Material damage identification method and device, equipment and storage medium

Abstract

The invention relates to a material damage identification method and device, equipment and a storage medium. The method comprises the following steps: acquiring a scattered light space distribution image corresponding to a test material under laser irradiation; performing image feature extraction on the scattered light spatial distribution image through a trained deep learning model to obtain deep features of the image; and analyzing the deep features of the image through a trained deep learning model, and outputting a damage identification result of each test point on the test material, the damage identification result including damage or non-damage. Therefore, the scattered light space distribution information can be fully utilized, automatic detection of the material test points is realized, manual intervention is not needed, the detection efficiency is improved, and the recognition precision of material damage is remarkably improved.

Description

Technical Field

[0001] The present application relates to the technical field of material testing, and in particular to a material damage identification method, device, computer equipment, computer-readable storage medium and computer program product. Background Art

[0002] With the development of material testing technology, laser damage testing technology has been widely used in the performance evaluation of optical materials.

[0003] In traditional technology, a laser source is generally used to illuminate the sample to be tested, and then the scattered light image of the sample to be tested is recorded. Finally, the image features of the scattered light image are extracted and analyzed with the help of manual observation or simple image processing algorithms to determine whether each test point is damaged.

[0004] However, the image features extracted by the above traditional methods have a single dimension and are heavily dependent on the personal experience of the testers. They have the disadvantages of strong subjectivity and non-repeatability, resulting in low accuracy in material damage testing. Summary of the invention

[0005] Based on this, it is necessary to provide a material damage identification method, device, computer equipment, computer-readable storage medium and computer program product that can realize automatic detection of material damage and significantly improve test accuracy in response to the above technical problems.

[0006] In a first aspect, the present application provides a material damage identification method, comprising:

[0007] Collect the scattered light spatial distribution image corresponding to the test material under laser irradiation;

[0008] Extracting image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image;

[0009] The deep features of the image are analyzed by a trained deep learning model, and damage identification results of each test point on the test material are output, where the damage identification results include: damaged or undamaged.

[0010] In one embodiment, the collecting of the scattered light spatial distribution image corresponding to the test material under laser irradiation includes:

[0011] The scattered light pattern generated on the surface of the test material under laser irradiation is collected by a photosensitive coupled camera to obtain a scattered light spatial distribution image corresponding to the test material.

[0012] In one embodiment, the deep features of the image include at least one of the following:

[0013] Number, shape and size of light spots;

[0014] The distribution pattern of the light spot;

[0015] Texture details.

[0016] In one of the embodiments, before extracting image features from the scattered light spatial distribution image using a trained deep learning model, the method further includes:

[0017] The scattered light spatial distribution image corresponding to the test material is preprocessed, wherein the preprocessing method includes: denoising processing and / or normalization processing.

[0018] In one of the embodiments, before extracting image features from the scattered light spatial distribution image using a trained deep learning model, the method further includes:

[0019] Collecting a scattered light spatial distribution image of the material under laser irradiation, and marking the collected scattered light spatial distribution image as damaged or undamaged to obtain a marked image;

[0020] Constructing a training set and a test set based on the annotated images;

[0021] Constructing an initial deep learning model; wherein the network structure of the initial deep learning model includes: a convolutional layer, a pooling layer, and a fully connected layer;

[0022] The initial deep learning model is trained using the training set, the classification error is calculated using a cross entropy loss function, and the model parameters in the initial deep learning model are optimized using a back propagation algorithm to obtain the trained initial deep learning model.

[0023] In one embodiment, after obtaining the trained initial deep learning model, the method further includes:

[0024] Verifying the trained initial deep learning model through the test set to obtain a verification result;

[0025] Based on the verification result, the model parameters of the trained initial deep learning model are adjusted to obtain a trained deep learning model.

[0026] In one embodiment, after outputting the damage identification results of each test point on the test material, the method further includes:

[0027] The damage identification result is saved as a damage area feature image corresponding to each test point of the damage;

[0028] A visual test report is generated based on the damage identification results of each test point on the test material and the characteristic image of the damaged area.

[0029] In a second aspect, the present application further provides a material damage identification device, the device comprising:

[0030] A data acquisition module, used to acquire a spatial distribution image of scattered light corresponding to the test material under laser irradiation;

[0031] A feature extraction module is used to extract image features from the scattered light spatial distribution image through a trained deep learning model to obtain deep features of the image;

[0032] The damage identification module is used to analyze the deep features of the image through a trained deep learning model, and output the damage identification results of each test point on the test material, and the damage identification results include: damaged or undamaged.

