Method, device, equipment and storage medium for identifying components in power grid wiring diagram

Abstract

The present invention discloses a method, device, equipment and storage medium for identifying components in a power grid wiring diagram. The method includes: creating a ledger for recording component information in the power grid wiring diagram, and setting multiple feature categories for the components in the ledger; based on the feature categories, using a pre-trained convolutional neural network to extract multi-dimensional semantic features of each component in the power grid wiring diagram, and obtaining a semantic feature matrix of each component; for each component, adopting a multi-criteria evaluation method, designing weights and reliability for each semantic feature of the component, and using weights and reliability to construct the evidence credibility of each semantic feature of each component; for each semantic feature of each component, inputting the evidence credibility and the semantic feature matrix into a pre-trained evidence reasoning classifier for the component, and using the evidence reasoning classifier to perform multiple binary classifications on the component, obtaining the classification results of the component and the feature category, and saving the classification results to the feature category of the ledger.

Description

Technical Field

[0001] The invention belongs to the technical field of pattern recognition, and in particular relates to a method for identifying components in a power grid wiring diagram and related equipment. Background Art

[0002] In the related technologies of digitalization of power grid wiring diagrams, the saved power grid wiring diagram blueprint is usually identified by equipment, and various components and related information are saved in a ledger for use in subsequent intelligent mapping.

[0003] However, in the relevant identification methods, the extraction of component features is not accurate enough, or the extracted component features cannot obtain sufficiently accurate identification and sorting results for each component in the ledger.

[0004] Based on this, a solution is needed that can accurately identify each component in the power grid wiring diagram and record it in the ledger. Summary of the invention

[0005] In view of this, the purpose of the present application is to provide a method, device and electronic equipment for identifying components in a power grid wiring diagram.

[0006] Based on the above purpose, the present application provides a method for identifying components in a power grid wiring diagram, including:

[0007] Creating a ledger to record the component information in the power grid wiring diagram, and setting multiple feature categories for the components in the ledger;

[0008] Based on the feature categories, a pre-trained convolutional neural network is used to extract multi-dimensional semantic features of each component in the power grid wiring diagram to obtain a semantic feature matrix of each component;

[0009] For each of the components, a multi-criteria evaluation method is adopted to design a weight and reliability for each of the semantic features of the component, and the weight and reliability are used to construct the evidence credibility of each of the semantic features of each of the components;

[0010] For each semantic feature of each component, the evidence credibility and the semantic feature matrix are input into the pre-trained evidence reasoning classifier for the component, and the evidence reasoning classifier is used to perform multiple binary classifications on the component to obtain the classification results of the component and the feature category, and the classification results are saved in the feature category of the ledger.

[0011] Furthermore, the method further comprises:

[0012] Placing the power grid wiring diagram in a coordinate system, and determining the position of each of the components in the coordinate system;

[0013] A position label matrix is ​​established using a Gaussian mask, and a position label of each of the components is obtained.

[0014] Furthermore, a pre-trained convolutional neural network is used to extract multi-dimensional semantic features of each element in the power grid wiring diagram, including:

[0015] Constructing the convolutional neural network by connecting a feature extraction network and a feature fusion network in series;

[0016] The shape, color, component type, text label, component name, scheduling label, angle and equipment information of the components in the power grid wiring diagram are used as dimensions of the semantic features to extract the semantic features.

[0017] Furthermore, a multi-criteria evaluation method is adopted to design weights and reliability for each semantic feature of each element, including:

[0018] For each of the elements, calculating the covariance matrix of all the semantic features;

[0019] Calculate the mean and variance of each semantic feature of the component;

[0020] Calculating the regularized Euclidean distance using the covariance matrix, and using the calculation result as the first component of each of the semantic features of the element;

[0021] Calculating a Fisher score using the mean and the variance as a second component of each of the semantic features of the element;

[0022] For each of the semantic features of the element, weight the first component and the second component to obtain reliability of the semantic feature;

[0023] A weight factor is designed for each of the semantic features of the element, and the weight of each of the semantic features is obtained by adjusting the weight factor.

