Tank bottom plate defect identification method, device and equipment and storage medium

Abstract

The application discloses a kind of storage tank bottom plate defect identification method, device, equipment and storage medium, it is related to detection technical field.The method includes obtaining the defect data set of storage tank bottom plate, defect data set includes defect data and the defect type corresponding to defect data, defect data refers to the data obtained by detecting the defect of storage tank bottom plate by ultrasonic phased array.The defect data set is divided, and training set data and test set data are obtained, based on training set data, training defect type identification model, based on test set data, the accuracy of the trained defect type identification model is tested, and the defect type of storage tank bottom plate is identified based on the defect type identification model tested by accuracy.

Description

Technical Field

[0001] This application relates to the field of detection technology, and in particular to a method, apparatus, equipment and storage medium for identifying defects in the bottom plate of a storage tank. Background Technology

[0002] A storage tank is a large container used to store liquids, gases, or solid materials, and is widely used in industries such as petroleum, chemical, energy, food, and pharmaceuticals. For storage tanks that are in long-term contact with liquids, corrosive media, or subjected to alternating loads, defects in the tank bottom plate (such as corrosion, cracks, perforations, etc.) may lead to leaks, environmental pollution, and safety accidents.

[0003] Existing methods for identifying defects in tank bottom plates include using ultrasonic testing to detect common defects such as cracks, corrosion, and delamination. For example, ultrasonic phased array testing can be used to obtain defect images of the tank bottom plate, and the location and type of defects can be identified based on these images.

[0004] However, determining the type of defect in the tank bottom plate based on inspection results mainly relies on the experience and judgment of the staff, which is costly and carries risks such as misjudgment and missed detection. Moreover, when the tank bottom plate structure is complex and multiple defect types are mixed, the accuracy of defect type identification will be further reduced. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus, equipment and storage medium for identifying defects in the bottom plate of a storage tank, thereby improving the efficiency and accuracy of defect type identification.

[0006] To achieve the above objectives, this application adopts the following technical solution:

[0007] In a first aspect, this application provides a method for identifying defects in the bottom plate of a storage tank, the method comprising:

[0008] Obtain the defect dataset of the tank bottom plate.

[0009] The defect dataset is divided into training set data and test set data.

[0010] A defect type identification model is trained based on the training set data.

[0011] The accuracy of the trained defect type identification model is tested based on the test set data.

[0012] The defect types of the tank bottom plate are identified based on a defect type identification model that has passed accuracy testing.

[0013] The defect dataset includes defect data and the corresponding defect types. The defect data refers to the data obtained by detecting defects in the bottom plate of the storage tank using an ultrasonic phased array.

[0014] The technical solution provided in this application embodiment trains a defect type recognition model using a defect dataset, and then uses the trained defect type recognition model to identify the defect type of the tank bottom plate. This enables efficient and accurate identification of the defect type represented by the damage detection result after the tank bottom plate is damaged by ultrasonic phased array, thereby improving the efficiency and accuracy of defect type identification.

[0015] In some embodiments, the defect dataset is divided into training and test sets. Specifically, the defect dataset is divided into at least two parts, with data from one part used as the test set and all other data in the defect dataset used as the training set. Both the test and training sets contain data from the defect dataset, exhibiting similar patterns. By using test set data, which the defect type identification model has not previously encountered, the prediction performance is evaluated, ensuring the reliability and accuracy of the defect type identification model's performance.

[0016] In some embodiments, data from one defect dataset is used as the test set, and data from other defect datasets are used as the training set. Specifically, this can be implemented by using each data point in the defect dataset as the test set and all data in the defect datasets excluding the test set data as the training set, resulting in multiple sets of training and test data. By training and testing the defect type recognition model using multiple sets of training and test data, each data point in the defect dataset is used as both test and training data. Compared to a single division of training and test data, this method achieves higher data utilization and helps identify performance differences of the defect type recognition model across different sets of data, avoiding overfitting or underfitting.

[0017] In some embodiments, training a defect type identification model based on training set data can be specifically implemented as follows: training the defect type identification model sequentially based on each set of training set data. By training the defect type identification model separately according to different sets of training set data, the defect type identification model learns more general data patterns, thereby improving the prediction accuracy of the defect type identification model.

