Equipment anticorrosive coating falling identification system and method thereof

Abstract

The embodiment of the invention relates to the technical field of equipment anti-corrosion coating falling identification, and provides an equipment anti-corrosion coating falling identification system and method, and the method comprises the steps: obtaining a reflected sound wave signal, collected by a sound wave receiver, of an anti-corrosion coating of to-be-detected equipment; performing data preprocessing on the reflected sound wave signal to obtain a sequence of coating reflected sound wave signal fragments; performing feature extraction and adaptive reinforcement on the sequence of the coating reflection sound wave signal fragments to obtain an adaptive reinforcement coating reflection sound wave waveform feature map; and determining whether the shedding degree of the anti-corrosion coating of the to-be-detected equipment exceeds a preset requirement or not based on the self-adaptive reinforced coating reflection sound wave waveform characteristic pattern. According to the embodiment of the invention, the state of the anticorrosive coating of the equipment can be monitored and recognized in time, the influence of human subjective factors is reduced, the accuracy and reliability of anticorrosive coating recognition are improved, the automation of anticorrosive coating falling recognition is also realized, the recognition efficiency is greatly improved, and time and human resources are saved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of equipment anti-corrosion coating shedding identification, and in particular to an equipment anti-corrosion coating shedding identification system and method. Background Art

[0002] Corrosion protection coatings are applied to metal surfaces to protect them from corrosion and oxidation. When metals are exposed to harsh environmental conditions such as high humidity, salt spray, and chemicals, they are susceptible to corrosion, leading to damage and reduced functionality. The primary function of an anti-corrosion coating is to form a protective barrier on the metal surface, preventing direct contact between corrosive substances like water, oxygen, and salt, thereby extending the lifespan and safety of the equipment and reducing maintenance cycles.

[0003] Anti-corrosion coatings typically consist of multiple components, including resins, pigments, solvents, and additives. Resins are the primary component of anti-corrosion coatings, providing adhesion and corrosion resistance. Pigments provide the coating's color and appearance, while also enhancing its durability and UV resistance. Solvents adjust the coating's viscosity and fluidity, ensuring uniform application on the metal surface. Additives enhance the coating's properties, such as UV resistance, abrasion resistance, and scratch resistance.

[0004] The selection and application of anti-corrosion coatings require consideration of multiple factors, including environmental conditions, metal materials, and coating thickness. Different environmental conditions and metal materials place varying demands on anti-corrosion coatings. For example, in marine environments, salt spray and humidity accelerate metal corrosion, necessitating the selection of an anti-corrosion coating with high resistance to salt spray and moisture. Coating thickness is also a crucial consideration; a coating that is too thin may not provide adequate protection, while a coating that is too thick may cause cracking and flaking.

[0005] In wind farms and wind turbines, equipment is often exposed to harsh environments such as high temperature, high humidity, strong winds, and salt spray. Therefore, regular inspections are necessary to ensure the integrity of the equipment's anti-corrosion coatings and to check for signs of peeling or damage. Currently, the identification of equipment anti-corrosion coatings relies primarily on manual inspections, whereby the user visually observes the equipment's surface for abnormalities, such as color and gloss, and detects cracks. However, this method is highly subjective, and the results of manual inspections are susceptible to individual judgment, resulting in varying judgments between observers. Furthermore, the results of this method are susceptible to the influence of the observer's technical level and experience. This is because accurately determining whether an anti-corrosion coating is abnormal requires a certain level of technical expertise and experience, and for inexperienced or novice observers, misjudgments or omissions may occur. Furthermore, manual inspections require significant time and human resources, and cannot efficiently cover large equipment surfaces. Manual inspections are often inefficient for large equipment or large coating surfaces. Summary of the Invention

[0006] The present disclosure aims to solve at least one of the problems existing in the prior art and provides a system and method for identifying equipment anti-corrosion coating shedding.

[0007] In one aspect of the present disclosure, a method for identifying equipment anti-corrosion coating shedding is provided, the method comprising:

[0008] Acquiring a reflected acoustic wave signal of the anti-corrosion coating of the equipment to be tested collected by an acoustic wave receiver;

[0009] performing data preprocessing on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments;

[0010] Performing feature extraction and adaptive enhancement on a sequence of the coating reflected acoustic wave signal segments to obtain an adaptively enhanced coating reflected acoustic wave waveform feature map;

[0011] Based on the waveform characteristic diagram of the acoustic wave reflected by the adaptive strengthening coating, it is determined whether the degree of shedding of the anti-corrosion coating of the equipment to be inspected exceeds a predetermined requirement.

[0012] Optionally, performing data preprocessing on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments includes:

[0013] The reflected acoustic wave signal is image-blocked to obtain a sequence of segments of the coating reflected acoustic wave signal.

[0014] Optionally, feature extraction and adaptive enhancement are performed on the sequence of the coating reflected acoustic wave signal segments to obtain an adaptively enhanced coating reflected acoustic wave waveform feature map, including:

[0015] Using a deep learning network model, waveform feature extraction is performed on the sequence of coating reflected acoustic wave signal segments to obtain a sequence of coating reflected acoustic wave waveform feature maps;

[0016] The sequence of the coating reflected sound wave waveform characteristic graphs is passed through a region separation adaptive attention layer to obtain the adaptive enhanced coating reflected sound wave waveform characteristic graph.