[0033] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0034] Collect the scattered light spatial distribution image corresponding to the test material under laser irradiation;

[0035] Extracting image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image;

[0036] The deep features of the image are analyzed by a trained deep learning model, and damage identification results of each test point on the test material are output, where the damage identification results include: damaged or undamaged.

[0037] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0038] Collect the scattered light spatial distribution image corresponding to the test material under laser irradiation;

[0039] Extracting image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image;

[0040] The deep features of the image are analyzed by a trained deep learning model, and damage identification results of each test point on the test material are output, where the damage identification results include: damaged or undamaged.

[0041] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0042] Collect the scattered light spatial distribution image corresponding to the test material under laser irradiation;

[0043] Extracting image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image;

[0044] The deep features of the image are analyzed by a trained deep learning model, and damage identification results of each test point on the test material are output, where the damage identification results include: damaged or undamaged.

[0045] The above-mentioned material damage identification method, device, computer equipment, computer-readable storage medium and computer program product collect the scattered light spatial distribution image corresponding to the test material under laser irradiation; thus, the scattering result of the laser on the surface of the test material can be used as the basis for damage identification. Compared with naked eye observation and direct image acquisition, it can have more deep features, which is convenient for subsequent improvement of test accuracy. The image features of the scattered light spatial distribution image are extracted by a trained deep learning model to obtain the deep features of the image; thus, fully automatic image feature extraction can be achieved without manual intervention. The deep features of the image are analyzed by a trained deep learning model, and the damage identification results of each test point on the test material are output, and the damage identification results include: damaged or undamaged. Thus, the scattered light spatial distribution information can be fully utilized to realize the automatic detection of material test points without manual intervention, thereby improving the detection efficiency and significantly improving the recognition accuracy of material damage. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 is an application environment diagram of a material damage identification method in an embodiment;

[0048] Figure 2 It is a schematic diagram of the flow of the material damage identification method in the first embodiment;

[0049] Figure 3 A schematic diagram of a flow chart of a material damage identification method in a second embodiment;

[0050] Figure 4 A schematic diagram of a process flow of a material damage identification method in a third embodiment;

[0051] Figure 5 is a schematic flow chart of a material damage identification method in a fourth embodiment;

[0052] Figure 6 is a structural block diagram of a material damage identification device in one embodiment;

[0053] Figure 7 is a structural block diagram of a material damage identification device in another embodiment;

[0054] Figure 8 is a structural block diagram of a material damage identification device in yet another embodiment;

[0055] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0057] The material damage identification method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the laser light source 101, the attenuation system 102, and the lens 103 are located on the same optical axis. The laser emitted by the laser light source 101 is attenuated by the attenuation system 102 in turn, and then gathered on the test surface of the test material 105 through the lens 103. The scattered light generated by the laser on the test material 105 is collected by the CCD (Charge Coupled Device) camera 104, so that the CCD camera 104 can collect the scattered light spatial distribution image corresponding to the test material under laser irradiation. Further, the CCD camera 104 transmits the collected scattered light spatial distribution image to the computer device, in which the trained deep learning model is pre-loaded. The image features of the scattered light spatial distribution image are extracted by the trained deep learning model to obtain the deep features of the image; the deep features of the image are analyzed by the trained deep learning model, and the damage recognition results of each test point on the test material are output, and the damage recognition results include: damage or no damage.

[0058] It should be understood that the above Figure 1The illustrated embodiment only provides a possible position structure. In a more common case, the test material 105 is placed on the same optical axis of the lens and is directly in front of the lens. At this time, the CCD camera 104 is located on the side of the test material 105 close to the lens to collect the scattered light formed by the laser irradiation on the surface of the test material 105. Optionally, when the position of the test material 105 is not directly in front of the lens (for example, the surface of the test material 105 is not perpendicular to the optical axis of the lens, or the test material 105 is located in front of the side of the lens, etc.), the position of the CCD camera 104 can be adaptively adjusted at this time, so that the CCD camera 104 can fully collect the scattered light formed by the laser irradiation on the surface of the test material 105. That is, the setting position of the CCD camera 104 in the embodiment of the present application is flexible and adjustable. In addition, the embodiment of the present application does not limit the setting distance between the laser light source 101, the attenuation system 102, the lens 103, and the test material 105.

