Aircraft structure fatigue test-oriented crack automatic detection method

Abstract

The application belongs to the field of structural health monitoring, and particularly relates to a crack automatic detection method for aircraft structure fatigue test. The method comprises the following steps: step one, determining a detection position, performing fatigue test, and collecting a template image and a detection image of the detection position under the same test condition, wherein the template image is a non-crack image; step two, detecting the template image by using a crack target segmentation algorithm to determine the endpoint coordinate range of interference items in the template image; and step three, detecting the detection image by using the crack target segmentation algorithm, and if the detected crack position overlaps with any of the interference items, the crack is considered as an interference, otherwise, the crack is a real crack, and a warning is given. The application compares the non-crack template with the real-time detection image, finds the crack, and distinguishes the scratch and the stain, thereby reducing the false alarm rate and reducing the influence of the interference factors on the crack detection accuracy.

Description

Technical Field

[0001] This application belongs to the field of structural health monitoring, and specifically relates to an automatic crack detection method for fatigue testing of aircraft structures. Background Technology

[0002] Metal cracking is a common form of damage in aerospace structures. Timely detection and early warning of damage during fatigue testing of aerospace structures can expose weaknesses in the structural design, support the assessment of structural strength and integrity, and also provide a basis for the compilation of aerospace structural maintenance manuals.

[0003] Currently, crack detection methods in full-scale aircraft fatigue testing mainly include visual inspection, eddy current testing, and ultrasonic testing. These methods all heavily rely on expert experience, and due to the complex testing environment, high detection risks during loading, and difficulties in operation within confined spaces, crack detection suffers from high labor costs, long processing times, and low reliability. Therefore, achieving automated and intelligent high-reliability crack detection in aerospace structures is a pressing issue that needs to be addressed in full-scale aircraft fatigue testing.

[0004] With the rapid development of robotics and artificial intelligence technologies and their application in the civilian sector over the past decade, machine vision has provided a new solution for automated crack detection in aircraft fatigue testing. By acquiring high-definition images of the inspection area through high-precision motion systems (such as crawling robots and robotic arms) and industrial cameras, and then applying target detection algorithms for automatic crack identification and damage warning, the adverse effects of manual labor in terms of cost, real-time performance, and risks can be significantly reduced.

[0005] Currently, deep learning object detection algorithms, such as Faster R-CNN and the YOLO series, are widely used in object detection and segmentation due to their advantages of speed and high accuracy. However, in aircraft structural fatigue tests, due to the complexity of the environment, interference such as dirt and scratches with high similarity to crack features can easily appear in the detection area. If existing algorithms are used directly for crack detection, there will be a high false alarm rate.

[0006] Therefore, it is desirable to have a technical solution to overcome or at least mitigate one of the aforementioned defects of the prior art. Summary of the Invention

[0007] The purpose of this application is to provide an automatic crack detection method for aircraft structural fatigue testing, in order to solve at least one problem existing in the prior art.

[0008] The technical solution of this application is:

[0009] An automatic crack detection method for aircraft structural fatigue testing includes:

[0010] Step 1: Determine the part to be tested and conduct a fatigue test. Under the same test conditions, acquire a template image and a test image of the part to be tested. The template image is a crack-free image.

[0011] Step 2: Use a crack target segmentation algorithm to detect the template image and determine the endpoint coordinate range of interference items in the template image;

[0012] Step 3: Use a crack target segmentation algorithm to detect the image to be detected. If the detected crack position overlaps with any of the interference items, the crack is considered to be interference; otherwise, it is considered a real crack, and an early warning is issued.

[0013] In at least one embodiment of this application, step one, which involves determining the area to be tested, conducting a fatigue test, and acquiring a template image and a test image of the area to be tested under the same testing conditions, includes:

[0014] S11. Determine multiple parts to be inspected, number the parts to be inspected, and determine the corresponding inspection conditions for each part to be inspected.

[0015] S12. A fatigue test is conducted using a testing machine. When the test load reaches the test condition, the fatigue test control system automatically maintains the load. At the same time, the fatigue test control system sends a control command to the image acquisition system, which controls the industrial camera in the image acquisition system to acquire template images and images to be tested for each part to be tested.

[0016] In at least one embodiment of this application, in S12, the template image and the image to be tested for the same part to be tested are acquired at different times under the same examination conditions. The template image is a crack-free image acquired at the beginning of the fatigue test, and the image to be tested is acquired during the subsequent fatigue test.

[0017] In at least one embodiment of this application, in S12, at different time periods of the test run, a template images are collected for each detection site, where the value of a is 10 to 15.

[0018] In at least one embodiment of this application, step two, which involves using a crack target segmentation algorithm to detect the template image and determine the endpoint coordinate range of interference items in the template image, includes:

[0019] S21. Use the crack target segmentation algorithm to detect a template images, and classify the same interference item in different template images into one category. If a certain interference item is detected only in one template image, then classify it as a separate category.