[0017] Optionally, the deep learning network model is a coating reflected sound wave waveform feature extractor established based on a convolutional neural network model; the coating reflected sound wave waveform feature extractor includes: an input layer, a convolution layer, a pooling layer, an activation layer and an output layer.

[0018] Optionally, a deep learning network model is used to extract waveform features of the sequence of coating reflected acoustic wave signal segments to obtain a sequence of coating reflected acoustic wave waveform feature maps, including:

[0019] The sequence of the coating reflected sound wave signal segments is passed through the coating reflected sound wave waveform feature extractor to obtain a sequence of the coating reflected sound wave waveform feature graphs.

[0020] Optionally, the sequence of the coating reflected acoustic wave waveform characteristic graphs is passed through a region separation adaptive attention layer to obtain the adaptive enhanced coating reflected acoustic wave waveform characteristic graph, comprising:

[0021] performing a convolution operation on the sequence of waveform characteristic graphs of the acoustic wave reflected by the coating to obtain a sequence of waveform characteristic graphs of the acoustic wave reflected by the separation coating;

[0022] splicing the sequence of the separation coating layer reflected acoustic wave waveform characteristic graphs into a separation coating layer reflected acoustic wave waveform characteristic graph;

[0023] Performing a single-channel convolution operation on a sequence of waveform characteristic graphs of acoustic waves reflected by the coating to obtain an adaptive attention parameter matrix;

[0024] The separation coating reflected sound wave waveform characteristic diagram and the adaptive attention parameter matrix are subjected to sub-region attention learning to obtain the adaptive strengthening coating reflected sound wave waveform characteristic diagram.

[0025] Optionally, performing sub-region attention learning on the separation coating reflected sound wave waveform characteristic map and the adaptive attention parameter matrix to obtain the adaptive strengthening coating reflected sound wave waveform characteristic map includes:

[0026] The adaptive strengthening coating reflected sound wave waveform characteristic graph is obtained by performing a dot multiplication of each separation coating reflected sound wave waveform characteristic matrix in the separation coating reflected sound wave waveform characteristic graph and the adaptive attention parameter matrix.

[0027] Optionally, determining whether the degree of peeling of the anti-corrosion coating of the equipment to be inspected exceeds a predetermined requirement based on the waveform characteristic diagram of the acoustic wave reflected by the adaptive strengthening coating includes:

[0028] Optimizing the characteristic distribution of the waveform characteristic diagram of the reflected sound wave of the adaptive strengthening coating to obtain an optimized waveform characteristic diagram of the reflected sound wave of the adaptive strengthening coating;

[0029] The optimized adaptive strengthening coating reflected acoustic wave waveform characteristic diagram is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the degree of shedding of the anti-corrosion coating of the equipment to be tested exceeds the predetermined requirement.

[0030] Another aspect of the present disclosure provides a system for identifying equipment anti-corrosion coating shedding, the system comprising:

[0031] An acoustic wave signal acquisition module, used to acquire the reflected acoustic wave signal of the anti-corrosion coating of the equipment to be tested collected by the acoustic wave receiver;

[0032] a data preprocessing module, configured to perform data preprocessing on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments;

[0033] A feature extraction and adaptive enhancement module is used to extract features and adaptively enhance the sequence of the coating reflected acoustic wave signal segments to obtain an adaptively enhanced coating reflected acoustic wave waveform feature map;

[0034] The module for determining the degree of detachment of the anti-corrosion coating of the equipment to be detected is used to determine whether the degree of detachment of the anti-corrosion coating of the equipment to be detected exceeds a predetermined requirement based on the waveform characteristic diagram of the acoustic wave reflected by the adaptive strengthening coating.

[0035] Optionally, the data preprocessing module is specifically configured to:

[0036] The reflected acoustic wave signal is image-blocked to obtain a sequence of segments of the coating reflected acoustic wave signal.

[0037] Compared with the existing technology, the present invention reduces the influence of human subjective factors, improves the accuracy and reliability of anti-corrosion coating identification, and realizes the automation of anti-corrosion coating shedding identification, greatly improving the efficiency of identification, saving time and human resources, and can realize timely monitoring and identification of the status of equipment anti-corrosion coating, and can detect anti-corrosion coating shedding problems early, providing timely basis for equipment maintenance and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] One or more embodiments are exemplarily illustrated by pictures in the corresponding drawings, and these exemplifications do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0039] Figure 1 This is a flow chart of a method for identifying equipment anti-corrosion coating shedding provided in one embodiment of the present disclosure;

[0040] Figure 2 A data flow diagram of a method for identifying equipment anti-corrosion coating shedding provided by another embodiment of the present disclosure;

[0041] Figure 3 This is a schematic structural diagram of a system for identifying equipment anti-corrosion coating shedding provided by another embodiment of the present disclosure;

[0042] Figure 4 A schematic diagram of an application scenario of a method and system for identifying the detachment of anti-corrosion coating on equipment provided in another embodiment of the present disclosure. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the embodiments of the present disclosure clearer, the embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. However, it will be understood by those skilled in the art that in each embodiment of the present disclosure, many technical details are provided to enable readers to better understand the present disclosure. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solutions claimed in the present disclosure can be implemented. The division of the following embodiments is for the convenience of description and should not constitute any limitation on the specific implementation of the present disclosure. The various embodiments can be combined and referenced with each other under the premise that there is no contradiction.