[0024] Furthermore, using the evidence credibility and the semantic feature matrix, an evidence reasoning classifier is established, including:

[0025] For each of the elements, construct an evidence matrix about the element using the evidence credibility and the semantic feature matrix;

[0026] Wherein, the evidence matrix includes: each of the semantic features, the evidence reliability of each of the semantic features and all of the feature categories;

[0027] The evidence matrix is ​​input into a pre-trained evidence reasoning classifier to perform fusion weighting on the evidence confidence of all the semantic features of the component.

[0028] Furthermore, the component is subjected to multiple binary classifications using the evidence reasoning classifier, including:

[0029] For each of the components, using the two semantic features, roughly classify the component;

[0030] In response to determining that any of the feature categories includes more than one element, adding a different semantic feature to each of all the elements in the feature category and classifying them again;

[0031] In response to determining that all of the feature categories contain only one of the components, the classification of the components is terminated.

[0032] Based on the same inventive concept, the present application also provides a device for identifying components in a power grid wiring diagram, comprising:

[0033] The ledger creation module is configured to create a ledger that records the component information in the power grid wiring diagram, and set multiple feature categories for the components in the ledger.

[0034] The semantic feature extraction module is configured to extract multi-dimensional semantic features of each component in the power grid wiring diagram based on the feature category using a pre-trained convolutional neural network to obtain a semantic feature matrix of each component.

[0035] The evidence credibility building module is configured to adopt a multi-criteria evaluation method for each of the components, design weights and reliability for each of the semantic features of the component, and use the weights and reliability to build the evidence credibility of each of the semantic features of each of the components.

[0036] The classification module is configured to input the evidence credibility and the semantic feature matrix into a pre-trained evidence reasoning classifier for each semantic feature of each component, and use the evidence reasoning classifier to perform multiple binary classifications on the component to obtain the classification results of the component and the feature category, and save the classification results to the feature category of the ledger.

[0037] Based on the same inventive concept, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for identifying elements in a power grid wiring diagram as described in any one of the above items is implemented.

[0038] Based on the same inventive concept, the present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method for identifying elements in the above-mentioned power grid wiring diagram.

[0039] From the above, it can be seen that the method, device, equipment and storage medium for identifying components in the power grid wiring diagram provided by the present application are based on convolutional neural networks and evidence reasoning classifiers, and comprehensively consider the extraction and classification of each component in the power grid wiring diagram during the identification process to organize the ledger, so that while ensuring the accuracy of the extracted features, the accuracy of component classification is improved, and by performing multiple binary classifications on the different semantic features of the components, the operation of components from coarse identification to detailed identification is realized, thereby improving the accuracy of component identification in the power grid wiring diagram. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings constituting a part of the present application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0041] Figure 1 A flowchart of a method for identifying components in a power grid wiring diagram according to an embodiment of the present application;

[0042] Figure 2 This is a schematic diagram of the structure of a device for identifying components in a power grid wiring diagram according to an embodiment of the present application;

[0043] Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0044] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0045] The following detailed description is an exemplary description, which is intended to provide further detailed description of the present invention. Unless otherwise specified, all technical terms used in the present invention have the same meaning as those generally understood by those skilled in the art to which the present application belongs. The terms used in the present invention are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention.

[0046] As described in the background technology section, the method for identifying components in the related power grid wiring diagram is still difficult to meet the needs of actual mapping.

[0047] In the process of implementing this application, the applicant discovered that the main problem with the method for identifying components in the relevant power grid wiring diagram is that the relevant identification methods often use a large amount of training of convolutional neural networks to identify the semantic features of different components in the power grid wiring diagram, or achieve target detection by semantically segmenting different components in the image, but simple neural network recognition often makes it difficult to accurately restore the components in the ledger.