[0059] In a first exemplary embodiment, if Figure 2 As shown, a material damage identification method is provided, which is applied to Figure 1 The application scenario shown is used as an example for description, including the following steps 201 to 203. Among them:

[0060] Step 201 : collecting a scattered light spatial distribution image corresponding to the test material under laser irradiation.

[0061] In this embodiment, the spatial distribution image of the scattered light corresponding to the test material under laser irradiation can contain more high-dimensional features. These high-dimensional features can be used as a basis for identifying whether there is damage on the surface of the material. Compared with directly collecting surface images or observing with the naked eye, it has higher recognition accuracy.

[0062] Step 202: extract image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image.

[0063] The deep features of the image include at least one of the following:

[0064] Number, shape and size of light spots;

[0065] The distribution pattern of the light spot;

[0066] Texture details.

[0067] In this embodiment, the trained deep learning model can automatically extract the deep features of the scattered light spatial distribution image through multiple convolutional layers and pooling operations without human intervention. Based on these deep features, it can accurately identify whether there is damage on the material surface.

[0068] Step 203: Analyze the deep features of the image through the trained deep learning model, and output the damage identification results of each test point on the test material.

[0069] The damage identification result includes: damaged or not damaged.

[0070] In this embodiment, by analyzing the deep features of the image through the trained deep learning model, the spatial distribution information of the scattered light image can be fully utilized to realize the automation and detection of damage points, and the test time can be significantly shortened to improve the detection efficiency.

[0071] It should be noted that the deep learning model trained in this embodiment is obtained by training with a large amount of sample data, so it can maintain high recognition accuracy and stability under different experimental conditions (such as laser intensity fluctuations, ambient light changes). In other words, the deep learning model trained in this embodiment can adapt to different types of samples and experimental conditions, meet the actual needs of complex optical material performance evaluation, and has a wider range of applications.

[0072] In the above-mentioned material damage identification method, by collecting the scattered light spatial distribution image corresponding to the test material under laser irradiation; thus, the scattering result of the laser on the surface of the test material can be used as the basis for damage identification. Compared with naked eye observation and direct image acquisition, it can have more deep features, which is convenient for subsequent improvement of test accuracy. The image features of the scattered light spatial distribution image are extracted by a trained deep learning model to obtain the deep features of the image; thus, fully automatic image feature extraction can be achieved without human intervention. The deep features of the image are analyzed by a trained deep learning model, and the damage identification results of each test point on the test material are output, and the damage identification results include: damaged or undamaged. Therefore, the spatial distribution information of scattered light can be fully utilized to realize the automatic detection of material test points without human intervention, thereby improving the detection efficiency and significantly improving the recognition accuracy of material damage.

[0073] In a second exemplary embodiment, a material damage identification method is provided, wherein the method is applied to Figure 1 The application scenario shown is used as an example for description, including the following steps 301 to 304. Among them:

[0074] Step 301 : collecting a scattered light spatial distribution image corresponding to the test material under laser irradiation.

[0075] In this embodiment, the specific implementation process and technical effects of step 301 are shown in Figure 2 The description of step 201 in the illustrated method embodiment will not be repeated here.

[0076] Step 302: pre-process the scattered light spatial distribution image corresponding to the test material.

[0077] The preprocessing method includes: denoising processing and / or normalization processing.

[0078] In this embodiment, taking a CCD camera as an example, the collected scattered light spatial distribution image may introduce some environmental interference, resulting in poor image quality. At this time, the collected scattered light spatial distribution image can be denoised and / or normalized to improve the image quality, so that the subsequent model can better identify the damage results.

[0079] It should be noted that this embodiment does not limit the specific method of denoising. For example, spatial domain denoising methods (neighborhood averaging method, median filtering method, low-pass filtering method, mean filtering method, morphological denoising method, etc.), variable domain denoising methods, and spatial domain and transform domain collaborative filtering methods can all be applied in this embodiment.

[0080] Step 303: extract image features from the scattered light spatial distribution image using the trained deep learning model to obtain deep features of the image.

[0081] Step 304: Analyze the deep features of the image through the trained deep learning model and output the damage identification results of each test point on the test material.

[0082] In this embodiment, the specific implementation process and technical effects of steps 303 to 304 are shown in Figure 2 The descriptions of steps 202 to 203 in the illustrated method embodiment are not repeated here.