[0020] S22. Select a key structure in the area to be detected as the origin of the template image, calculate the endpoint coordinate range of each type of interference, and form an interference library.

[0021] In at least one embodiment of this application, in step three, the crack target segmentation algorithm is used to detect the image to be detected. If the detected crack position overlaps with any of the interference items, the crack is considered to be interference; otherwise, it is considered a real crack, and an early warning is issued. This includes:

[0022] S31. The crack target segmentation algorithm is used to detect the image to be detected;

[0023] S32. Define the origin of the coordinates of the image to be detected and the template image as the same;

[0024] S33. Obtain the two endpoints of the crack detected in the image to be detected, and determine whether the coordinates of the two endpoints are both within the range of the endpoint coordinates of the interference term.

[0025] If so, the detected crack will be marked as interference and no warning will be issued;

[0026] If not, the detected crack will be marked as a real crack, and an early warning will be issued.

[0027] In at least one embodiment of this application, step four is further included: during the automatic crack detection process using the crack detection system, if a crack automatically detected by the system is manually identified as interference, the interference is added to the interference library to avoid subsequent continuous false alarms.

[0028] The invention has at least the following beneficial technical effects:

[0029] The automatic crack detection method for aircraft structural fatigue testing proposed in this application is based on a comparison strategy. It applies a mature target segmentation algorithm to replace the target detection algorithm. While detecting the presence or absence of cracks, it obtains the coordinate information of the crack endpoints. Taking advantage of the characteristic that "interference information exists from the beginning of the test, while cracks are generated later", it discovers cracks and distinguishes between scratches, dirt and other interference by comparing crack-free templates and real-time detection images, thereby reducing the false alarm rate and minimizing the impact of interference factors on the accuracy of crack detection. Attached Figure Description

[0030] Figure 1 This is a flowchart of an image acquisition process according to one embodiment of this application;

[0031] Figure 2 This is a schematic diagram of an interference item in one embodiment of this application. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some, but not all, embodiments of this application. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0033] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the scope of protection of this application.

[0034] The following is in conjunction with the appendix Figures 1 to 2 This application will be described in further detail.

[0035] This application provides an automatic crack detection method for aircraft structural fatigue testing, including load-controlled image acquisition, interference location marking of template images, and detection of the image to be detected. Specifically, it includes the following steps:

[0036] Step 1: Determine the part to be tested and conduct a fatigue test. Under the same test conditions, acquire the template image and the image to be tested of the part to be tested. The template image is an image without cracks.

[0037] Step 2: Use the crack target segmentation algorithm to detect the template image and determine the range of endpoint coordinates of interference items in the template image;

[0038] Step 3: Use the crack target segmentation algorithm to detect the image to be detected. If the detected crack position overlaps with any of the interference items, the crack is considered to be interference; otherwise, it is considered a real crack, and an early warning is issued.

[0039] The automatic crack detection method for aircraft structural fatigue testing proposed in this application is based on load-controlled image acquisition. It acquires template images (i.e., crack-free images) and images to be tested (measured images at any time during the fatigue test) under the same load conditions. This avoids interference from structural elastic deformation caused by load differences on image alignment, ensuring that the same key structure in the template image and the image to be tested does not undergo significant shape and position changes, thereby reducing damage misjudgment.

[0040] In a preferred embodiment of this application, the image acquisition process based on load control is as follows:

[0041] S11. Identify multiple parts to be inspected, number them, and determine the corresponding inspection conditions for each part to be inspected.

[0042] S12. Fatigue tests are conducted using a testing machine. When the test load reaches the test condition, the fatigue test control system automatically maintains the load. At the same time, the fatigue test control system sends control commands to the image acquisition system, which controls the industrial camera in the image acquisition system to acquire template images and images to be tested for each part to be tested.

[0043] In this embodiment, the template image and the image to be tested for the same location are acquired at different times under the same testing conditions. The template image is a crack-free image acquired at the beginning of the fatigue test, while the image to be tested is acquired during subsequent fatigue tests. At different times during the test, *a* template images are acquired for each location, where *a* ranges from 10 to 15. The multiple template images acquired cover the influence of factors such as testing machine vibration on image brightness and clarity.

[0044] This application presents an automatic crack detection method for aircraft structural fatigue testing. Based on structural mechanics analysis, it identifies potential damage points on the testing machine that require detection and assigns them sequential numbers. Then, for each detection point, a load condition with a high load and frequent occurrence is selected from the fatigue test load spectrum as the examination condition. The fatigue test control system in the crack detection system is then connected to an image acquisition system. Once the test reaches the examination condition, the fatigue test control system automatically maintains the load and simultaneously sends instructions to the image acquisition system to guide the industrial camera in image acquisition. After image acquisition is complete, the test resumes operation. The process is as follows: Figure 1 As shown.