[0018] In some embodiments, the accuracy test of the trained defect type identification model based on test set data can be specifically implemented as follows: calculate the identification accuracy of the defect type identification model based on each set of test set data, perform a weighted average of the calculated identification accuracy rates to obtain the target accuracy rate, and determine the defect type identification model whose target accuracy rate is greater than the accuracy rate threshold as the defect type identification model that has passed the accuracy test. Accuracy testing based on the target accuracy rate obtained from multiple training iterations results in more accurate and objective test results.

[0019] In some embodiments, training a defect type recognition model based on training set data can be specifically implemented by: extracting local binary pattern features from defect data, and training the defect type recognition model based on the local binary pattern features and the defect type. Pre-extracting local binary patterns is simple and efficient, reducing the computational load of the defect type recognition model. Simultaneously, the local binary patterns effectively capture the local texture structure of the image, better representing the damage detection effect of the tank bottom plate.

[0020] In some embodiments, the formula for calculating local binary pattern features includes:

[0021]

[0022] Among them, LBP r，p Let S(z) represent the local binary pattern feature, S(z) represent the sign function, p represent the number of neighboring pixels, r represent the neighborhood radius, and x represent the value of x. c This represents the grayscale value of the c-th neighboring pixel.

[0023] Secondly, a tank bottom plate defect identification device is provided, which includes: an acquisition module, a processing module, and an identification module.

[0024] The aforementioned acquisition module is used to acquire a defect dataset of the tank bottom plate.

[0025] The aforementioned processing module is used to divide the defect dataset into training set data and test set data.

[0026] The aforementioned processing module is also used to train a defect type identification model based on the training set data.

[0027] The aforementioned processing module is also used to perform accuracy testing on the trained defect type identification model based on the test set data.

[0028] The aforementioned identification module is used to identify the defect types of the tank bottom plate based on a defect type identification model that has passed an accuracy test.

[0029] The defect dataset includes defect data and the corresponding defect types. The defect data refers to the data obtained by detecting defects in the bottom plate of the storage tank using an ultrasonic phased array.

[0030] In some embodiments, the above processing module is further configured to: divide the defect dataset into at least two parts, use the data in one of the defect datasets as test set data, and use the data in the defect dataset other than the test set data as training set data.

[0031] In some embodiments, the above processing module is further configured to: use each piece of data in the defect dataset as test set data and the data in the defect dataset other than the test set data as training set data, to obtain multiple sets of training set data and test set data.

[0032] In some embodiments, the above-described processing module is further configured to: train a defect type identification model sequentially based on each set of training data.

[0033] In some embodiments, the above processing module is further configured to: calculate the recognition accuracy of the defect type recognition model based on each set of test data, perform a weighted average of the calculated recognition accuracy to obtain the target accuracy, and determine the defect type recognition model with a target accuracy greater than the accuracy threshold as the defect type recognition model that has passed the accuracy test.

[0034] In some embodiments, the above-described processing module is further configured to: extract local binary pattern features from the defect data, and train a defect type recognition model based on the local binary pattern features and the defect type.

[0035] In some embodiments, the formula for calculating local binary pattern features includes:

[0036]

[0037] Among them, LBP r，p Let S(z) represent the local binary pattern feature, S(z) represent the sign function, p represent the number of neighboring pixels, r represent the neighborhood radius, and x represent the value of x. c This represents the grayscale value of the c-th neighboring pixel.

[0038] The technical effects of any implementation method in the second aspect can be found in the technical effects of any implementation method in the first aspect mentioned above, and will not be repeated here.

[0039] Thirdly, a computer device is provided, comprising: a processor and a memory, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the above-mentioned method for identifying defects in the tank bottom plate.

[0040] Fourthly, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, and the at least one computer program is loaded and executed by a processor to implement the above-mentioned method for identifying defects in the tank bottom plate.

[0041] Fifthly, a computer program product is provided, which includes a computer program or instructions, which, when executed by a processor, implement the above-mentioned method for identifying defects in the tank bottom plate.

[0042] The solutions provided in aspects three through five above are used to implement the method provided in aspect one above, and their specific implementations will not be described in detail here. The technical effects corresponding to any implementation method of the solutions provided in aspects three through five above can be found in the technical effects corresponding to any implementation method in aspect one above, and will not be described in detail here.