[0083] In this embodiment, by preprocessing the collected scattered light spatial distribution image, the image quality for subsequent feature extraction can be improved, thereby indirectly improving the accuracy of the trained deep learning model in identifying material damage.

[0084] In a third exemplary embodiment, a material damage identification method is provided, wherein the method is applied to Figure 1 The application scenario shown is used as an example for description, including the following steps 401 to 409. Among them:

[0085] Step 401 : collecting a scattered light spatial distribution image corresponding to the test material under laser irradiation.

[0086] In this embodiment, the specific implementation process and technical effects of step 401 are shown in Figure 2 The description of step 201 in the illustrated method embodiment will not be repeated here.

[0087] Step 402 , collecting scattered light spatial distribution images of the material under laser irradiation, and marking the collected scattered light spatial distribution images as damaged or undamaged to obtain a marked image.

[0088] In this embodiment, previous historical data, such as scattered light spatial distribution images obtained from a large number of previous test experiments, can be used as images to be processed, and then these scattered light spatial distribution images can be annotated manually.

[0089] For example, as the test is continuously carried out, the amount of historical data is also increasing, so these historical data can be screened to select representative scattered light spatial distribution images for annotation.

[0090] For example, in addition to manual labeling, automatic labeling can also be used, such as using a high-precision camera to capture images of the material surface, and then analyzing the images captured by the high-precision camera based on a visual model to output the labeling results.

[0091] It should be noted that, considering the accuracy and efficiency of annotation, manual annotation and automatic annotation can also be combined. For example, the automatic annotation method is used for annotation first, and then the manual sampling method is used for review and correction, so as to obtain a large number of annotated images.

[0092] Step 403: construct a training set and a test set based on the annotated images.

[0093] In this embodiment, the annotated images are split into two parts: one part is used to construct a training set, and the other part is used to construct a test set.

[0094] Step 404, construct an initial deep learning model.

[0095] Among them, the network structure of the initial deep learning model includes: convolutional layer, pooling layer, and fully connected layer.

[0096] In this embodiment, a classification model based on CNN (convolutional neural network) can be constructed as an initial deep learning model. Exemplarily, in order to improve the performance of the model, a pre-trained model (such as ResNet, EfficientNet) can also be used for transfer learning.

[0097] Step 405: train the initial deep learning model using the training set, calculate the classification error using the cross entropy loss function, and optimize the model parameters in the initial deep learning model using the back propagation algorithm to obtain a trained initial deep learning model.

[0098] In this embodiment, the scattered light spatial distribution image in the training set is used as the input of the initial deep learning model, and the output result of the initial deep learning model is compared with the pre-labeled result to determine whether the damage identification result of the initial deep learning model is correct. During the entire training process, the classification error can be calculated using the cross entropy loss function, and the model parameters in the initial deep learning model can be optimized by the back propagation algorithm to obtain a trained initial deep learning model.

[0099] Step 406: Verify the trained initial deep learning model through the test set to obtain a verification result.

[0100] In this embodiment, the scattered light spatial distribution image in the test set is used as the input of the trained initial deep learning model, and the output results of the trained initial deep learning model are compared with the pre-labeled results to determine whether the damage identification results of the trained deep learning model are correct.

[0101] Step 407: Adjust the model parameters of the trained initial deep learning model based on the verification result to obtain a trained deep learning model.

[0102] In this embodiment, during the entire verification process, the model parameters of the trained initial deep learning model can be continuously adjusted to avoid overfitting. Exemplarily, when the accuracy of the output result of the trained initial deep learning model exceeds a preset threshold, the corresponding model parameters are saved to obtain a trained deep learning model.

[0103] Step 408: extract image features from the scattered light spatial distribution image using the trained deep learning model to obtain deep features of the image.

[0104] Step 409: Analyze the deep features of the image through the trained deep learning model, and output the damage identification results of each test point on the test material.

[0105] In this embodiment, the specific implementation process and technical effects of steps 408 to 409 are shown in Figure 2 The descriptions of steps 202 to 203 in the illustrated method embodiment are not repeated here.

[0106] In this embodiment, an initial deep learning model is constructed, trained based on a training set, and verified by a test set, thereby obtaining a trained deep learning model. Based on the trained deep learning model, image features can be automatically extracted from scattered light spatial distribution images and the damage recognition results can be analyzed and output. The entire process does not require manual intervention and has high stability, greatly improving detection efficiency and recognition accuracy.

