Crack detection positioning method and system, electronic equipment and storage medium

Abstract

The invention discloses a crack detection positioning method and system, electronic equipment and a storage medium. The method comprises the following steps: acquiring an image and temperature of the surface of a to-be-detected workpiece and structural acoustic data of the to-be-detected workpiece; processing the acquired image and temperature of the surface of the to-be-measured workpiece and the acoustic data of the structure of the to-be-measured workpiece; constructing a deep neural network model and a self-adaptive learning model, performing feature extraction on the processed image and temperature of the surface of the to-be-detected workpiece and acoustic data of the structure of the to-be-detected workpiece by using the constructed deep neural network model, and detecting the extracted features by using the self-adaptive learning model. Obtaining a crack detection result of the to-be-detected workpiece; the augmented reality technology is utilized to map the crack detection result in real time, rapid positioning of the crack of the to-be-detected workpiece is realized, and the technical defects that comprehensive data collection of the blade is difficult and the crack position cannot be accurately positioned by adopting a single type of sensor are overcome.

Description

Technical Field

[0001] The present invention belongs to the technical field of crack detection and positioning, and in particular relates to a crack detection and positioning method, system, electronic equipment and storage medium. Background Art

[0002] As a clean energy source, offshore wind power requires crucial maintenance. During operation, wind turbine blades are subject to a variety of external factors, including wind and seawater corrosion, which can easily cause cracks. Crack detection and location are key technologies in wind turbine blade maintenance, involving monitoring both the blade surface and internal structure to ensure safe and efficient operation of wind turbines.

[0003] At present, the detection of cracks in wind turbine blades is mainly based on the detection of a single type of sensor or relies on regular manual inspections. Although a single type of sensor can automatically collect data on certain aspects of the blade, it is often limited by the single type of sensor, sensor performance and environmental factors. In the complex offshore environment, it is difficult to collect comprehensive data on the blade and cannot accurately locate the position of the crack. During the regular manual inspection process, affected by the blade structure and environmental factors, not only is the labor intensity high, but the efficiency is also low, and it is difficult to cover the entire area of the blade.

[0004] Finally, existing technologies for detecting and locating cracks in offshore wind turbine blades cannot effectively predict the development trend of cracks, resulting in longer response times for maintenance work, increased risks and maintenance costs. Summary of the Invention

[0005] The purpose of the present invention is to provide a crack detection and positioning method, system, electronic device and storage medium. On the one hand, it is used to solve the technical defects that existing detection methods are difficult to collect comprehensive data on blades and cannot accurately locate the position of cracks; on the other hand, it is used to solve the technical defects that manual regular inspections are labor-intensive, inefficient, and difficult to cover the entire area of the blade; on the third hand, it is used to solve the technical defects that existing detection methods cannot effectively predict the development trend of cracks, resulting in a long response time for maintenance work, increased risks and maintenance costs.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, a crack detection and positioning method is provided, comprising:

[0008] Acquire the image, temperature and structural acoustic data of the workpiece to be measured;

[0009] Processing the acquired image, temperature and structural acoustic data of the workpiece to be measured;

[0010] Constructing a deep neural network model and an adaptive learning model, using the constructed deep neural network model to extract features from the processed image, temperature, and structural acoustic data of the workpiece to be tested, using the adaptive learning model to detect the extracted features, and obtaining a crack detection result of the workpiece to be tested;

[0011] By using augmented reality technology, the crack detection results are mapped in real time to achieve rapid positioning of cracks in the workpiece to be tested.

[0012] Furthermore, the step of obtaining the image, temperature and structural acoustic data of the workpiece surface to be measured specifically includes:

[0013] A high-resolution camera, infrared sensor and acoustic wave detector are mounted on a drone, and the drone is flown above the workpiece to be tested. The image, temperature and acoustic data of the workpiece structure are acquired along a preset detection route.

[0014] Furthermore, the acquired image, temperature and structural acoustic data of the workpiece to be measured are processed, specifically including:

[0015] De-noising the acquired surface image data of the workpiece to be tested to eliminate interference caused by wind and rain. High-pass filtering and histogram equalization are used to enhance the visibility of cracks in the surface image of the workpiece to be tested and reduce image quality fluctuations caused by lighting changes.