[0043] It should be noted that any of the possible implementations of any of the above aspects can be combined, provided that the solutions do not contradict each other. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A schematic diagram of the structure of a computer system provided as an exemplary embodiment;

[0046] Figure 2 A flowchart illustrating a method for identifying defects in the bottom plate of a storage tank, provided as an exemplary embodiment;

[0047] Figure 3 A schematic diagram of the structure of a random forest model is provided as an exemplary embodiment;

[0048] Figure 4 A flowchart illustrating another method for identifying defects in the bottom plate of a storage tank, provided as an exemplary embodiment;

[0049] Figure 5 A schematic diagram of a data-enhanced image provided for an exemplary embodiment;

[0050] Figure 6 A schematic diagram of the accuracy curve of a defect identification model provided for an exemplary embodiment;

[0051] Figure 7 A schematic diagram of a local binary pattern feature image provided as an exemplary embodiment;

[0052] Figure 8 A schematic diagram of an estimated fractional image of the average recognition rate of cracks and porosity defects provided for an exemplary embodiment;

[0053] Figure 9 A schematic diagram of a tank bottom plate defect identification device provided as an exemplary embodiment;

[0054] Figure 10 A schematic diagram of the structure of a computer device provided for an exemplary embodiment. Detailed Implementation

[0055] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0056] In the description of this application, it should be understood that the terms "upper," "lower," "left," "right," "front," "rear," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or relative positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Unless otherwise specified, the above orientational descriptions can be flexibly set in actual application, provided that the relative positional relationship shown in the accompanying drawings is satisfied.

[0057] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0058] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "communication" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection. They can refer to a direct connection or an indirect connection through an intermediate medium, or a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0059] In embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, article, or apparatus that includes that element.

[0060] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0061] In the embodiments of this application, at least one can also be described as one or more, and multiple can be two, three, four or more, and this application does not impose any restrictions.

[0062] In the description of this specification, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0063] To facilitate understanding, the terms used in the embodiments of this application will be explained first.

[0064] Damage detection refers to the identification, location, and assessment of damage or defects in materials, structures, or systems through a series of techniques and methods. Specifically, it includes non-destructive testing (NDT), vibration analysis testing, machine learning and artificial intelligence (AI) testing, and sensor testing. NDT does not damage the object being tested; examples include ultrasonic testing, X-ray imaging, eddy current testing, and infrared thermography. Vibration analysis testing identifies damage by analyzing changes in structural vibration signals, such as frequency response analysis. Machine learning and AI testing automatically detects surface cracks or internal defects through image recognition (such as deep learning). Sensor testing monitors damage development in real time using deployed strain gauges, fiber optic sensors, etc.

[0065] Ultrasonic phased array (ESA) is an ultrasonic testing technology that achieves directional emission and focusing of sound waves by controlling the excitation time and sequence of multiple independent piezoelectric crystals. It is primarily used in industrial non-destructive testing and medical imaging. The electronic scanning speed of an ultrasonic phased array is much faster than mechanical scanning, and it can achieve multi-angle and multi-depth detection with a single probe. It can detect irregular shapes or hard-to-reach areas, offering high flexibility and adaptability. Furthermore, the dynamic focusing of the ultrasonic phased array improves the defect detection rate and provides high damage detection accuracy.

[0066] It should be noted that all information (including but not limited to device information, personal information of the target), data (including but not limited to data used for analysis, stored data, and displayed data), and signals involved in this application have been authorized by the target or fully authorized by all parties, and the collection, use, and processing of related data must comply with relevant laws, regulations, and standards. For example, the defect data, defect types, etc. involved in this application were all obtained with full authorization.

[0067] The industry-standard method for identifying defects in tank bottom plates mainly involves locating defects through damage detection data and then manually determining the defect type. This will be briefly explained below.

[0068] For example, a magnetic gradient testing unit is set up to acquire magnetic measurement data of the tank wall in real time to determine the structural health status of the tank. For instance, based on the principle that stress concentration caused by early defects alters the self-leaking magnetic field, localized stress concentration areas can be identified, thereby enabling the detection and location of early defects. Simultaneously, by utilizing liquid level raising and lowering operations, cyclic loads are applied to the tank wall, increasing the difference in self-leaking magnetic field strength between tank wall defects and the main body, enabling accurate detection of defects with weak initial magnetic signal strength and improving the accuracy of tank wall defect identification.