[0045] The automatic crack detection method for aircraft structural fatigue testing disclosed in this application uses a pre-designed crack target segmentation algorithm to detect interference locations in the template image. Since the template image may contain scratches, stains, or other interference similar to crack features, the algorithm may misreport these interferences as cracks. Therefore, the crack detection system records the locations of interferences in the template image that are misreported as cracks, providing a basis for subsequently determining the authenticity of the warning based on the location information.

[0046] In a preferred embodiment of this application, the interference location marking of the template image specifically includes:

[0047] S21. Use the crack target segmentation algorithm to detect a template images, and classify the same interference item in different template images into one category. If a certain interference item is detected only in one template image, then classify it as a separate category.

[0048] S22. Select a key structure (such as a rivet) in the area to be detected as the origin of the template image, calculate the endpoint coordinate range of each type of interference, and form an interference library.

[0049] like Figure 2 As shown, in this embodiment, there are three key structural rivet holes 1, 2, and 3. Considering image distortion and other factors, rivet 2 in the middle of the field of view is selected as the matching object for each template image in this detection area. The midpoint of rivet 2 is the origin of the template image coordinates. Assuming that there is a scratch on the upper part of rivet 2 with high similarity to a crack, it is detected as a crack b times in a template images. Due to factors such as experimental vibration and slight elastic deformation of the structure, the relative position of the scratch and the origin changes slightly in b misjudged images. The scratch positions closest to and farthest from the origin are position 1 and position 2, respectively. Then, the scratch is defined as an interference item. At the same time, the endpoint coordinates of the scratch at positions 1 and 2 can be obtained by the crack target segmentation algorithm as [(x1,y1),(x2,y2)], [(x'1,y'1),(x'2,y'2)], respectively. Then, the coordinate range of the two endpoints of the interference item is defined as [(x1,y1),(x'1,y'1)](...). Figure 2 In the range 1), [(x2,y2),(x'2,y'2)]( Figure 2 In this embodiment, and in accordance with Figure 2 The method shown calibrates the endpoint coordinate range of all interference items in the area to be detected, forming a set of interference item coordinates P. i ={M1,M2,...,M... j M n}, where i represents the number of the testing part in the testing machine, M j This represents the range of endpoint coordinates of the j-th interference item in the i-th detection location, where there are a total of n interference items in the detection location.

[0050] The automatic crack detection method for aircraft structural fatigue testing disclosed in this application uses a crack target segmentation algorithm (such as Mask-RCNN) to segment the cracks in the image to be detected during the test. Figure 1 The process involves collecting images to be inspected. If the location of a detected crack overlaps with any of the interference items, the crack is considered an interference; otherwise, it is considered a real crack and an alert is required.

[0051] In a preferred embodiment of this application, determining whether a detected crack is an interference specifically involves:

[0052] S31. The crack target segmentation algorithm is used to detect the image to be detected;

[0053] S32. Define the origin of the coordinate system of the image to be detected and the template image as the same.

[0054] S33. Obtain the two endpoints of the crack detected in the image to be detected, and determine whether the coordinates of the two endpoints are both within the range of the endpoint coordinates of the interference term.

[0055] If so, the detected crack will be marked as interference and no warning will be issued;

[0056] If not, the detected crack will be marked as a real crack, and an early warning will be issued.

[0057] The automatic crack detection method for aircraft structural fatigue testing in this application specifies that the origin of the coordinate markers of the image to be detected and the template image are the same, for example, for... Figure 2 The locations to be inspected shown are all referenced to the midpoint of rivet hole 2 as the origin of the coordinate system. If the coordinates of the two endpoints of a crack detected in real time are within the range of the endpoint coordinates of any interference item, the detected crack is recorded as interference and no warning is issued. For example, regarding... Figure 2 If the coordinates of the two ends of the detected crack are located in range 1 and range 2 respectively, they are considered as interference.

[0058] The automatic crack detection method for aircraft structural fatigue testing in this application also includes interference data updating. Specifically, in step four, during the automatic crack detection process using the crack detection system, if a crack automatically detected by the system is manually identified as interference, the interference is added to the interference library to avoid subsequent false alarms. That is, during real-time detection, if a misjudged crack is found, it needs to be treated as a new interference item, and its coordinates stored in the original interference item coordinate set.