[0016] The surface temperature data of the workpiece to be measured and the structural acoustic data of the workpiece to be measured are cleaned to improve the accuracy of the surface temperature data of the workpiece to be measured and the structural acoustic data of the workpiece to be measured.

[0017] Furthermore, feature extraction is performed on the processed image, temperature and workpiece structure acoustic data of the surface of the workpiece to be measured, specifically:

[0018] The processed surface image, temperature and structural acoustic data of the workpiece to be measured are simultaneously input into the neural network model to form multi-source data, and joint feature extraction is performed from the multi-source data.

[0019] Furthermore, the extracted features are detected using the adaptive learning model, and crack detection results of the workpiece to be tested are obtained, specifically:

[0020] The adaptive learning model automatically adjusts model parameters based on real-time environmental data, performs crack detection on the extracted features, and achieves accurate crack identification in various environments.

[0021] Furthermore, the crack detection results are mapped in real time using augmented reality technology to achieve rapid location of cracks in the workpiece to be tested, specifically including:

[0022] Augmented reality technology is used to map the crack detection results in the operator's field of view in real time, assisting the operator to quickly locate the crack position.

[0023] Furthermore, it also includes:

[0024] Build a predictive maintenance model that combines historical crack data of the workpiece under test with machine learning algorithms to analyze crack development patterns and predict future maintenance needs of the workpiece under test;

[0025] An interactive platform is constructed, and a user interface on the interactive platform is used to display real-time progress of crack detection of the workpiece to be tested, historical crack data, prediction results and early warning information to the operator.

[0026] In a second aspect, a crack detection and positioning system is provided, comprising:

[0027] A data acquisition module is used to obtain the image, temperature and structural acoustic data of the workpiece surface to be measured;

[0028] A data processing module, configured to process the acquired image, temperature and workpiece structural acoustic data of the workpiece surface to be measured;

[0029] Model building module for building deep neural network models, adaptive learning models, and predictive maintenance models;

[0030] Positioning module, used for rapid positioning of cracks in the workpiece to be tested;

[0031] Predictive maintenance module, used to predict future maintenance needs of the workpiece to be tested;

[0032] The interactive module is used to show the operator the real-time progress of crack detection of the workpiece to be tested, historical crack data, prediction results and early warning information.

[0033] In a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of the crack detection and positioning method as described above when executing the computer program.

[0034] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the crack detection and positioning method as described above is implemented.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. By acquiring the image, temperature and structural acoustic data of the surface of the workpiece to be tested, comprehensive data collection can be performed on the blade, and subtle changes in the workpiece to be tested at different physical levels can be captured, which helps to more comprehensively evaluate the state of the workpiece to be tested, reduce misjudgments or missed detections that may be caused by a single data source, and thus improve the accuracy of crack detection. Secondly, the use of deep neural network models for feature extraction can automatically learn and identify complex feature patterns related to cracks, which are often difficult for the human eye to directly perceive. At the same time, the adaptive learning model can continuously optimize the detection algorithm based on real-time data, further improving the sensitivity and accuracy of detection; finally, the use of augmented reality technology can quickly locate the position of the crack, making it easier for operators to find the crack location and perform timely maintenance, solving the technical defects of using a single type of sensor that is difficult to collect comprehensive data on the blade and cannot accurately locate the crack location.

[0037] 2. Using drones equipped with high-resolution cameras, infrared sensors, and acoustic wave detectors to inspect blades according to preset inspection routes effectively solves the technical problem of manual regular inspections, which are affected by blade structure and environmental factors, resulting in high labor intensity, low efficiency, and difficulty in covering the entire blade area. This realizes automated inspection operations and improves the efficiency of crack detection and positioning.

[0038] 3. Image denoising effectively eliminates interference caused by external environmental factors and internal system noise, thereby improving data accuracy. For image data, denoising and enhancement reduce the impact of lighting variations on image quality, making subtle features such as cracks more visible. For temperature and acoustic data, de-noising removes irrelevant signals, improving data purity and reliability.

[0039] 4. Joint feature extraction of multi-source data can make full use of the complementary information in image, temperature and acoustic data, help identify local temperature changes caused by cracks, and form a more comprehensive and rich feature expression, thereby improving the accuracy of crack detection.