[0069] However, the above technical solutions only achieve the location of tank defects. The improvement in defect identification accuracy refers to the improvement in defect location accuracy. They cannot directly analyze and identify the specific type of defect by analyzing the damage detection data. Especially for mixed defects, the damage detection data of various defects interfere with each other and become mixed, making accurate identification impossible.

[0070] Based on this, this application provides a method for identifying defects in tank bottom plates. The method trains a defect type identification model using a defect dataset, and then uses the trained defect type identification model to identify the defect type of the tank bottom plate. This enables efficient and accurate identification of the defect type represented by the damage detection result after the tank bottom plate is damaged by ultrasonic phased array, thereby improving the efficiency and accuracy of defect type identification.

[0071] The solution provided in this application can be applied to Figure 1 In the computer system shown, Figure 1 A schematic diagram of the structure of a computer system provided for an exemplary embodiment.

[0072] For example, the computer system includes a computer device 100. The computer device 100 can be a high-performance server, serving as the core of the computer system, used to identify the defect detection results 101 of the tank bottom plate and output the defect type 102. The computer device 100 can directly / indirectly acquire the defect detection results 101 and perform identification, analysis, and reasoning on them. The computer device 100 can deploy a defect identification model, using it to identify, analyze, and reason on the defect detection results 101, and output the defect type 102. The computer device 100 can be configured with a database including a defect dataset, and can train and test the defect identification model based on the defect dataset. The computer device 100 can accept instructions from external personnel and flexibly adjust the configuration and training parameters of the defect identification model based on those instructions. In this application, "acquisition" in the context of the computer device 100 includes any term with acquisition functionality such as querying, discovering, and extracting, and is not limited thereto.

[0073] Optionally, the computer device 100 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and basic cloud computing services such as big data. This application embodiment does not limit the implementation method or application scenario of the computer device 100.

[0074] Figure 2 This is a flowchart illustrating a method for identifying defects in the bottom plate of a storage tank, provided as an exemplary embodiment. The method can be executed by a computer device. The computer device can be... Figure 1 The computer equipment mentioned can also be other equipment capable of data analysis and processing.

[0075] like Figure 2 As shown in the embodiments of this application, the method for identifying defects in the tank bottom plate may include:

[0076] Step S201: The computer equipment acquires the defect dataset of the tank bottom plate.

[0077] The defect dataset includes defect data and the defect types corresponding to the defect data.

[0078] Defect data refers to the data obtained by detecting defects in the bottom plate of a storage tank using an ultrasonic phased array.

[0079] Optionally, a defect in the tank bottom plate refers to physical or chemical damage to the tank bottom plate caused by long-term use, environmental factors, or improper operation, or a defect that occurs during the manufacturing or installation of the tank bottom plate.

[0080] For example, the types of defects in the tank bottom plate include uneven / reduced bottom plate thickness, partial bulges, depressions, cracks, pores, pitting, weld corrosion, etc.

[0081] Optionally, the defect data can be represented as waveforms, signals, images, data values, etc.

[0082] Ultrasonic phased array is a non-destructive testing technology that uses multi-element probes and electronic control to achieve dynamic focusing and scanning of the sound beam, and can be used to detect defects / damage on the surface of storage tanks.

[0083] Specifically, the ultrasonic phased array's beam scanning modes include A-scan, B-scan, and C-scan. The A-scan displays the amplitude-time signal at a single location, used for defect depth and amplitude analysis. The B-scan is a cross-sectional view along the probe's movement direction, showing the vertical distribution of defects. The C-scan is a top-view projection, showing the planar distribution of defects (positions need to be recorded by the encoder). Therefore, the length of a defect / damage can be determined by the signal extension range in the C-scan or B-scan, and the depth of the defect / damage is calculated using time-of-flight (TOF) measurement in the A-scan. The height of the defect / damage (e.g., cracks) is measured using time-of-flight diffraction (TOFD).

[0084] For example, the defect data corresponding to cracks is a high-amplitude echo, and the signal shifts significantly with changes in the beam angle. The defect data corresponding to pores is an isolated circular echo, and the signal position is fixed during multi-angle scanning. The defect data corresponding to incomplete penetration is a continuous linear echo at the weld root with a consistent depth. The defect data corresponding to corrosion thinning is a forward shift or disappearance of the bottom echo, accompanied by a decrease in the wall thickness measurement. The defect data corresponding to delamination is a high-amplitude echo appearing inside the plate, separated from the bottom echo.