[0044] Unless otherwise specified, all technical and scientific terms used in the embodiments of the present disclosure have the same meaning as those commonly understood by those skilled in the art. The terms used in the present disclosure are only for the purpose of describing specific embodiments and are not intended to limit the scope of the present disclosure.

[0045] In the description of the embodiments of the present disclosure, it should be noted that, unless otherwise specified and limited, the term "connection" should be understood in a broad sense. For example, it can be an electrical connection, or it can be the internal connection between two components. It can be a direct connection, or it can be an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meaning of the above terms can be understood according to the specific circumstances.

[0046] It should be noted that the terms "first, second, and third" used in the embodiments of the present disclosure are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It is understood that the terms "first, second, and third" may interchangeably represent a specific order or precedence, where permitted. It should be understood that the terms "first, second, and third" may interchangeably represent objects, where appropriate, such that the embodiments of the present disclosure described herein may be implemented in an order other than that illustrated or described herein.

[0047] In order to solve the problems existing in the prior art in the identification and detection of equipment anti-corrosion coatings, the embodiments of the present disclosure provide a system and method for identifying the shedding of equipment anti-corrosion coatings. The technical concept is: using an acoustic wave receiver to receive the reflected acoustic wave signal of the coating of the equipment, and combining artificial intelligence technology based on deep learning to process and analyze the reflected acoustic wave signal of the coating, so as to realize the identification of the shedding of the anti-corrosion coating. Specifically, the system and method for identifying the shedding of equipment anti-corrosion coatings provided by the embodiments of the present disclosure first obtain the reflected acoustic wave signal of the anti-corrosion coating of the equipment to be detected collected by the acoustic wave receiver, and perform data preprocessing on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments, and then perform feature extraction and adaptive enhancement on the sequence of coating reflected acoustic wave signal segments to obtain an adaptive enhanced coating reflected acoustic wave waveform feature map, and based on the adaptive enhanced coating reflected acoustic wave waveform feature map, determine whether the degree of shedding of the anti-corrosion coating of the equipment to be detected exceeds the predetermined requirements, thereby realizing the identification of the shedding of the anti-corrosion coating.

[0048] Compared to existing methods of manually identifying equipment anti-corrosion coatings, deep learning-based artificial intelligence technology can objectively analyze and process the coating's reflected acoustic wave signals, learn and identify complex acoustic wave signal patterns, and accurately determine whether the anti-corrosion coating has fallen off based on the coating's reflected acoustic wave signals. This reduces the influence of human subjective factors, improves the accuracy and reliability of anti-corrosion coating identification, and automates the identification of anti-corrosion coating shedding, greatly improving identification efficiency and saving time and human resources. In addition, the coating's reflected acoustic wave signals collected by the acoustic wave receiver can be processed and analyzed in real time, enabling timely monitoring and identification of the equipment's anti-corrosion coating status. This allows for early detection of anti-corrosion coating shedding problems and provides a timely basis for equipment maintenance and safety.

[0049] One embodiment of the present disclosure provides a method for identifying equipment anti-corrosion coating shedding, the process of which is as follows: Figure 1 As shown, the data flow is as follows Figure 2 shown.

[0050] like Figure 1 As shown, the method for identifying equipment anti-corrosion coating shedding provided in this embodiment includes the following steps:

[0051] Step S110 , obtaining a reflected sound wave signal of the anti-corrosion coating of the equipment to be tested collected by a sound wave receiver.

[0052] Specifically, to obtain accurate coating-reflected acoustic wave signals, the acoustic receiver must be positioned and angled appropriately, and interference and noise must be avoided during acquisition to ensure signal quality. By using an acoustic receiver to capture the reflected acoustic wave signals from the anti-corrosion coating, the acoustic characteristics of the coating on the equipment surface can be determined, providing foundational data for subsequent analysis and processing.

[0053] In actual application scenarios, different acoustic wave receiver types can be selected based on different needs, and corresponding device parameters can be set. For example, an ultrasonic receiver can be used to collect reflected acoustic wave signals. Specifically, when ultrasonic waves encounter the anti-corrosion coating on the surface of the equipment, they are reflected. The reflected sound waves are captured by the receiver and converted into electrical signals, which serve as the reflected acoustic wave signals of the anti-corrosion coating. This reflected acoustic wave signal contains information such as the thickness, density, and elasticity of the anti-corrosion coating, which is related to the degree of corrosion of the coating. By analyzing the reflected acoustic wave signals from the coating, abnormalities in the equipment's anti-corrosion coating can be identified.

[0054] Step S120 , performing data preprocessing on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments.

[0055] Specifically, data preprocessing can include filtering, denoising, and signal segmentation to extract a sequence of coating-reflected acoustic wave signal segments from the anti-corrosion coating's reflected acoustic wave signal. This data preprocessing removes noise and interference, extracting the valid portion of the coating-reflected acoustic wave signal and providing accurate data for subsequent feature extraction and analysis.

[0056] Step S130 , performing feature extraction and adaptive enhancement on the sequence of coating reflected acoustic wave signal segments to obtain an adaptively enhanced coating reflected acoustic wave waveform feature map.