[0048] The applicant further discovered in the research that by combining a neural network with an evidence-based reasoning classifier, the semantic features of components can be effectively and accurately mapped to the categories of the ledger, thereby achieving effective identification of components in power grid wiring diagrams.

[0049] It can be understood that the method can be executed by any device, equipment, platform, or device cluster having computing and processing capabilities.

[0050] The technical method of the present application is described in detail below through specific embodiments.

[0051] refer to Figure 1 A method for identifying components in a power grid wiring diagram according to an embodiment of the present application comprises the following steps:

[0052] Step S101: create a ledger to record the component information in the power grid wiring diagram, and set multiple feature categories for the components in the ledger.

[0053] In an embodiment of the present application, a ledger is first created for the power grid wiring diagram to be processed in the database, the ledger is used to record relevant information of each component in the power grid wiring diagram, and multiple feature categories are set for each component in the ledger.

[0054] In this embodiment, for further processing, each element in the power grid wiring diagram may be used as a data sample, and the multiple feature categories set above may be used as attributes of each data sample.

[0055] Furthermore, the coordinate system is used to determine the location label of each component.

[0056] Specifically, the power grid connection diagram may be placed in a coordinate system, and the coordinate position of each component in the coordinate system may be determined according to the coordinate system.

[0057] Furthermore, based on the coordinate position, the position label of each component is calculated using a Gaussian mask, and the position label matrix of all components is formed by splicing.

[0058] Specifically, the following formula is used for the position label of each component:

[0059]

[0060]

[0061] Wherein, d represents the location label of each component, x and y are the center coordinates of the component in the coordinate system, and w and h are the width and height of the component in the circuit wiring diagram.

[0062] Furthermore, the above-mentioned position label Gaussian masks of each element are spliced ​​together to obtain the position label matrix D of all elements.

[0063] Step S102: Based on the feature categories, a pre-trained convolutional neural network is used to extract multi-dimensional semantic features of each component in the power grid wiring diagram to obtain a semantic feature matrix of each component.

[0064] In an embodiment of the present application, according to the feature categories set above, semantic features of the corresponding category are extracted for each element, wherein the semantic features extracted for each element are the same.

[0065] Specifically, a pre-trained convolutional neural network for feature extraction is used to extract the semantic features of the element. The convolutional neural network can be a DLA (Deep Learning Accelerator)-34 convolutional neural network composed of a feature extraction network and a scale feature fusion network connected in series.

[0066] Furthermore, the dimensions of the semantic features extracted by the convolutional neural network may include: shape, color, component type, text label, component name, scheduling label, angle and equipment information of components in the power grid wiring diagram.

[0067] The dispatch label is the dispatch number of the component; the angle is the direction in which the component is placed in the power grid wiring diagram.

[0068] Furthermore, the convolutional neural network will output a semantic feature matrix corresponding to each component.

[0069] Step S103: for each of the components, a multi-criteria evaluation method is adopted to design a weight and reliability for each of the semantic features of the component, and the weight and reliability are used to construct the evidence credibility of each of the semantic features of each of the components.

[0070] In the embodiment of the present application, for each element, each semantic feature of the element is used as an attribute in the evidence reasoning algorithm, and a weight and reliability are designed for each semantic feature, that is, each attribute.

[0071] Specifically, for reliability design, a multi-criteria evaluation approach can be adopted.

[0072] In this embodiment, the Fisher score of the attribute is first calculated using the following formula:

[0073]

[0074] Among them, k represents the kth attribute of the component, represents the inter-class variance of the attribute in all components, It represents the intra-class variance of the attribute in the specified category, and calculates the inter-class variance and intra-class variance according to the following formula:

[0075]

[0076]

[0077] Among them, x k Indicates the value of the component at the kth attribute. represents the mean value of the kth attribute of the i-th element in C categories, m k Represents the mean value of the kth attribute of all categories of components.