[0085] For example, a computer device acquires detection data of known / typical defects in the bottom plate of a storage tank using an ultrasonic phased array, performs full-focus imaging, and forms a defect dataset.

[0086] Step S202: The computer equipment divides the defect dataset to obtain training set data and test set data.

[0087] Optionally, the defect dataset can be partitioned in various ways, including random partitioning, stratified sampling partitioning, and time-series partitioning. For example, 70%–80% of the defect data and their corresponding defect types in the defect dataset can be used as the training set data, while 20%–30% of the defect data and their corresponding defect types can be used as the test set data. Another example is selecting defect types from the defect dataset, and extracting a portion of the defect data from multiple defect data points corresponding to the same defect type as the training set data, with the remaining data used as the test set data.

[0088] Step S203: The computer device trains a defect type identification model based on the training set data.

[0089] Among them, the defect type identification model is used to predict / identify the defect type corresponding to the defect data based on the defect data.

[0090] Optionally, defect type identification models include Random Forest models, Support Vector Machine (SVM) models, convolutional neural network models, etc.

[0091] In some embodiments, the computer device trains a random forest model based on training set data to obtain a trained defect type recognition model.

[0092] For example, such as Figure 3 As shown, the defect data in the training set is input into the random forest model 300, and the output is the identified predicted defect type. The model parameters of the random forest model are optimized based on the defect type and the loss of the predicted defect type, such as: the number of decision trees 301 (nestimators), the maximum depth of decision trees 301 (max_depth), the feature selection method (max_features), the combiner 302, etc., until the preset number of training times is reached, or the loss of the defect type and the predicted defect type converges, and the defect type identification model training is completed.

[0093] Step S204: The computer equipment performs an accuracy test on the trained defect type identification model based on the test set data.

[0094] The accuracy test refers to evaluating the recognition accuracy of the trained defect type identification model. An accuracy rate higher than the threshold indicates that the model has passed the accuracy test. An accuracy rate lower than or equal to the threshold requires retraining of the defect type identification model.

[0095] For example, the formula for calculating recognition accuracy is as follows:

[0096]

[0097] Where accuracy represents the recognition accuracy, TP represents correctly classified positive samples, TN represents correctly classified negative samples, FP represents incorrectly classified positive samples, and FN represents incorrectly classified negative samples. For example, if crack defects are positive samples and pore defects are negative samples, TP indicates that crack defects are correctly identified, TN indicates that pore defects are correctly identified, FP indicates that crack defects are incorrectly identified, and FN indicates that pore defects are incorrectly identified.

[0098] Step S205: The computer equipment identifies the defect type of the tank bottom plate based on the defect type identification model that has passed the accuracy test.

[0099] For example, a computer device identifies the defect type based on the damage detection results of the tank bottom plate using a defect type identification model that has passed an accuracy test. For instance, the tank bottom plate is scanned using an ultrasonic phased array, the resulting focused image is input into the defect type identification model that has passed an accuracy test, and the identified defect type is output.

[0100] In summary, the technical solution provided in this application training method uses a defect type identification model to train a defect type identification model, and then uses the trained defect type identification model to identify the defect type of the tank bottom plate. This enables efficient and accurate identification of the defect type represented by the damage detection result after the tank bottom plate is damaged by ultrasonic phased array, thus improving the efficiency and accuracy of defect type identification.

[0101] Figure 4 This is a flowchart illustrating another method for identifying defects in the bottom plate of a storage tank, provided as an exemplary embodiment. The method can be executed by a computer device. The computer device can be... Figure 1 The computer equipment mentioned can also be other equipment capable of data analysis and processing.

[0102] like Figure 4 As shown in the embodiments of this application, the method for identifying defects in the tank bottom plate may include:

[0103] Step S401: The computer equipment acquires the defect dataset of the tank bottom plate.

[0104] In some embodiments, a set of focused images of defects in the tank bottom plate is obtained, and the defect types corresponding to the focused images are labeled to obtain a defect dataset.