[0057] Specifically, the features extracted in step S130 can include time-domain features, frequency-domain features, wavelet features, and the like of the sequence of coating-reflected acoustic wave signal segments. After feature extraction, step S130 can process the extracted features using an adaptive enhancement algorithm to produce an adaptively enhanced coating-reflected acoustic wave waveform feature map. Feature extraction and adaptive enhancement can highlight features in the reflected acoustic wave signal of the anti-corrosion coating that are associated with coating delamination, enhancing its detectability and distinguishability. The adaptively enhanced coating-reflected acoustic wave waveform feature map can provide clearer and more accurate information, facilitating subsequent determination and analysis of the extent of coating delamination.

[0058] Step S140 , based on the waveform characteristic diagram of the acoustic wave reflected by the adaptive strengthening coating, it is determined whether the degree of peeling of the anti-corrosion coating of the equipment to be inspected exceeds a predetermined requirement.

[0059] Specifically, the adaptive enhanced coating reflected acoustic waveform characteristic diagram can be used to determine and analyze the degree of anti-corrosion coating shedding. For example, step S140 can set a certain threshold or model and determine whether the degree of anti-corrosion coating shedding exceeds the predetermined requirement by comparing the characteristic diagram with the predetermined requirement.

[0060] By judging the degree of anti-corrosion coating shedding based on the adaptive enhanced coating reflected acoustic waveform characteristic diagram, the status of the equipment's anti-corrosion coating can be automatically evaluated, thereby improving the objectivity and accuracy of the judgment, while reducing the subjectivity and error rate of manual observation, and providing a reliable basis for timely maintenance measures.

[0061] Compared with the existing technology, the method for identifying the shedding of anti-corrosion coating on equipment provided by the embodiment of the present disclosure reduces the influence of human subjective factors, improves the accuracy and reliability of anti-corrosion coating identification, and realizes the automation of anti-corrosion coating shedding identification, greatly improving the efficiency of identification, saving time and human resources, and can realize timely monitoring and identification of the status of equipment anti-corrosion coating, and can detect the problem of anti-corrosion coating shedding early, providing a timely basis for equipment maintenance and safety.

[0062] Exemplarily, step S120 includes: performing image segmentation on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments.

[0063] Specifically, since the anti-corrosion coating on the surface of the equipment may have different corrosion areas, and different corrosion areas may correspond to different reflected sound wave signal characteristics, this embodiment divides the reflected sound wave signal of the anti-corrosion coating in the form of image data into blocks to facilitate separate feature extraction of each area to avoid details being ignored.

[0064] By segmenting the reflected acoustic wave signal from the anti-corrosion coating into distinct segments, each representing the acoustic wave signal characteristics at a different location or time period. This allows for better extraction and analysis of characteristic information from the reflected acoustic wave signal, facilitating subsequent processing and identification. After segmenting the reflected acoustic wave signal, each segment can be processed and analyzed individually, reducing data processing complexity and improving efficiency. Furthermore, targeted processing can be performed on each segment to better capture and identify characteristics of anti-corrosion coating shedding. The sequence of coating reflected acoustic wave signal segments provides temporal information. By analyzing this sequence, temporal changes and trends in the coating reflected acoustic wave signal can be observed, helping to detect and identify the evolution of anti-corrosion coating shedding, proactively identifying potential problems and enabling appropriate maintenance measures. The sequence of coating reflected acoustic wave signal segments also provides information on the spatial distribution of the coating surface. By analyzing this sequence, differences in the reflected acoustic wave signal characteristics at different locations can be observed, helping to determine the location and extent of coating shedding, providing guidance for equipment maintenance and repair.

[0065] By dividing the reflected acoustic wave signals of the anti-corrosion coating into image blocks, more feature information can be provided, data processing can be simplified, and time series analysis and spatial distribution analysis can be performed, which will help improve the accuracy and reliability of identifying anti-corrosion coating shedding and provide more comprehensive support for equipment maintenance and safety.

[0066] Exemplarily, step S130 includes: using a deep learning network model to extract waveform features of a sequence of coating reflected sound wave signal fragments to obtain a sequence of coating reflected sound wave waveform feature maps; passing the sequence of coating reflected sound wave waveform feature maps through a regional separation adaptive attention layer to obtain an adaptive enhanced coating reflected sound wave waveform feature map.

[0067] Specifically, step S130 is mainly to capture the waveform characteristics of the coating reflected sound wave from the sequence of the coating reflected sound wave signal segments to identify abnormal conditions of the coating.

[0068] Exemplarily, the deep learning network model here is a coating reflected acoustic wave waveform feature extractor established based on a convolutional neural network model. The coating reflected acoustic wave waveform feature extractor includes: an input layer, a convolution layer, a pooling layer, an activation layer, and an output layer.

[0069] Exemplarily, in step S130, a deep learning network model is used to extract waveform features of a sequence of coating reflected acoustic wave signal segments to obtain a sequence of coating reflected acoustic wave waveform feature graphs, including: passing the sequence of coating reflected acoustic wave signal segments through a coating reflected acoustic wave waveform feature extractor to obtain a sequence of coating reflected acoustic wave waveform feature graphs.