[0078] Furthermore, the Mahalanobis distance can be used as one of the criteria for multi-criteria evaluation. Specifically, the Mahalanobis distance can be regarded as a regularized Euclidean distance, so it can be calculated using the following formula:

[0079]

[0080] Among them, S is the covariance matrix between the attribute to be calculated and another attribute.

[0081] Furthermore, the Fisher score and Mahalanobis distance are weighted using the following formula to obtain the reliability of the attribute:

[0082]

[0083] Furthermore, the weight of each attribute can be calculated according to the following weight formula:

[0084]

[0085] Among them, w k represents the weight of the kth attribute, γ k It represents the weight factor of the attribute, which is set according to the actual situation and expert experience. k It indicates that the number of components belonging to each category is determined by the kth reliability. When the weight is 1, the attribute is the most reliable.

[0086] Furthermore, the weights of all attributes of the component follow the following constraints:

[0087] 1=w1+w2+…+w k

[0088] Furthermore, for this component, based on the reliability and weight of each attribute obtained as described above, the evidence credibility of each attribute is constructed. The evidence credibility describes the credibility of the attribute alone in supporting the classification of the component into a certain category. Specifically, the reliability of each attribute is adjusted according to the weight ratio of each attribute, and the adjusted reliability is used as the evidence credibility of the attribute.

[0089] Step S104: for each of the semantic features of each of the components, the evidence credibility and the semantic feature matrix are input into a pre-trained evidence reasoning classifier for the component, and the evidence reasoning classifier is used to perform multiple binary classifications on the component to obtain the classification results of the component and the feature category, and the classification results are saved in the feature category of the ledger.

[0090] In the embodiment of the present application, for each component, based on the evidence credibility of each attribute obtained above and combined with the semantic feature matrix, it can be simply merged into an evidence matrix about the component.

[0091] The evidence matrix includes all semantic features of the component and all feature categories in the circuit wiring diagram, as well as the corresponding evidence reliability under each category.

[0092] Furthermore, the evidence matrix is ​​input into a pre-trained evidence reasoning classifier to integrate all the attributes of the component and its evidence credibility, and then classify it.

[0093] Specifically, it can be described by the following formula:

[0094] P(x i )={(C n , β n，K ), n=1, ....C C}

[0095] Among them, P(x i ) represents the fusion result of all attributes of the i-th element, β n，K represents the fusion confidence of classifying the i-th element as category n after fusing K attributes, C n It represents the nth category among all C feature categories.

[0096] Furthermore, the category corresponding to the highest fusion confidence is used as the category for classifying the component.

[0097] Furthermore, based on the above-mentioned evidential reasoning classifier, each component may be roughly classified first, and in each rough classification, only the component is classified.

[0098] Specifically, for example, in a common power grid wiring diagram, there are types of components such as switches, power supply stations, transformers, etc., and the components can be classified into categories based on whether they are switches.

[0099] Furthermore, based on the classification of component types, the scheduling label of the component, that is, the scheduling number, can also be divided, the components under the same bus can be divided into one category, and other components not under the bus can be divided into another category. Furthermore, the text label of the component can be divided.

[0100] Multiple binary classification operations can be performed according to the actual situation of the required ledger. In the present embodiment, the feature category of the ledger is set so that each component corresponds to one feature category. Therefore, when all components and all feature categories reach a classification that uniquely corresponds to each other, the classification can be ended and the classification result can be saved in the value ledger.

[0101] It can be seen that the method for identifying components in the power grid wiring diagram of the embodiment of the present application is based on a convolutional neural network and an evidence reasoning classifier, and comprehensively considers the extraction and classification of each component in the power grid wiring diagram during the identification process to organize the ledger, thereby ensuring the accuracy of the extracted features while improving the accuracy of component classification, and performing multiple binary classifications on the different semantic features of the components, thereby realizing the operation from coarse identification to detailed identification of the components. By combining the neural network with the evidence reasoning classifier, the semantic features of the components can be effectively and accurately mapped to the categories of the ledger, thereby realizing the effective identification of the components in the power grid wiring diagram and improving the accuracy of the identification.