[0107] In a fourth exemplary embodiment, a material damage identification method is provided, wherein the method is applied to Figure 1 The application scenario shown is used as an example for description, including the following steps 501 to 505. Among them:

[0108] Step 501 : collecting a scattered light spatial distribution image corresponding to the test material under laser irradiation.

[0109] Step 502: extract image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image.

[0110] Step 503: Analyze the deep features of the image through the trained deep learning model, and output the damage identification results of each test point on the test material.

[0111] In this embodiment, the specific implementation process and technical effects of steps 501 to 503 are shown in Figure 2 The descriptions of steps 201 to 203 in the illustrated method embodiment are not repeated here.

[0112] Step 504, saving the damage identification result as the damage area feature image corresponding to each test point of the damage.

[0113] In this embodiment, the damaged area feature image corresponding to each test point with a damaged identification result is saved, so that during subsequent inspection, the staff can quickly retrieve the corresponding feature image for analysis.

[0114] Step 505: Generate a visual test report based on the damage identification results of each test point on the test material and the damage area feature image.

[0115] In this embodiment, a visual test report can be automatically generated based on the damage identification results of each test point on the test material and the characteristic image of the damaged area, so that the final test result is more intuitive and convenient for subsequent damage result analysis.

[0116] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0117] Based on the same inventive concept, the embodiment of the present application also provides a material damage identification device for implementing the material damage identification method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more material damage identification device embodiments provided below can refer to the limitations of the material damage identification method above, and will not be repeated here.

[0118] In an exemplary embodiment, Figure 6 As shown, a material damage identification device is provided, including: a data acquisition module 601, a feature extraction module 602 and a damage identification module 603, wherein:

[0119] The data acquisition module 601 is used to acquire the scattered light spatial distribution image corresponding to the test material under laser irradiation;

[0120] A feature extraction module 602 is used to extract image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image;

[0121] The damage identification module 603 is used to analyze the deep features of the image through a trained deep learning model, and output the damage identification results of each test point on the test material, and the damage identification results include: damaged or undamaged.

[0122] Exemplarily, the data acquisition module 601 is specifically used to: acquire a scattered light pattern generated on the surface of the test material under laser irradiation by a photosensitive coupling camera to obtain a scattered light spatial distribution image corresponding to the test material.

[0123] Exemplarily, the deep features of the image include at least one of the following:

[0124] Number, shape and size of light spots;

[0125] The distribution pattern of the light spot;

[0126] Texture details.

[0127] In this embodiment, by collecting the scattered light spatial distribution image corresponding to the test material under laser irradiation; thus, the scattering result of the laser on the surface of the test material can be used as the basis for damage identification. Compared with naked eye observation and direct image acquisition, it can have more deep features, which is convenient for improving the test accuracy in the future. The image features of the scattered light spatial distribution image are extracted by a trained deep learning model to obtain the deep features of the image; thus, fully automatic image feature extraction can be achieved without human intervention. The deep features of the image are analyzed by a trained deep learning model, and the damage identification results of each test point on the test material are output, and the damage identification results include: damaged or undamaged. Therefore, the spatial distribution information of scattered light can be fully utilized to realize the automatic detection of material test points without human intervention, thereby improving the detection efficiency and significantly improving the recognition accuracy of material damage.

[0128] In another exemplary embodiment, Figure 7 As shown, a material damage identification device is provided. Figure 6 Based on the device shown, it can also include: a preprocessing module 604, which is used to preprocess the scattered light spatial distribution image corresponding to the test material, wherein the preprocessing method includes: denoising processing and / or normalization processing.

[0129] Optionally, the above device may further include:

[0130] The marking module 605 is used to collect the scattered light spatial distribution image of the material under laser irradiation, and mark the collected scattered light spatial distribution image as damaged or undamaged to obtain a marked image;

[0131] A data set construction module 606 is used to construct a training set and a test set based on the annotated images;

[0132] A model building module 607 is used to build an initial deep learning model; wherein the network structure of the initial deep learning model includes: a convolutional layer, a pooling layer, and a fully connected layer;

[0133] The training module 608 is used to train the initial deep learning model through the training set, calculate the classification error using the cross entropy loss function, and optimize the model parameters in the initial deep learning model through the back propagation algorithm to obtain the trained initial deep learning model.