[0105] Optionally, defect types can be labeled using single-label classification or mixed-label classification. For example, each defect data point can uniquely correspond to one defect type, or each defect data point can be labeled as a combination of defect types (e.g., simultaneously labeling cracks and uneven thickness).

[0106] In some embodiments, the computer device performs data augmentation on the defect data to expand the defect dataset.

[0107] For example, a computer device performs data augmentation operations such as flipping, adding noise, and rotating on the focused image of the damage detection to expand the defect dataset. For instance: Figure 5 A luminance image 501 with enhanced brightness, a noise image 502 with increased noise, a rotated image 503 rotated by 90 degrees, and a flipped image 504 are shown.

[0108] For further details on this step, please refer to step S201; it will not be repeated here.

[0109] Step S402: The computer device divides the defect dataset into at least two parts, using the data in one of the defect datasets as the test set data, and the data in the defect dataset other than the test set data as the training set data.

[0110] For example, each piece of data in the defect dataset is used as the test set data, and the data in the defect dataset other than the test set data is used as the training set data, resulting in multiple sets of training set data and test set data.

[0111] For example, the defect dataset is divided into 10 equal parts: X1, X2, X3, X4, X5, X6, X7, X8, X9, X10. One part of the dataset is used as the training set, and the remaining parts are used as the test set, resulting in 10 sets of training and test set data, as shown below:

[0112] Training set data Train1: X1, X2, X3, X4, X5, X6, X7, X8, X9;

[0113] Test set data Test1: X10.

[0114] Training set data Train2: X1, X2, X3, X4, X5, X6, X7, X8, X10;

[0115] Test set data Test2:X9.

[0116] Training set data Train3: X1, X2, X3, X4, X5, X6, X7, X9, X10;

[0117] Test set data Test3:X8.

[0118] Training set data Train4: X1, X2, X3, X4, X5, X6, X8, X9, X10;

[0119] Test set data Test4:X7.

[0120] Training set data Train5: X1, X2, X3, X4, X5, X7, X8, X9, X10;

[0121] Test set data Test5:X6.

[0122] Training set data Train6: X1, X2, X3, X4, X6, X7, X8, X9, X10;

[0123] Test set data Test6:X5.

[0124] Training set data Train7: X1, X2, X3, X5, X6, X7, X8, X9, X10;

[0125] Test set data Test7:X4.

[0126] Training set data Train8: X1, X2, X4, X5, X6, X7, X8, X9, X10;

[0127] Test set data Test8:X3.

[0128] Training set data Train9: X1, X3, X4, X5, X6, X7, X8, X9, X10;

[0129] Test set data Test9:X2.

[0130] Training set data Train10: X2, X3, X4, X5, X6, X7, X8, X9, X10; Test set data Test10: X1.

[0131] Step S403: The computer equipment extracts local binary pattern features from the defect data, and trains a defect type recognition model based on the local binary pattern features and the defect type.

[0132] Local binary pattern features refer to operators that describe the local texture features of an image, mainly by analyzing the gray-level change patterns of pixel neighborhoods to extract the texture information of the image.

[0133] For example, the formula for calculating local binary pattern features is expressed as follows:

[0134]

[0135] Among them, LBP r，p Let S(z) represent the local binary pattern feature, S(z) represent the sign function, p represent the number of neighboring pixels, r represent the neighborhood radius, and x represent the value of x. c This represents the grayscale value of the c-th neighboring pixel.

[0136] For example, local binary pattern features are extracted from defect data, and the local binary pattern features and defect types are input into the defect recognition model for training.

[0137] Step S404: The computer equipment performs an accuracy test on the trained defect type identification model based on the test set data.

[0138] In some embodiments, the defect identification model is trained using each set of training data, and the accuracy of the defect identification model is tested using the corresponding test set data.

[0139] For example, the recognition accuracy of the defect type identification model is calculated based on each set of test data, and the target accuracy is obtained by weighted averaging of the calculated multiple recognition accuracies.

[0140] For example, the formula for target accuracy is expressed as:

[0141] acc=1 / 10(a1+a2+a3+a4+a5+a6+a7+a8+a9+10)

[0142] Where acc represents the target accuracy, and a1…a10 represent the recognition accuracy of the defect recognition model trained on the training set data from group 1 to group 10.