[0102] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0103] It should be noted that the above describes some embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0104] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, an embodiment of the present application further provides a device for identifying components in a power grid wiring diagram.

[0105] refer to Figure 2 The device for identifying components in a power grid wiring diagram includes: a ledger creation module 201, a semantic feature extraction module 202, an evidence credibility construction module 203 and a classification module 204.

[0106] The ledger creation module 201 is configured to: create a ledger for recording the component information in the power grid wiring diagram, and set multiple feature categories for the components in the ledger.

[0107] The semantic feature extraction module 202 is configured to: based on the feature category, use a pre-trained convolutional neural network to extract multi-dimensional semantic features of each component in the power grid wiring diagram to obtain a semantic feature matrix of each component.

[0108] The evidence credibility building module 203 is configured to: for each of the components, adopt a multi-criteria evaluation method to design weights and reliability for each of the semantic features of the component, and use the weights and reliability to build the evidence credibility of each of the semantic features of each of the components.

[0109] The classification module 204 is configured as follows: for each semantic feature of each component, the evidence credibility and the semantic feature matrix are input into a pre-trained evidence reasoning classifier for the component, and the evidence reasoning classifier is used to perform multiple binary classifications on the component to obtain the classification results of the component and the feature category, and the classification results are saved in the feature category of the ledger.

[0110] Furthermore, the above device may also include a position matching module, which is configured to:

[0111] The power grid wiring diagram is placed in a coordinate system, the position label of each component in the coordinate system is determined, a position label matrix is ​​established for each component using a Gaussian mask, and the position label matrix is ​​merged with the semantic feature matrix to obtain the position of each component in the ledger.

[0112] For the convenience of description, the above devices are described in terms of functions and are divided into various modules. Of course, when implementing the embodiments of the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0113] The device of the above embodiment is used to implement the corresponding method for identifying components in the power grid wiring diagram in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

[0114] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, an embodiment of the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for identifying elements in a power grid wiring diagram as described in any of the above-mentioned embodiments is implemented.

[0115] Figure 3 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0116] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.

[0117] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solution provided in the embodiment of the present application is implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0118] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0119] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0120] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0121] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiment of the present application, and does not necessarily include all the components shown in the figure.

[0122] The device of the above embodiment is used to implement the corresponding method for identifying components in the power grid wiring diagram in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

[0123] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method for identifying elements in the power grid wiring diagram as described in any of the above embodiments.

[0124] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0125] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the method for identifying components in a power grid wiring diagram as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0126] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0127] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present application will be implemented (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0128] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0129] It is known from common technical knowledge that the present invention can be implemented by other embodiments that do not deviate from its spirit or essential features. Therefore, the above disclosed embodiments are only illustrative in all respects and are not exclusive. All changes within the scope of the present invention or within the scope equivalent to the present invention are included in the present invention.