[0040] 5. The adaptive learning model can automatically adjust its internal parameters according to real-time environmental data to adapt to different detection environments. This dynamic adjustment capability enables the model to maintain stable performance in complex and changing environments and improve the accuracy of crack detection.

[0041] 6. By directly superimposing the crack detection results in the operator's field of view, the specific location and shape of the crack can be clearly seen, thereby greatly improving the positioning speed and accuracy.

[0042] 7. By building a predictive maintenance model and combining the historical crack data of the workpiece to be tested with machine learning algorithms to analyze the crack development pattern, it is possible to effectively predict the future crack development trend changes of the workpiece to be tested. Based on the trend changes, future maintenance needs can be predicted more accurately, maintenance technology can be prepared in advance, and the response time, risk and maintenance cost of future maintenance work can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 A flow chart of a crack detection and positioning method provided by the present invention;

[0045] Figure 2 This is a schematic diagram of the principle of a crack detection and positioning system provided by the present invention. DETAILED DESCRIPTION

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0047] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0048] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0049] In the description of the embodiments of the present invention, it should be noted that if the terms "upper," "lower," "horizontal," "inner," etc. appear, the orientation or positional relationship indicated is based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the inventive product is typically placed when in use. These terms are merely for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. In addition, the terms "first," "second," etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0050] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0051] In the description of the embodiments of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0052] As a clean energy source, offshore wind power requires crucial maintenance. During operation, wind turbine blades are subject to a variety of external factors, including wind and seawater corrosion, which can easily cause cracks. Crack detection and location are key technologies in wind turbine blade maintenance, involving monitoring both the blade surface and internal structure to ensure safe and efficient operation of wind turbines.

[0053] At present, the detection of cracks in wind turbine blades is mainly based on the detection of a single type of sensor or relies on regular manual inspections. Although a single type of sensor can automatically collect data on certain aspects of the blade, it is often limited by the single type of sensor, sensor performance and environmental factors. In the complex offshore environment, it is difficult to collect comprehensive data on the blade and cannot accurately locate the position of the crack. During the regular manual inspection process, affected by the blade structure and environmental factors, not only is the labor intensity high, but the efficiency is also low, and it is difficult to cover the entire area of the blade.

[0054] Finally, existing technologies for detecting and locating cracks in offshore wind turbine blades cannot effectively predict the development trend of cracks, resulting in longer response times for maintenance work, increased risks and maintenance costs.

[0055] In order to solve the above technical defects, the inventor provides a crack detection and positioning method, system, electronic device and storage medium.

[0056] The present invention will be described in further detail below with reference to the accompanying drawings, taking wind turbine blade crack detection and positioning as an example:

[0057] In a first aspect, an embodiment of the present invention provides a crack detection and positioning method, such as Figure 1 As shown, including:

[0058] S101. Obtain the image, temperature and structural acoustic data of the workpiece to be measured on the surface. For example, since a single type of sensor cannot collect comprehensive data on the workpiece to be measured, i.e., the blade, this method analyzes the working environment of the blade and considers it from different physical levels during the process of blade crack detection. Then, the image, temperature and structural acoustic data of the blade surface to be measured are obtained first, thereby realizing comprehensive data collection on the blade. This helps to more comprehensively evaluate the status of the workpiece to be measured, reduce misjudgments or missed detections that may be caused by a single data source, thereby improving the accuracy of crack detection, and solving the technical defects that it is difficult to collect comprehensive data on the blade and accurately locate the crack position using a single type of sensor. Secondly, in the process of blade crack detection, priority is given to the use of drones. The drones are equipped with high-resolution cameras, infrared sensors and acoustic wave detectors. The drones fly above the blades to be tested and obtain images, temperatures and acoustic data of the blade structure on the surface according to the preset detection route. The operator only needs to stand on the ground or remotely control the drone through a controller, realizing automated detection operations. This effectively solves the technical problem that during manual regular inspections, the labor intensity is high, the efficiency is low, and it is difficult to cover the entire area of the blade due to the influence of blade structure and environmental factors, thereby improving the efficiency of crack detection and positioning.