[0134] Optionally, the above device may further include:

[0135] The verification module 609 is used to verify the trained initial deep learning model through the test set to obtain a verification result; based on the verification result, the model parameters of the trained initial deep learning model are adjusted to obtain a trained deep learning model.

[0136] In this embodiment, an initial deep learning model is constructed, and training is performed based on a training set, and verification is performed through a test set, thereby obtaining a trained deep learning model. Based on the trained deep learning model, image features can be automatically extracted from scattered light spatial distribution images and the damage recognition results can be analyzed and output. The entire process does not require manual intervention, has high stability, and greatly improves detection efficiency and recognition accuracy.

[0137] In yet another exemplary embodiment, Figure 8 As shown, a material damage identification device is provided. Figure 6 Based on the device shown, it can also include:

[0138] The storage module 610 is used to store the damage area characteristic image corresponding to each test point whose damage identification result is damage;

[0139] The test report generation module 611 is used to generate a visual test report based on the damage identification results of each test point on the test material and the damage area feature image.

[0140] In this embodiment, a visual test report can be automatically generated based on the damage identification results of each test point on the test material and the characteristic image of the damage area, so that the final test result is more intuitive and convenient for subsequent damage result analysis.

[0141] Each module in the above-mentioned material damage identification device can be implemented in whole or in part by software, hardware or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0142] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Fig. 9As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a material damage identification method is realized. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0143] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0144] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0145] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0146] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0147] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0148] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0149] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0150] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A material damage identification method, characterized in that: The method comprises: Collect the scattered light spatial distribution image corresponding to the test material under laser irradiation; Extracting image features from the scattered light spatial distribution image using a trained deep learning model to obtain deep features of the image; The deep features of the image are analyzed by a trained deep learning model, and damage identification results of each test point on the test material are output, where the damage identification results include: damaged or undamaged.

2. The method according to claim 1, characterized in that The collecting of the scattered light spatial distribution image corresponding to the test material under laser irradiation includes: The scattered light pattern generated on the surface of the test material under laser irradiation is collected by a photosensitive coupled camera to obtain a scattered light spatial distribution image corresponding to the test material.

3. The method according to claim 1, characterized in that The deep features of the image include at least one of the following: Number, shape and size of light spots; The distribution pattern of the light spot; Texture details.

4. The method according to claim 1, characterized in that: Before extracting image features from the scattered light spatial distribution image using the trained deep learning model, the method further includes: The scattered light spatial distribution image corresponding to the test material is preprocessed, wherein the preprocessing method includes: denoising processing and / or normalization processing.

5. The method according to any one of claims 1 to 4, characterized in that: Before extracting image features from the scattered light spatial distribution image using the trained deep learning model, the method further includes: Collecting a scattered light spatial distribution image of the material under laser irradiation, and marking the collected scattered light spatial distribution image as damaged or undamaged to obtain a marked image; Constructing a training set and a test set based on the annotated images; Constructing an initial deep learning model; wherein the network structure of the initial deep learning model includes: a convolutional layer, a pooling layer, and a fully connected layer; The initial deep learning model is trained using the training set, the classification error is calculated using a cross entropy loss function, and the model parameters in the initial deep learning model are optimized using a back propagation algorithm to obtain the trained initial deep learning model.

6. The method according to claim 5, characterized in that After obtaining the trained initial deep learning model, the method further includes: Verifying the trained initial deep learning model through the test set to obtain a verification result; Based on the verification result, the model parameters of the trained initial deep learning model are adjusted to obtain a trained deep learning model.

7. The method according to any one of claims 1 to 4, characterized in that: After outputting the damage identification results of each test point on the test material, the method further includes: The damage identification result is saved as a damage area feature image corresponding to each test point of the damage; A visual test report is generated based on the damage identification results of each test point on the test material and the characteristic image of the damaged area.

8. A material damage identification device, characterized in that: The device comprises: A data acquisition module, used to acquire a spatial distribution image of scattered light corresponding to the test material under laser irradiation; A feature extraction module is used to extract image features from the scattered light spatial distribution image through a trained deep learning model to obtain deep features of the image; The damage identification module is used to analyze the deep features of the image through a trained deep learning model, and output the damage identification results of each test point on the test material, and the damage identification results include: damaged or undamaged.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

11. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 7.

Images