[0070] Specifically, convolutional neural networks are powerful deep learning models that can effectively extract features from image and waveform data. By inputting a sequence of coating-reflected acoustic wave signal fragments into a convolutional neural network-based coating-reflected acoustic wave waveform feature extractor, the extractor can automatically learn and extract important features of the anti-corrosion coating's reflected acoustic wave signal, capturing relevant patterns and characteristics of anti-corrosion coating shedding, facilitating subsequent shedding identification and analysis. Through multiple layers of convolution and pooling operations, the convolutional neural network model can gradually extract higher-level feature representations. By inputting a sequence of coating-reflected acoustic wave signal fragments into the convolutional neural network-based coating-reflected acoustic wave waveform feature extractor, more abstract and discriminative waveform features are extracted layer by layer, better distinguishing the acoustic wave patterns of normal coating and anti-corrosion coating shedding, thereby improving the accuracy of shedding identification. After processing the sequence of coating-reflected acoustic wave signal fragments through the coating-reflected acoustic wave waveform feature extractor, a sequence of coating-reflected acoustic wave waveform feature maps is generated, providing a more intuitive and visual representation of the acoustic wave characteristics at different time periods or locations. By observing the sequence of characteristic graphs, the temporal changes and spatial distribution of the reflected acoustic wave signals of the anti-corrosion coating can be better analyzed, further supporting the identification and analysis of anti-corrosion coating shedding.

[0071] By using a coating reflected acoustic wave waveform feature extractor based on a convolutional neural network model for feature extraction, a meaningful waveform feature graph sequence can be extracted from a sequence of coating reflected acoustic wave signal fragments. The feature graph sequence can provide more advanced feature representation, more intuitive visualization, and better adaptive enhancement, thereby further improving the recognition accuracy and performance of anti-corrosion coating peeling.

[0072] Exemplarily, in step S130, the sequence of coating reflection sound wave waveform feature maps is separated by a regional adaptive attention layer to obtain an adaptive enhanced coating reflection sound wave waveform feature map, including: performing a convolution operation on the sequence of coating reflection sound wave waveform feature maps to obtain a sequence of separated coating reflection sound wave waveform feature maps; splicing the sequence of separated coating reflection sound wave waveform feature maps into a separated coating reflection sound wave waveform feature map; performing a single-channel convolution operation on the sequence of coating reflection sound wave waveform feature maps to obtain an adaptive attention parameter matrix; performing sub-region attention learning on the separated coating reflection sound wave waveform feature map and the adaptive attention parameter matrix to obtain an adaptive enhanced coating reflection sound wave waveform feature map.

[0073] Among them, the separation coating reflected sound wave waveform feature map and the adaptive attention parameter matrix are subjected to sub-region attention learning to obtain the adaptive strengthening coating reflected sound wave waveform feature map, including: performing point multiplication on each separation coating reflected sound wave waveform feature matrix in the separation coating reflected sound wave waveform feature map and the adaptive attention parameter matrix to obtain the adaptive strengthening coating reflected sound wave waveform feature map.

[0074] Specifically, the region-separated adaptive attention layer can focus on specific regions of the coating's reflected acoustic waveform feature map, helping to increase attention to key areas, thereby better capturing and emphasizing the characteristics of anti-corrosion coating shedding. This regional attention reduces interference from irrelevant areas and improves the accuracy and robustness of shedding detection. The region-separated adaptive attention layer also enhances the features of key regions in the coating's reflected acoustic waveform feature map. By focusing on these key regions, the characteristics of anti-corrosion coating shedding can be highlighted and enhanced, making them easier to identify and analyze. This feature enhancement helps improve the accuracy and reliability of shedding detection. The region-separated adaptive attention layer also learns the spatial relationships between different regions. By modeling the associations between different regions, it can capture the spatial distribution patterns of anti-corrosion coating shedding, helping to better understand and analyze the characteristics of anti-corrosion coating shedding and improve the performance of shedding detection. By using the region-separated adaptive attention layer for adaptive reinforcement, the computational and storage overhead of the network can also be reduced. Compared to performing attention calculations on the entire feature map sequence, region separation can reduce computational complexity and improve network efficiency and speed, facilitating real-time anti-corrosion coating shedding detection and analysis.

[0075] Exemplarily, step S140 includes: optimizing the characteristic distribution of the waveform characteristic diagram of the reflected sound wave of the adaptive strengthening coating to obtain an optimized waveform characteristic diagram of the reflected sound wave of the adaptive strengthening coating; passing the optimized waveform characteristic diagram of the reflected sound wave of the adaptive strengthening coating through a classifier to obtain a classification result, and the classification result is used to indicate whether the degree of detachment of the anti-corrosion coating of the equipment to be tested exceeds predetermined requirements.