[0059] This application presents an automatic crack detection method for aircraft structural fatigue testing. Based on the completed crack segmentation model design, it is applied to the task of detecting the presence or absence of cracks during fatigue testing. Due to interference from factors such as scratches and dirt, crack segmentation models often experience false alarms in practical applications. This application aims to reduce false alarms and, targeting real-world engineering applications, designs four working steps: First, before the fatigue test begins, based on structural mechanics analysis, n locations requiring crack detection are determined; then, in the initial stage of the fatigue test, when no cracks are found at the detection locations, an inspection robot or a fixed industrial camera is used to... Figure 1 The process shown involves acquiring *a* template images of all detected areas and calibrating the location range of all interference items. It's important to note that the camera position and angle must be fixed for each image acquisition to ensure the image field of view remains unchanged. Furthermore, the template image acquisition time needs to be spaced out to ensure the template library covers the influence of random factors such as vibration during fatigue testing. After template image acquisition, the real-time crack detection stage begins. If the target segmentation algorithm detects a crack, interference identification is performed first (specific methods are as follows). Figure 2 As shown in the figure, it also alarms for non-interference cracks. Since full-scale fatigue testing of aircraft is generally carried out for 5 to 10 years, and manual cabin inspections and wiring rearrangements are carried out during this process, which may introduce new interference items, the last step is to update the interference data. That is, if the cracks automatically warned by the system are identified as interference by manual detection during the real-time detection process, they are added to the interference data set to avoid continuous false alarms.

[0060] This application presents an automatic crack detection method for aircraft structural fatigue testing. Firstly, the effectiveness of a crack detection and segmentation system depends on both the target detection and segmentation algorithm itself and the quality of the acquired images. From an engineering application perspective, this application designs an image acquisition method that connects the fatigue test control system to an industrial camera. Image acquisition is triggered by load spectrum conditions, enabling imaging of cracks in their open state. This reduces the difficulty of crack detection and avoids the influence of structural elastic deformation and other factors on image alignment, ensuring the consistency between the template and the real-time detection image, thereby reducing false alarms caused by differences in image quality. Secondly, due to the complexity of full-scale aircraft fatigue testing, existing algorithms struggle to completely distinguish between cracks and scratches. Considering practical engineering scenarios such as test aircraft vibration, this application utilizes a comparison mechanism to design an interference elimination method. This improves the overall early warning accuracy of the system while reducing the system's requirements for target detection and segmentation algorithms.

[0061] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An automatic crack detection method for aircraft structural fatigue testing, characterized in that, include: Step 1: Determine the part to be tested and conduct a fatigue test. Under the same test conditions, acquire a template image and a test image of the part to be tested. The template image is a crack-free image. Step 2: Detect the template image using a crack target segmentation algorithm to determine the endpoint coordinate range of interference items in the template image, including: S21. Use the crack target segmentation algorithm to detect a template images, and classify the same interference item in different template images into one category. If a certain interference item is detected only in one template image, then classify it as a separate category. S22. Select a key structure in the area to be detected as the origin of the template image, calculate the endpoint coordinate range of each type of interference item, and form an interference library. Step 3: The crack target segmentation algorithm is used to detect the image to be detected. If the detected crack location overlaps with any of the interference items, the crack is considered interference; otherwise, it is a real crack, and an early warning is issued, including: S31. The crack target segmentation algorithm is used to detect the image to be detected; S32. Define the origin of the coordinates of the image to be detected and the template image as the same; S33. Obtain the two endpoints of the crack detected in the image to be detected, and determine whether the coordinates of the two endpoints are both within the range of the endpoint coordinates of the interference term. If so, the detected crack will be marked as interference and no warning will be issued; If not, the detected crack will be marked as a real crack, and an early warning will be issued.

2. The automatic crack detection method for aircraft structural fatigue testing according to claim 1, characterized in that, In step one, determining the area to be tested, conducting fatigue tests, and acquiring template images and test images of the area to be tested under the same testing conditions includes: S11. Determine multiple parts to be inspected, number the parts to be inspected, and determine the corresponding inspection conditions for each part to be inspected. S12. A fatigue test is conducted using a testing machine. When the test load reaches the test condition, the fatigue test control system automatically maintains the load. At the same time, the fatigue test control system sends a control command to the image acquisition system, which controls the industrial camera in the image acquisition system to acquire template images and images to be tested for each part to be tested.

3. The automatic crack detection method for aircraft structural fatigue testing according to claim 2, characterized in that, In S12, the template image and the image to be tested for the same part to be tested are collected under the same working conditions at different times. The template image is a crack-free image collected at the beginning of the fatigue test, and the image to be tested is collected during the subsequent fatigue test.

4. The automatic crack detection method for aircraft structural fatigue testing according to claim 3, characterized in that, In S12, at different time periods during the test run, a template images are collected for each detection site, where the value of a ranges from 10 to 15.

5. The automatic crack detection method for aircraft structural fatigue testing according to claim 4, characterized in that, It also includes step four: during the automatic crack detection process using the crack detection system, if a crack automatically detected by the system is manually identified as interference, then the interference is added to the interference library to avoid subsequent false alarms.

Images