[0059] S102: Processing the acquired image, temperature, and structural acoustic data of the workpiece surface. For example, since the drone may be affected by rain and wind when acquiring blade surface images, causing the drone to shake and thus interfere with the acquired image, the blade surface image is first processed by denoising to eliminate interference caused by wind and rain. High-pass filtering and histogram equalization are then performed to enhance the visibility of cracks in the workpiece surface image and reduce image quality fluctuations caused by illumination variations. Next, the workpiece surface temperature data and structural acoustic data are cleaned to improve their accuracy. During the processing, the image denoising step effectively eliminates interference caused by external environmental factors and internal system noise, thereby improving data accuracy. For image data, denoising and enhancement reduce the impact of illumination variations on image quality, making subtle features such as cracks more clearly visible. For temperature and acoustic data, cleanup removes irrelevant signals, improving data purity and reliability.

[0060] S103. Construct a deep neural network model and an adaptive learning model. Use the constructed deep neural network model to extract features from the processed surface image, temperature, and structural acoustic data of the workpiece to be tested. Use the adaptive learning model to detect the extracted features and obtain crack detection results for the workpiece to be tested. For example, a convolutional neural network (CNN) is used as the main architecture of the deep neural network. CNNs have excellent performance in image feature extraction and can effectively capture crack-related features such as texture and edges from image data. The CNN network structure includes basic components such as convolutional layers, pooling layers, and fully connected layers. The convolutional layer extracts local features by sliding convolution kernels across the image and extracting features. The pooling layer reduces and abstracts the features, enhancing their invariance. The fully connected layer performs comprehensive classification of the extracted high-level features. Furthermore, an attention mechanism is introduced based on the CNN, enabling the model to adaptively focus on more representative feature regions, thereby further improving detection accuracy. During the construction process, a transfer learning approach was employed. This approach uses a CNN model pre-trained on a large-scale dataset as initialization, and then fine-tunes it on data from the specific application scenario. This approach fully utilizes existing network parameters and feature extraction capabilities, accelerates model convergence, and reduces the amount of data required for training. During the construction of the adaptive learning model, an adaptive learning model based on transfer learning and reinforcement learning was employed for crack detection. This adaptive model has the following characteristics: It uses general object detection models pre-trained on large-scale datasets, such as YOLO and Faster R-CNN, as initialization models. These models have already learned relatively general feature representations and can quickly adapt to new object detection tasks. Building on transfer learning, reinforcement learning is employed to continuously optimize the model. Specifically, an intelligent agent is designed that adjusts the model's parameters and hyperparameters based on environmental feedback (such as detection accuracy and speed), enabling the model to better adapt to complex offshore environments. The model then continuously collects new blade data during operation and uses this data to further optimize itself. For example, through active learning strategies, the model can proactively query manually labeled "important" samples to update itself. Simultaneously, through transfer learning, the model can also learn from accumulated experience in different environments. Furthermore, reinforcement learning strategies make the model highly robust, enabling it to maintain high detection accuracy across diverse environments and conditions. Furthermore, continuous self-optimization enables the model to adapt to dynamic environmental changes. In summary, a reinforcement learning algorithm based on a deep Q-network (DQN) was employed to construct an adaptive learning model. This algorithm learns optimal decision-making strategies through interaction with the environment, enabling the model to autonomously adjust parameters to adapt to different detection scenarios.After the two models are constructed, the processed images, temperatures, and acoustic data of the blade surface to be tested are simultaneously fed into the neural network model, generating multi-source data. Joint feature extraction is then performed from this multi-source data. The adaptive learning model then automatically adjusts model parameters based on real-time environmental data, performing crack detection on the extracted features. This enables accurate crack identification in various environments and yields crack detection results. This end-to-end feature extraction approach better mines correlations between multi-source data, avoiding information loss. Furthermore, this designed deep neural network model is also easier to train and optimize end-to-end, improving overall performance.

[0061] S104. Using augmented reality technology, the crack detection results are mapped in real time to achieve rapid crack location in the workpiece to be tested. For example, the crack detection results are mapped in real time in the operator's field of view using augmented reality technology. By directly superimposing the crack detection results in the operator's field of view, the specific location and morphology of the crack can be clearly seen, thereby greatly improving positioning speed and accuracy, and assisting the operator in quickly locating the crack.