[0076] Specifically, feature distribution optimization can adjust the feature distribution in the adaptively enhanced coating reflection acoustic waveform feature map, making key features more prominent and distinct. This enhances the characteristics of anti-corrosion coating shedding, making them easier to identify and analyze. Feature enhancement helps improve the accuracy and reliability of shedding detection. Feature distribution optimization can also reduce noise and interference in the feature map. By adjusting the feature distribution, irrelevant features can be suppressed, making the characteristics of anti-corrosion coating shedding more distinct and clear, thereby improving the robustness of shedding detection and making it more robust to interference and noise. Feature distribution optimization can also make the features in the coating reflection acoustic waveform feature map more spatially consistent, helping to improve the consistency and stability of shedding detection. By optimizing the feature distribution, it can also ensure that the same type of anti-corrosion coating shedding has similar appearance in the feature map, further improving the performance of shedding detection. Feature distribution optimization can also make the features in the coating reflection acoustic waveform feature map more interpretable. By adjusting the feature distribution, the features in the feature map can be made more consistent with human cognition and understanding of anti-corrosion coating shedding, thereby improving the interpretability of shedding detection and making the results easier to understand and apply.

[0077] Through feature distribution optimization, the waveform characteristic diagram of the reflected acoustic wave of the adaptive enhanced coating can be adjusted and optimized to make it more prominent, clear, robust and interpretable, which will help improve the recognition and analysis capabilities of anti-corrosion coating shedding and further enhance the accuracy, reliability and practicality of shedding identification.

[0078] In the above embodiment, when the sequence of coating reflection acoustic waveform feature maps passes through the regional separation adaptive attention layer, the regional separation adaptive attention layer performs single-channel spatial attention encoding on each coating reflection acoustic waveform feature map in the sequence of coating reflection acoustic waveform feature maps, and splices the spatial attention information in the spatial dimension and acts on the sequence of coating reflection acoustic waveform feature maps to obtain a global coating reflection acoustic waveform feature map. In this way, the local features of certain predetermined areas in the coating reflection acoustic wave signal are adaptively emphasized, thereby obtaining an adaptive enhanced coating reflection acoustic waveform feature map. However, this will also cause the spatial dimension of the feature distribution of the adaptive enhanced coating reflection acoustic waveform feature map to be sparse, resulting in poor convergence of the probability density distribution of the class regression probability when the adaptive enhanced coating reflection acoustic waveform feature map is classified and regressed through the classifier, affecting the effect of classification regression and the accuracy of the classification results.

[0079] Therefore, preferably, in step S140, when classifying the adaptive strengthening coating reflected acoustic wave waveform characteristic graph through the classifier, the adaptive strengthening coating reflected acoustic wave waveform characteristic vector obtained after expansion can be subjected to class coding optimization, which is specifically expressed as follows: the adaptive strengthening coating reflected acoustic wave waveform characteristic vector obtained after expansion of the adaptive strengthening coating reflected acoustic wave waveform characteristic graph is subjected to class coding optimization using the following optimization formula to obtain the optimized adaptive strengthening coating reflected acoustic wave waveform characteristic vector obtained after expansion of the optimized adaptive strengthening coating reflected acoustic wave waveform characteristic graph; wherein the optimization formula is:

[0080]

[0081] Among them, v i and v j are the i-th eigenvalue and j-th eigenvalue of the waveform feature vector of the reflected acoustic wave of the adaptive strengthening coating obtained after the waveform feature map of the reflected acoustic wave of the adaptive strengthening coating is expanded, and V represents the set of all eigenvalues ​​of the waveform feature vector of the reflected acoustic wave of the adaptive strengthening coating obtained after the waveform feature map of the reflected acoustic wave of the adaptive strengthening coating is expanded. is the global characteristic mean of the characteristic vector of the reflected acoustic wave of the adaptive strengthening coating, v i ′ is the i-th eigenvalue of the waveform characteristic vector of the reflected acoustic wave of the optimized adaptive strengthening coating obtained by expanding the waveform characteristic graph of the reflected acoustic wave of the optimized adaptive strengthening coating, and exp(·) represents the value of the natural exponential function raised to a numerical value.

[0082] Specifically, to address the problem of local probability density mismatch in the probability density distribution in the probability space caused by the sparse distribution of the adaptive enhanced coating reflected acoustic waveform feature vector in the high-dimensional feature space, a global self-consistent class coding based on the soft L1 regularization can be used to imitate the global self-consistent relationship of the category coding behavior of the high-dimensional feature manifold of the adaptive enhanced coating reflected acoustic waveform feature vector in the probability space, so as to adjust the error landscape of the feature manifold in the high-dimensional open space domain and realize the self-consistent matching class coding of the high-dimensional feature manifold of the adaptive enhanced coating reflected acoustic waveform feature vector embedded in the explicit probability space, thereby improving the convergence of the probability density distribution of the class regression probability of the adaptive enhanced coating reflected acoustic waveform feature map, and improving the effect of classification regression and the accuracy of classification results.

[0083] By using a classifier to classify the optimized feature map, i.e., the optimized adaptive enhanced coating reflected acoustic waveform feature map, the degree of anti-corrosion coating shedding on the equipment can be assessed. The classification result can provide a quantitative indicator to help determine whether the anti-corrosion coating shedding exceeds the predetermined requirements, which helps to promptly detect and address corrosion coating problems and ensure the normal operation and safety of the equipment. By using the optimized feature map and classifier to classify the degree of anti-corrosion coating shedding, highly accurate results can be obtained. The optimized feature map has undergone the data processing, feature extraction, and feature distribution optimization mentioned above, and can better capture the characteristics of anti-corrosion coating shedding. After training and optimization, the classifier can effectively distinguish between different degrees of corrosion coating shedding. Therefore, the classification results are highly accurate and reliably reflect the condition of the anti-corrosion coating.