[0143] For example, such as Figure 6 The figure shows the recognition accuracy and target accuracy (average) of the defect recognition model trained on training data from group 1 to group 10. Figure 6 The horizontal axis represents the false positive rate, and the vertical axis represents the true positive rate (recall rate).

[0144] The false positive rate refers to the proportion of samples that are actually negative but are incorrectly identified as positive. The true positive rate refers to the proportion of samples that are actually positive but are correctly identified as positive.

[0145] For example, a defect type identification model with a target accuracy greater than an accuracy threshold is identified as a defect type identification model that passes the accuracy test.

[0146] For further details on this step, please refer to step S204; it will not be repeated here.

[0147] Step S405: The computer equipment identifies the defect type of the tank bottom plate based on the defect type identification model that has passed the accuracy test.

[0148] For example, the bottom plate of the storage tank is scanned using an ultrasonic phased array to obtain a focused image. Local binary pattern features are extracted from the focused image, and these features are input into a defect type identification model that has passed an accuracy test. The model then outputs the identified defect type. For example: Figure 7 The image shown is a focused image 700, a grayscale image 701 of the focused image 700, and a local binary pattern feature 702 extracted from the focused image 700.

[0149] Experiments have verified that, for example Figure 8 The image shows the estimated score images of the average recognition rate of cracks and pores by different estimators. The technical solution of this application has an average recognition rate of about 95% for cracks and pores.

[0150] In summary, the technical solution provided in this application improves the training effect of the defect type recognition model by sequentially training it with multiple sets of training data. Furthermore, when training, testing, and using the defect type recognition model to identify defect types in the tank bottom plate, local binary pattern features of the tank bottom plate damage detection results are extracted. This extraction process is simple and efficient, reducing the computational load of the defect type recognition model. Simultaneously, the local binary pattern better represents the damage detection effect of the tank bottom plate, further improving the efficiency and accuracy of defect type recognition.

[0151] The foregoing mainly describes the solution provided in this application. Accordingly, this application also provides a tank bottom plate defect identification device, which is used to implement the above-described method embodiments.

[0152] like Figure 9 The diagram shows a structural schematic of a tank bottom plate defect identification device. The device may include an acquisition module 901, a processing module 902, and an identification module 903. The acquisition module 901 is used to perform… Figure 2 The operation of step S201 in the illustrated method, and Figure 4 The illustrated method includes step S401; the processing module 902 is used to execute... Figure 2 The operations of steps S202, S203 and S204, and Figure 4 The operations in steps S402, S403, and S404; the identification module 903 is used to execute Figure 2 The operation of step S205, and Figure 4 The operation of step S405.

[0153] In some embodiments, the tank bottom plate defect identification device includes hardware structures and / or software modules corresponding to the execution of each function in order to achieve the above-mentioned functions. Those skilled in the art will readily recognize that, based on the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0154] This application embodiment can divide the tank bottom plate defect identification device into functional modules according to the above method embodiment. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into a calibration module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0155] like Figure 10 As shown, the computer device provided in this application embodiment may include a processor 1001, a bus 1002, a communication interface 1003, and a memory 1004. The processor 1001, memory 1004, and communication interface 1003 communicate with each other via the bus 1002. It should be understood that this application does not limit the number of processors and memories in the network device.

[0156] Bus 1002 can be a PCI bus, an Extended Industry Standard Architecture (EISA) bus, or a UB bus, etc. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 10 The bus 1002 may be represented by a single line, but this does not mean that there is only one bus or one type of bus. The bus 1002 may include a path for transmitting information between various components of the network device (e.g., memory 1004, processor 1001, communication interface 1003).

[0157] The processor 1001 may include any one or more processors such as a CPU, graphics processing unit (GPU), microprocessor (MP), or digital signal processor (DSP).

[0158] The memory 1004 may include volatile memory, such as random access memory (RAM). The processor 1001 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0159] The communication interface 1003 uses transceiver modules, such as, but not limited to, network interface cards and transceivers, to enable communication between network devices and other devices or communication networks.

[0160] The memory 1004 stores executable program code, and the processor 1001 executes the executable program code to implement the functions of the aforementioned method embodiments. That is, the memory 1004 stores instructions for executing the above-described tank bottom plate defect identification method.

[0161] In another aspect, a computer-readable storage medium is provided, which stores at least one computer program, which is loaded and executed by a processor to implement the tank bottom plate defect identification method provided in the above-described method embodiments.