Claims

1. A method for identifying components in a power grid wiring diagram, characterized in that: include: Creating a ledger to record the component information in the power grid wiring diagram, and setting multiple feature categories for the components in the ledger; Based on the feature categories, a pre-trained convolutional neural network is used to extract multi-dimensional semantic features of each component in the power grid wiring diagram to obtain a semantic feature matrix of each component; For each of the components, a multi-criteria evaluation method is adopted to design a weight and reliability for each of the semantic features of the component, and the weight and reliability are used to construct the evidence credibility of each of the semantic features of each of the components; For each of the semantic features of each of the components, the evidence credibility and the semantic feature matrix are input into a pre-trained evidence reasoning classifier for the component, and the evidence reasoning classifier is used to perform multiple binary classifications on the component to obtain classification results of the component and the feature category, and the classification results are saved in the feature category of the ledger; Also includes: The power grid connection diagram is placed in a coordinate system, and the position label of each element in the coordinate system is determined, wherein the position label of each element is calculated using the following formula: d represents the location label of each component, x and y are the center coordinates of the component in the coordinate system, w and h are the width and height of the component in the circuit wiring diagram; Using a Gaussian mask to establish a position label matrix for each of the components; The position label matrix is ​​merged with the semantic feature matrix to obtain the position of each component in the ledger; The multi-criteria evaluation method is adopted to design weights and reliability for each semantic feature of each element, including: For each of the elements, calculating the covariance matrix of all the semantic features; Calculate the mean and variance of each semantic feature of the component: The covariance matrix is ​​used to calculate the regularized Euclidean distance E M (k), and taking the calculation result as the first component of each of the semantic features of the element, wherein, k represents the kth attribute of the component, m is the mean, and S is the covariance matrix between the attribute to be calculated and another attribute; Calculating a Fisher score using the mean and the variance as a second component of each of the semantic features of the element; For each of the semantic features of the element, weighting the first component and the second component to obtain reliability of the semantic feature; A weight factor is designed for each of the semantic features of the element, and the weight of each of the semantic features is obtained by adjusting the weight factor, wherein the formula is: w k represents the weight of the kth attribute, γ k represents the weight factor of the kth attribute, Q k It means that the number of components belonging to each category is determined by the kth reliability, and the constraint condition of the weight of all attributes of the components is 1=w1+w2+...+w k .

2. The method according to claim 1, characterized in that The method of extracting multi-dimensional semantic features of each element in the power grid wiring diagram by using a pre-trained convolutional neural network includes: Constructing the convolutional neural network by connecting a feature extraction network and a feature fusion network in series; The shape, color, component type, text label, component name, scheduling label, angle and equipment information of the components in the power grid wiring diagram are used as dimensions of the semantic features to extract the semantic features.

3. The method according to claim 1, characterized in that The method of establishing an evidence reasoning classifier by using the evidence credibility and the semantic feature matrix includes: For each of the elements, construct an evidence matrix about the element using the evidence credibility and the semantic feature matrix; Wherein, the evidence matrix includes: each of the semantic features, the evidence reliability of each of the semantic features and all of the feature categories; The evidence matrix is ​​input into a pre-trained evidence reasoning classifier to perform fusion weighting on the evidence confidence of all the semantic features of the component.

4. The method according to claim 1, characterized in that: The using the evidential reasoning classifier to perform multiple binary classifications on the components includes: For each of the components, using the two semantic features, roughly classify the component; In response to determining that any of the feature categories includes more than one element, adding a different semantic feature to each of all the elements in the feature category and classifying them again; In response to determining that all of the feature categories contain only one of the components, the classification of the components is terminated.

5. A device for identifying components in a power grid wiring diagram, characterized in that: A method for identifying a component in a power grid wiring diagram according to any one of claims 1 to 4, comprising: A ledger creation module is configured to create a ledger that records the component information in the power grid wiring diagram, and set multiple feature categories for the components in the ledger; A semantic feature extraction module is configured to extract multi-dimensional semantic features of each of the components in the power grid wiring diagram based on the feature categories using a pre-trained convolutional neural network to obtain a semantic feature matrix for each of the components; an evidence credibility building module, configured to adopt a multi-criteria evaluation method for each of the components, design a weight and reliability for each of the semantic features of the component, and build the evidence credibility of each of the semantic features of each of the components using the weight and the reliability; The classification module is configured to input the evidence credibility and the semantic feature matrix into a pre-trained evidence reasoning classifier for each semantic feature of each component, and use the evidence reasoning classifier to perform multiple binary classifications on the component to obtain the classification results of the component and the feature category, and save the classification results to the feature category of the ledger.

6. The device according to claim 5, characterized in that Also includes: The position matching module is configured to place the power grid wiring diagram in a coordinate system, determine the position label of each of the components in the coordinate system, establish a position label matrix for each of the components using a Gaussian mask, and fuse the position label matrix with the semantic feature matrix to obtain the position of each of the components in the ledger.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable by the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 4 is implemented.

8. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions for causing a computer to execute the method according to any one of claims 1 to 4.

Images