[0062] In addition, to effectively predict crack development trends, a predictive maintenance model is first constructed. This model combines historical crack data from the workpiece under test with machine learning algorithms to analyze crack development patterns and predict future maintenance needs for the workpiece under test. An interactive platform is then constructed, and its user interface displays the operator with real-time progress, historical crack data, prediction results, and early warning information on the workpiece under test. By constructing a predictive maintenance model and analyzing crack development patterns based on historical crack data from the workpiece under test and machine learning algorithms, future crack development trends can be effectively predicted. Based on these trends, future maintenance needs can be more accurately predicted, enabling maintenance techniques to be implemented in advance and reducing response time, risk, and maintenance costs for future maintenance work.

[0063] In the process of using the blade crack detection method, the following implementation methods may also be considered:

[0064] In terms of data collection, in addition to drones, sensor networks composed of various types of sensors mounted on wind turbine towers, or sensors mounted on ground vehicles, can also be used to collect data to suit different operating environments and cost requirements. After data collection is complete, various image processing algorithms, such as edge detection and texture analysis, can be combined to improve the accuracy and robustness of crack identification. For deep learning optimization of feature extraction, different deep learning models, such as recurrent neural networks (RNNs) or variational autoencoders (VAEs), can be employed to better handle time series data or complex image features. These models have been improved in the following ways: For example, for RNN models processing time series data, their network structure and hyperparameters can be adjusted to the characteristics of wind turbine blade monitoring data to better capture time-dependent features. For VAE models used to extract complex image features, transfer learning can be used to further fine-tune and optimize the general VAE model to better suit the characteristics of wind turbine blade images. Different types of deep learning models (such as CNN, RNN, VAE, etc.) can also be integrated to fully utilize the advantages of multi-source heterogeneous data and improve the overall feature extraction capability.

[0065] Among ensemble methods for crack detection, ensemble learning methods, such as bagging or boosting, can be used to combine the prediction results of multiple models (how exactly are they combined?). In bagging, multiple base learners (such as neural networks or decision trees) are trained, each trained on a random subset of the original training set. During prediction, the outputs of these base learners are voted or averaged to produce the final prediction. This method can improve the model's stability and generalization ability, as the differences between the base learners can compensate for each other's shortcomings. In boosting, multiple weak learners are chained together to gradually improve the performance of the overall model. Specifically, an initial base learner is trained, and then the sample weights are adjusted based on its predictions, focusing on previously misclassified samples. The next base learner is then trained, and so on, until a satisfactory accuracy is achieved. Boosting excels at handling complex patterns that are difficult to learn and can significantly improve detection accuracy. In practical implementation, bagging and boosting can be combined: first, multiple base learners are trained using bagging, and then, using the boosting approach, these base learners are chained together to further improve overall performance. In addition, the selection and combination of base learners can also be experimented with and optimized to achieve the best detection results. For example, different types of neural network models can be used as base learners, or their outputs can be combined using weighted averaging or other methods.

[0066] During feature extraction, we first use the aforementioned deep neural network model to extract features from the preprocessed data. The feature vectors generated in this step serve as input for subsequent crack detection. We then use ensemble learning algorithms such as bagging and boosting on these feature vectors to construct multiple crack detection models. These models can be different types of machine learning models, such as neural networks, decision trees, and support vector machines. During prediction, we combine the outputs of these basic models, for example, through voting or weighted averaging. This ensemble approach further improves final detection accuracy and reliability. It leverages the strengths of existing algorithms while also customizing and optimizing them to enhance detection accuracy and reliability.

[0067] like Figure 2 As shown, in a second aspect, a crack detection and positioning system is provided, comprising:

[0068] A data acquisition module is used to obtain the image, temperature and structural acoustic data of the workpiece surface to be measured;

[0069] A data processing module, configured to process the acquired image, temperature and workpiece structural acoustic data of the workpiece surface to be measured;

[0070] Model building module for building deep neural network models, adaptive learning models, and predictive maintenance models;

[0071] Positioning module, used for rapid positioning of cracks in the workpiece to be tested;

[0072] Predictive maintenance module, used to predict future maintenance needs of the workpiece to be tested;

[0073] The interactive module is used to show the operator the real-time progress of crack detection of the workpiece to be tested, historical crack data, prediction results and early warning information.

