the area of ​​the actual pixels to obtain the actual area of ​​the defect, thereby achieving quantitative analysis of the defect.

[0034] Example The present embodiment is a quantitative analysis method for internal defects of carbon fiber honeycomb sandwich composite materials based on artificial intelligence. The tested sample is a carbon fiber honeycomb sandwich composite material. The carbon fiber skin is made of M40J, the panel and the honeycomb core are glued with J-78B adhesive film, the skin thickness is 0.5mm, and the honeycomb core is spliced ​​with J-78D foam glue; the honeycomb core height is 15mm; a porous honeycomb core is used, and the honeycomb core specification is LF2Y, 0.03mm×5mm.

[0035] The method for making standard test pieces with artificial prefabricated defects is as follows: referring to the defect requirements of GJB1719-93 "General Specification for Aluminum Honeycomb Sandwich Structures", the types of internal defects that may occur during the manufacturing process of carbon fiber sandwich composites are mainly pores, which can be simulated using the film missing method. The film corresponding to the prefabricated defect area is removed, and a layer of release agent is applied to the film missing area. The defect size is and , respectively used for calibration and testing of quantitative analysis of internal defects in carbon fiber honeycomb sandwich composites.

[0036] like Figure 1 As shown, the internal defect quantitative analysis method of the carbon fiber honeycomb sandwich based on artificial intelligence in the embodiment of the present invention comprises the following steps: Step S1: Use an infrared nondestructive testing system to detect defects with a size of The standard test block is used for testing. The infrared non-destructive testing system generally includes: infrared thermal excitation module, infrared thermal image acquisition module and control and post-processing module.

[0037] In this embodiment, the infrared acquisition module needs to face the standard test block to ensure that the workpiece and defects in the infrared image are not distorted or deformed, and to avoid changes in the shape and proportion of the workpiece and defect images.

[0038] Step S2: Before image segmentation, the YOLO neural network needs to be trained using existing infrared data. There are a large number of infrared nondestructive testing images in a carbon fiber honeycomb sandwich composite material manufacturing plant. First, these images are used to create a data set, and the defects in the data set are framed and marked using X-AnyLabeling software. The polygon is selected in the way of framed defects, and the polygon should be slightly larger than the defect boundary. The created data set is divided into training set, validation set and test set in a ratio of 7:1.5:1.5.

[0039] Next, build the YOLO neural network, use Python version 3.10, and the neural network framework use ultralytics version 8.3.25. After using Anaconda and VS Code to build the YOLO neural network environment, select the yolov8m-segment neural network version for training, and obtain the trained best.pt neural network parameters.

[0040] Finally, the infrared nondestructive testing image obtained in step S1 is input into the trained neural network, and after the defect image is segmented, the defect mask matrix and the number of pixels of the defect mask are output.

[0041] Step S3: define the mask matrix in step S2 as , and convert the values ​​in the mask matrix into values ​​between 0 and 1. The converted matrix is ​​used Indicates that the conversion formula is as follows: ; in, Indicates the matrix Row, No. The index of the column.

[0042] When the mask threshold is set to MaskThreshold, the resulting mask matrix , as shown below: ; In SegmentMask, 1 means generating a mask, and 0 means not generating a mask. The number of pixels generated by the mask can be controlled by the mask threshold MaskThreshold.

[0043] The defect size in step S1 is The infrared image is input into the neural network, the number of pixels under different mask thresholds is counted, and compared with the actual pixel area 0.8721mm 2 By multiplying, we can get the defect area under different mask thresholds, such as Figure 2 As shown, the larger the mask threshold, the smaller the defect area.

[0044] Figure 2 The mask threshold range is 0 to 0.99 with an interval of 0.01. The defect size is The actual area is 1256.6mm 2 , the point closest to the true value is recorded as B, and by calculating the minimum adjacent point, we can get point B at Figure 2 The coordinates in are (0.66, 1255.68), so the calibrated mask threshold MaskThreshold is 0.66.

[0045] Step S4: Under the same detection conditions as step S1, the defect size is obtained as The infrared non-destructive testing image is fed into the trained neural network, and the calibrated mask threshold of 0.66 is used to control the size of the mask generation.

[0046] Step S5: Based on the number of pixels masked in step S4, multiply it by the actual pixel area to obtain The area of ​​the defect is 312.30mm 2 , calculated according to the circle area formula The area of ​​the defect is 314.16mm 2 , and the error is -0.59%.

[0047] Figure 3 for The original infrared nondestructive detection image of the defect and the mask under different mask thresholds. It can be seen that the larger the mask threshold, the smaller the mask area, which proves the necessity of mask threshold calibration.

[0048] When the mask threshold is set to 0.1, the defect area is 361.94 mm 2 , the error is 15.21%; when the mask threshold is set to 0.9, the defect area is 259.22mm 2 , with an error of -17.49%.

[0049] When the mask threshold setting deviates from the calibrated value, the defect area error is large, while the error after calibration is only -0.59%, which further proves that the quantitative analysis method of calibration before measurement is necessary and effective.

[0050] Among them, step S4 can obtain a large number of infrared non-destructive testing images of carbon fiber honeycomb sandwich composite materials to be tested. After executing the quantitative analysis of step S5, fast and large-scale detection and defect quantitative analysis tasks can be achieved, providing technical support for the quality control of composite materials.