[0162] On another front, a computer program product is provided, which includes a computer program or instructions. When the computer program or instructions are executed by a processor, the tank bottom plate defect identification method provided in the above-described method embodiments is implemented.

[0163] Through the above description of the implementation methods, those skilled in the art will clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the module can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, modules, and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0164] Since the tank bottom plate defect identification module, computer-readable storage medium, and computer program product in the embodiments of the present invention can be applied to the above method, the technical effects obtained can also be referred to the above method embodiments. The embodiments of the present invention will not be repeated here.

[0165] The method steps in this embodiment can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Alternatively, the ASIC can reside in a network device. Of course, the processor and storage medium can also exist as discrete components in the network device.

[0166] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions of the embodiments of this application are performed, in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user equipment, or other programmable module. The computer program or instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, a computer program or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; it can also be an optical medium, such as a digital video disc (DVD); or it can be a semiconductor medium, such as a solid-state drive (SSD).

[0167] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for identifying defects in the bottom plate of a storage tank, characterized in that, The method includes: Obtain the defect dataset of the tank bottom plate. The defect dataset includes defect data and the defect type corresponding to the defect data. The defect data refers to the data obtained by detecting defects in the tank bottom plate using an ultrasonic phased array. The defect dataset is divided into training set data and test set data; Based on the training set data, a defect type identification model is trained; The accuracy of the trained defect type identification model is tested based on the test set data. The defect types of the tank bottom plate are identified based on a defect type identification model that has passed an accuracy test.

2. The method according to claim 1, characterized in that, The step of dividing the defect dataset to obtain training set data and test set data includes: The defect dataset is divided into at least two parts; The data in one of the defect datasets is used as the test set data, and the data in the defect dataset other than the test set data is used as the training set data.

3. The method according to claim 2, characterized in that, The step of using data from one of the defect datasets as the test set data and data from the other defect datasets as the training set data includes: Each piece of data in the defect dataset is used as the test set data, and the data in the defect dataset other than the test set data is used as the training set data, resulting in multiple sets of training set data and test set data.

4. The method according to claim 3, characterized in that, The training of the defect type identification model based on the training set data includes: Based on each set of training data, the defect type identification model is trained sequentially.

5. The method according to claim 3, characterized in that, The accuracy test of the trained defect type identification model based on the test set data includes: The recognition accuracy of the defect type recognition model is calculated based on the test set data for each group. The target accuracy is obtained by taking a weighted average of the calculated recognition accuracy rates. The defect type identification model whose target accuracy is greater than the accuracy threshold is identified as the defect type identification model that has passed the accuracy test.

6. The method according to any one of claims 1-5, characterized in that, The training of the defect type identification model based on the training set data includes: Extract local binary pattern features from the defect data; Based on the local binary pattern features and the defect type, the defect type recognition model is trained.

7. The method according to claim 6, characterized in that, The calculation formula for the local binary pattern features includes: Among them, LBP r，p Let S(z) represent the local binary pattern feature, S(z) represent the sign function, p represent the number of neighboring pixels, r represent the neighborhood radius, and x represent the value of x. c This represents the grayscale value of the c-th neighboring pixel.

8. A device for identifying defects in the bottom plate of a storage tank, characterized in that, The device includes: The acquisition module is used to acquire the defect dataset of the tank bottom plate. The defect dataset includes defect data and the defect type corresponding to the defect data. The defect data refers to the data obtained by detecting defects in the tank bottom plate through an ultrasonic phased array. The processing module is used to divide the defect dataset to obtain training set data and test set data; The processing module is also used to train a defect type identification model based on the training set data; The processing module is also used to perform accuracy testing on the trained defect type identification model based on the test set data; The identification module is used to identify the defect type of the tank bottom plate based on a defect type identification model that has passed an accuracy test.

9. A computer device, characterized in that, The computer device includes a processor and a memory, wherein the memory stores at least one computer program, and the at least one computer program is loaded and executed by the processor to implement the tank bottom plate defect identification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which is loaded and executed by a processor to implement the tank bottom plate defect identification method as described in any one of claims 1 to 7.

11. A computer program product, characterized in that, The computer program product includes a computer program or instructions that, when executed by a processor, implement the tank bottom plate defect identification method as described in any one of claims 1 to 7.

Images