[0074] In a third aspect, an embodiment of the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of the crack detection and positioning method as described above when executing the computer program.

[0075] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the crack detection and positioning method as described above is implemented.

[0076] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be considered as equivalent replacement methods and are included in the scope of protection of the present invention.

[0077] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0078] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0079] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0080] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0081] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A crack detection and positioning method, characterized in that: include: Acquire the image, temperature and structural acoustic data of the workpiece to be measured; Processing the acquired image, temperature and structural acoustic data of the workpiece to be measured; Constructing a deep neural network model and an adaptive learning model, using the constructed deep neural network model to extract features from the processed image, temperature, and structural acoustic data of the workpiece to be tested, using the adaptive learning model to detect the extracted features, and obtaining a crack detection result of the workpiece to be tested; By using augmented reality technology, the crack detection results are mapped in real time to achieve rapid positioning of cracks in the workpiece to be tested.

2. The detection and positioning method according to claim 1, characterized in that: The step of obtaining the image, temperature and structural acoustic data of the workpiece surface to be measured specifically includes: A high-resolution camera, infrared sensor and acoustic wave detector are mounted on a drone, and the drone is flown above the workpiece to be tested. The image, temperature and acoustic data of the workpiece structure are acquired along a preset detection route.

3. The detection and positioning method according to claim 1, characterized in that: The acquired image, temperature and structural acoustic data of the workpiece to be measured are processed, specifically including: De-noising the acquired surface image data of the workpiece to be tested to eliminate interference caused by wind and rain. High-pass filtering and histogram equalization are used to enhance the visibility of cracks in the surface image of the workpiece to be tested and reduce image quality fluctuations caused by lighting changes. The surface temperature data of the workpiece to be measured and the structural acoustic data of the workpiece to be measured are cleaned to improve the accuracy of the surface temperature data of the workpiece to be measured and the structural acoustic data of the workpiece to be measured.

4. The detection and positioning method according to claim 1, characterized in that: Feature extraction is performed on the processed image, temperature and workpiece structure acoustic data of the workpiece surface to be measured, specifically: The processed surface image, temperature and structural acoustic data of the workpiece to be measured are simultaneously input into the neural network model to form multi-source data, and joint feature extraction is performed from the multi-source data.

5. The detection and positioning method according to claim 1, characterized in that: The extracted features are detected using the adaptive learning model, and crack detection results of the workpiece to be tested are obtained, specifically: The adaptive learning model automatically adjusts model parameters based on real-time environmental data, performs crack detection on the extracted features, and achieves accurate crack identification in various environments.

6. The detection and positioning method according to claim 1, characterized in that: Using augmented reality technology, the crack detection results are mapped in real time to achieve rapid location of cracks in the workpiece to be tested, specifically including: Augmented reality technology is used to map the crack detection results in the operator's field of view in real time, assisting the operator to quickly locate the crack position.

7. The detection and positioning method according to claim 1, characterized in that: Also includes: Build a predictive maintenance model that combines historical crack data of the workpiece under test with machine learning algorithms to analyze crack development patterns and predict future maintenance needs of the workpiece under test; An interactive platform is constructed, and a user interface on the interactive platform is used to display real-time progress of crack detection of the workpiece to be tested, historical crack data, prediction results and early warning information to the operator.

8. A crack detection and positioning system, characterized in that: include: A data acquisition module is used to obtain the image, temperature and structural acoustic data of the workpiece surface to be measured; A data processing module, configured to process the acquired image, temperature and workpiece structural acoustic data of the workpiece surface to be measured; Model building module for building deep neural network models, adaptive learning models, and predictive maintenance models; Positioning module, used for rapid positioning of cracks in the workpiece to be tested; Predictive maintenance module, used to predict future maintenance needs of the workpiece to be tested; The interactive module is used to show the operator the real-time progress of crack detection of the workpiece to be tested, historical crack data, prediction results and early warning information.

9. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of the crack detection and positioning method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the crack detection and positioning method according to any one of claims 1 to 7 is implemented.

Images