[0051] Therefore, the present invention adopts the above-mentioned artificial intelligence-based internal defect quantitative analysis method of carbon fiber honeycomb sandwich, utilizes infrared non-destructive testing technology combined with artificial intelligence algorithm, and performs defect image segmentation on infrared images through YOLO neural network. After using the mask threshold calibration method, it can more accurately perform quantitative analysis of defects; using the artificial intelligence method, it only needs to set the calibrated threshold, and under the same detection conditions, it can achieve fast, large-scale detection and defect quantitative analysis tasks, providing technical support for quality control of composite materials.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A quantitative analysis method for internal defects of carbon fiber honeycomb sandwich based on artificial intelligence, characterized in that: The following steps are involved: Step S1, obtaining an infrared nondestructive testing image of a standard test block of a carbon fiber sandwich composite material with a known defect area; Step S2, using the YOLO neural network to perform defect image segmentation on the infrared nondestructive testing image obtained in step S1, and obtaining the value of each pixel point of the mask in the image segmentation result; Step S3, controlling the size of the mask generated by setting a mask threshold, calibrating the mask threshold using a defect of known area, and obtaining a calibrated mask threshold; Step S4, under the same infrared nondestructive testing conditions, obtaining an infrared nondestructive testing image of the workpiece to be tested, inputting it into the neural network and applying the calibrated mask threshold to obtain an accurate defect boundary; Step S5: Calculate the number of mask pixels and multiply it by the actual area of ​​the pixels to obtain an accurate defect area, thereby achieving quantitative analysis of internal defects of the carbon fiber honeycomb sandwich composite material.

2. The method for quantitative analysis of internal defects of carbon fiber honeycomb sandwich based on artificial intelligence according to claim 1, characterized in that: In step S1, an infrared nondestructive testing device is used to test a standard test block with a known defect area, and an infrared nondestructive testing image is obtained; When acquiring infrared nondestructive testing images, the infrared equipment should be used to test the standard test block to ensure that the acquired workpiece and defect images are not distorted or deformed, and to avoid changes in the shape and proportion of the workpiece and defect images.

3. The method for quantitative analysis of internal defects of carbon fiber honeycomb sandwich based on artificial intelligence according to claim 2, characterized in that: The infrared nondestructive testing system includes an infrared thermal excitation module, an infrared thermal image acquisition module, and a control and post-processing module.

4. The method for quantitative analysis of internal defects of carbon fiber honeycomb sandwich based on artificial intelligence according to claim 1, characterized in that: In step S2, the YOLO neural network is used to perform defect image segmentation on the infrared nondestructive testing image obtained in step S1, and the value of each pixel of the mask in the image segmentation result is obtained. The specific process is as follows: Step S21, first, a data set of infrared non-destructive defects is prepared, and the defects in the data set are framed and marked with polygons, and the polygons are larger than the defect boundaries; Step S22, then, input the data set into the YOLO image segmentation neural network, perform training and obtain the trained neural network; Step S23, finally, use the trained neural network to perform image segmentation on the infrared nondestructive testing image of the standard test block obtained in step S1, and the image segmentation result will include a mask of the defect; The essence of the mask is a matrix, which contains the value of each mask pixel.

5. The method for quantitative analysis of internal defects of carbon fiber honeycomb sandwich based on artificial intelligence according to claim 1, characterized in that: In step S3, the size of the mask generated is controlled by setting the mask threshold, and the mask threshold is calibrated using a defect of known area to obtain a calibrated mask threshold. The specific process is as follows: Step S31, based on the matrix of the defect mask obtained in step S2, a mask threshold is set, and when the value in the matrix is ​​less than the set defect threshold, mask pixels less than the threshold are removed to control the size of the final mask generated; Step S32, calibrating the mask threshold using a defect of known area to obtain a calibrated mask threshold; After setting a certain mask threshold, count the number of mask pixels and multiply it by the actual area of ​​the pixels to get the defect area under the threshold; after traversing the mask thresholds within the range, get the defect area under each threshold, and the threshold with the smallest error with the actual defect area is the calibrated mask threshold.

6. The method for quantitative analysis of internal defects of carbon fiber honeycomb sandwich based on artificial intelligence according to claim 1, characterized in that: In step S4, under the same infrared nondestructive testing conditions, the infrared nondestructive testing image of the workpiece to be tested is obtained, input into the neural network and the calibrated mask threshold is applied to obtain the accurate defect boundary. The specific process is as follows: Step S41, under the same infrared nondestructive testing conditions as step S1, obtaining an infrared nondestructive testing image of the workpiece to be tested; Among them, the detection parameters include detection distance and angle, ambient temperature, acquisition frequency, exposure time, focal length, field of view, filtering parameters, and imaging parameters; Step S42: input the acquired infrared nondestructive testing image into the neural network for prediction, and apply the calibrated mask threshold to obtain an accurate defect segmentation mask if defects exist.

7. The method for quantitative analysis of internal defects of carbon fiber honeycomb sandwich based on artificial intelligence according to claim 1, characterized in that: In step S5, the defect mask is obtained according to step S4, the number of mask pixels is calculated, and the number is multiplied by the actual area of ​​the pixels to obtain the accurate defect area, thereby achieving quantitative analysis of internal defects of the carbon fiber honeycomb sandwich composite material.

Images