[0084] By inputting the optimized characteristic map, i.e., the optimized adaptive enhanced coating reflected acoustic waveform characteristic map, into the classifier for classification, real-time monitoring of anti-corrosion coating shedding can be achieved. This means that real-time monitoring and evaluation can be performed during equipment operation, allowing anti-corrosion coating problems to be discovered promptly and appropriate measures to be taken. Real-time monitoring helps prevent equipment failures and accidents, and improves equipment reliability and safety. Classification results can be presented visually, such as through bar charts, heat maps, or color coding to indicate the degree of anti-corrosion coating shedding. Such visualizations make it easier for operators to quickly understand and judge the condition of the equipment's anti-corrosion coating, providing intuitive feedback information to support decision-making and maintenance work.

[0085] By inputting the optimized adaptive enhanced coating reflected acoustic waveform feature map into the classifier for classification, the degree of anti-corrosion coating shedding of the equipment to be inspected can be evaluated. This method has the advantages of high accuracy, real-time monitoring and visualization of results, which helps to improve the efficiency and reliability of equipment maintenance and safety management.

[0086] Another embodiment of the present disclosure relates to an equipment anti-corrosion coating peeling identification system 200, such as Figure 3 As shown, it includes an acoustic signal acquisition module 210, a data preprocessing module 220, a feature extraction and adaptive enhancement module 230, and a module 240 for determining the degree of peeling of the anti-corrosion coating of the equipment to be detected.

[0087] The acoustic wave signal acquisition module 210 is used to acquire the reflected acoustic wave signal of the anti-corrosion coating of the equipment to be tested collected by the acoustic wave receiver.

[0088] The data preprocessing module 220 is used to perform data preprocessing on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments.

[0089] The feature extraction and adaptive enhancement module 230 is used to perform feature extraction and adaptive enhancement on a sequence of coating reflected acoustic wave signal segments to obtain an adaptively enhanced coating reflected acoustic wave waveform feature map.

[0090] The module 240 for determining the degree of detachment of the anti-corrosion coating of the equipment to be inspected is used to determine whether the degree of detachment of the anti-corrosion coating of the equipment to be inspected exceeds a predetermined requirement based on the waveform characteristic diagram of the acoustic wave reflected by the adaptively enhanced coating.

[0091] Exemplarily, the data preprocessing module 220 is specifically configured to perform image segmentation on the reflected acoustic wave signal to obtain a sequence of coating reflected acoustic wave signal segments.

[0092] The specific implementation method of the equipment anti-corrosion coating shedding identification system provided by the embodiment of the present disclosure can be found in the equipment anti-corrosion coating shedding identification method provided by the embodiment of the present disclosure, and will not be repeated here.

[0093] Compared with the existing technology, the equipment anti-corrosion coating shedding identification system provided by the embodiment of the present disclosure reduces the influence of human subjective factors, improves the accuracy and reliability of anti-corrosion coating identification, and realizes the automation of anti-corrosion coating shedding identification, greatly improving the efficiency of identification, saving time and human resources, and can realize timely monitoring and identification of the status of equipment anti-corrosion coating, and can detect anti-corrosion coating shedding problems early, providing timely basis for equipment maintenance and safety.

[0094] The device anti-corrosion coating peeling identification system 200 provided in the embodiments of the present disclosure can be implemented in various terminal devices, such as a server for identifying device anti-corrosion coating peeling. In one example, the device anti-corrosion coating peeling identification system 200 can be integrated into the terminal device as a software module and / or hardware module. For example, the device anti-corrosion coating peeling identification system 200 can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the device anti-corrosion coating peeling identification system 200 can also be one of the many hardware modules of the terminal device.

[0095] Alternatively, in another example, the equipment anti-corrosion coating peeling identification system 200 and the terminal device can also be separate devices, and the equipment anti-corrosion coating peeling identification system 200 can be connected to the terminal device through a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0096] Figure 4 The following shows the application scenario of the equipment anti-corrosion coating peeling identification method and system provided by the embodiment of the present disclosure. Figure 4As shown, in this application scenario, first, step S110 or the acoustic wave signal acquisition module 210 acquires the reflected acoustic wave signal of the anti-corrosion coating of the device to be detected collected by the acoustic wave receiver. The reflected acoustic wave signal can be expressed as Figure 4 Then, the reflected sound wave signal is input into the server deployed with the equipment anti-corrosion coating peeling identification algorithm, where the server can be represented as Figure 4 As shown in S, the server can process the reflected sound wave signal of the anti-corrosion coating based on the equipment corrosion coating shedding identification algorithm corresponding to steps S120 to S140 or the equipment corrosion coating shedding identification algorithm corresponding to the data preprocessing module 220, the feature extraction and adaptive enhancement module 230, and the anti-corrosion coating shedding degree determination module 240 of the equipment to be detected to determine whether the degree of shedding of the anti-corrosion coating of the equipment to be detected exceeds the predetermined requirement.

[0097] Those skilled in the art will appreciate that the above-mentioned embodiments are specific embodiments for implementing the present disclosure, and that in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present disclosure.

Claims

1. A method for identifying equipment anti-corrosion coating shedding, characterized in that: The method comprises: Acquiring a reflected sound wave signal of the anti-corrosion coating of the equipment to be tested collected by a sound wave receiver; Performing data preprocessing on the reflected sound wave signal to obtain a sequence of coating reflected sound wave signal fragments; Performing feature extraction and adaptive enhancement on the sequence of the coating reflected sound wave signal fragments to obtain an adaptive enhanced coating reflected sound wave waveform feature map; Based on the characteristic waveform diagram of the acoustic wave reflected by the adaptive strengthening coating, it is determined whether the degree of shedding of the anti-corrosion coating of the equipment to be inspected exceeds a predetermined requirement.

2. The method according to claim 1, characterized in that Performing data preprocessing on the reflected sound wave signal to obtain a sequence of coating reflected sound wave signal fragments, including: The reflected sound wave signal is divided into image blocks to obtain a sequence of segments of the reflected sound wave signal of the coating.

3. The method according to claim 1, characterized in that Feature extraction and adaptive enhancement are performed on the sequence of the coating reflected sound wave signal fragments to obtain an adaptive enhanced coating reflected sound wave waveform feature map, including: Using a deep learning network model, waveform feature extraction is performed on the sequence of the coating reflected sound wave signal fragments to obtain a sequence of coating reflected sound wave waveform feature graphs; The sequence of the coating reflected sound wave waveform characteristic graphs is passed through a regional separation adaptive attention layer to obtain the adaptive enhanced coating reflected sound wave waveform characteristic graph.

4. The method according to claim 3, characterized in that The deep learning network model is a coating reflected sound wave waveform feature extractor established based on a convolutional neural network model; the coating reflected sound wave waveform feature extractor includes: an input layer, a convolution layer, a pooling layer, an activation layer and an output layer.

5. The method according to claim 4, characterized in that The deep learning network model is used to extract waveform features of the sequence of the coating reflected sound wave signal fragments to obtain a sequence of waveform feature graphs of the coating reflected sound wave, including: The sequence of the coating reflected sound wave signal segments is passed through the coating reflected sound wave waveform feature extractor to obtain a sequence of the coating reflected sound wave waveform feature graphs.

6. The method according to claim 3, characterized in that The sequence of the coating reflected sound wave waveform characteristic graphs is passed through a region separation adaptive attention layer to obtain the adaptive enhanced coating reflected sound wave waveform characteristic graph, including: Performing a convolution operation on the sequence of waveform characteristic graphs of the acoustic wave reflected by the coating to obtain a sequence of waveform characteristic graphs of the acoustic wave reflected by the separation coating; splicing the sequence of the separation coating layer reflected sound wave waveform characteristic graphs into a separation coating layer reflected sound wave waveform characteristic graph; Performing a single-channel convolution operation on a sequence of waveform characteristic graphs of acoustic waves reflected by the coating to obtain an adaptive attention parameter matrix; The separation coating reflected sound wave waveform characteristic diagram and the adaptive attention parameter matrix are subjected to sub-region attention learning to obtain the adaptive strengthening coating reflected sound wave waveform characteristic diagram.

7. The method according to claim 6, characterized in that The separation coating reflected sound wave waveform characteristic graph and the adaptive attention parameter matrix are subjected to sub-region attention learning to obtain the adaptive strengthening coating reflected sound wave waveform characteristic graph, including: The waveform characteristic map of the reflected sound wave of the adaptive strengthening coating is obtained by performing a dot multiplication of each separation coating reflected sound wave waveform characteristic matrix in the separation coating reflected sound wave waveform characteristic map with the adaptive attention parameter matrix.

8. The method according to any one of claims 1 to 7, characterized in that: Based on the waveform characteristic diagram of the acoustic wave reflected by the adaptive strengthening coating, determining whether the degree of the anti-corrosion coating of the equipment to be detected falls off exceeds a predetermined requirement, including: Optimizing the characteristic distribution of the waveform characteristic diagram of the reflected sound wave of the adaptive strengthening coating to obtain an optimized waveform characteristic diagram of the reflected sound wave of the adaptive strengthening coating; The optimized adaptive strengthening coating reflected acoustic wave waveform characteristic diagram is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the degree of detachment of the anti-corrosion coating of the equipment to be tested exceeds the predetermined requirement.

9. An equipment anti-corrosion coating peeling identification system, characterized in that: The system comprises: An acoustic wave signal acquisition module, used to acquire the reflected acoustic wave signal of the anti-corrosion coating of the equipment to be tested collected by the acoustic wave receiver; A data preprocessing module, used for performing data preprocessing on the reflected sound wave signal to obtain a sequence of coating reflected sound wave signal fragments; A feature extraction and adaptive enhancement module is used to extract features and adaptively enhance the sequence of the coating reflected sound wave signal fragments to obtain an adaptive enhanced coating reflected sound wave waveform feature map; The module for determining the degree of detachment of the anti-corrosion coating of the equipment to be detected is used to determine whether the degree of detachment of the anti-corrosion coating of the equipment to be detected exceeds a predetermined requirement based on the waveform characteristic diagram of the acoustic wave reflected by the adaptive strengthening coating.

10. The system according to claim 9, characterized in that The data preprocessing module is specifically used for: The reflected sound wave signal is divided into image blocks to obtain a sequence of segments of the reflected sound wave signal of the coating.

Images