Artificial intelligence supervision method and system for electric power information engineering quality

Abstract

The embodiment of the invention provides an artificial intelligence supervision method and system for electric power information engineering quality, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring project engineering data of an electric power information engineering project; analyzing the project engineering data to obtain project detailed rule requirements; determining a project quality standard according to the project detailed rule requirements; reading standard data from a preset standard file database according to the engineering quality standard; and executing supervision operation according to the standard data. Compared with the prior art, the method and the system provided by the invention have the advantages that the standard data is directly extracted from the project engineering data, and human intervention is not needed, so that the supervision efficiency and accuracy of the electric power information engineering project are improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, storage medium and processor for processing field information. Background Art

[0002] The construction of power information engineering projects requires the corresponding supervision engineers. The current supervision work is all done manually, lacking an effective unmanned supervision mechanism. Even if there is some video surveillance, human intervention is still required due to the lack of engineering standards. This obviously restricts the efficiency and accuracy of engineering supervision. Summary of the invention

[0003] The purpose of the embodiments of the present invention is to provide an artificial intelligence supervision method and system for the quality of electric power information engineering, which can improve the supervision efficiency of the quality of electric power information engineering.

[0004] In order to achieve the above-mentioned purpose, an embodiment of the present invention provides an artificial intelligence supervision method for power information engineering quality, comprising:

[0005] Obtain project engineering data for power information engineering projects;

[0006] Analyze the project engineering data to obtain project detailed requirements;

[0007] Determine the engineering quality standards according to the requirements of the project details;

[0008] Reading standard data from a preset standard file database according to the engineering quality standard;

[0009] The supervision operation is performed according to the standard data.

[0010] Optionally, obtaining project engineering data of the electric power information engineering project includes:

[0011] Obtaining a construction plan for the electric power information engineering project;

[0012] Obtaining a procurement plan for the electric power information engineering project;

[0013] Searching for construction types from the construction plan;

[0014] Search for material categories from the procurement plan.

[0015] Optionally, the engineering quality standards are determined according to the project details, including:

[0016] Determining whether construction requirements matching the construction type are recorded in the construction plan;

[0017] When it is determined that the construction plan records a construction requirement that matches the construction type, associating the construction requirement with the construction type;

[0018] If it is determined that the construction plan does not record a construction requirement matching the construction type, setting the construction requirement associated with the construction type to empty;

[0019] Determine whether material requirements for the material type are recorded in the procurement plan;

[0020] When it is determined that a material requirement matching the material type is recorded in the procurement plan, associating the material requirement with the material type;

[0021] When it is determined that the material requirement matching the material type is not recorded in the procurement plan, the material requirement associated with the material type is set to blank.

[0022] Optionally, reading standard data from a preset standard file database according to the engineering quality standard includes:

[0023] Reading a corresponding standard file in the standard file database according to the construction type;

[0024] Determine whether the construction requirement corresponding to the construction type is empty;

[0025] When it is determined that the construction requirement corresponding to the construction type is empty, using the construction standard in the standard file as the standard data;

[0026] When it is determined that the construction requirement corresponding to the construction type is not empty, selecting the higher one of the construction standard and the construction requirement as the standard data;

[0027] Reading a corresponding standard file in the standard file database according to the material type;

[0028] Determine whether the construction requirement corresponding to the material type is empty;

[0029] In the case where it is determined that the material requirement corresponding to the material type is empty, the material standard in the standard file is used as the standard data;

[0030] When it is determined that the material requirement corresponding to the material type is not empty, a higher one of the material standard and the material requirement is selected as the standard data.

[0031] Optionally, performing a supervision operation according to the standard data includes:

[0032] According to the construction requirements in the standard data, obtain on-site surveillance videos every day;

[0033] Extracting a sample image from the surveillance video;

[0034] The sampled image is input into a preset image recognition model to complete the supervision operation of the on-site construction situation.

[0035] Optionally, extracting a sample image from the surveillance video includes:

[0036] Detect whether someone triggers an entry or exit event in the preset image anchor frame;

[0037] When it is determined that a person triggers an entry or exit event in the image anchor frame, a sampling timestamp is generated;

[0038] Traversing every two sampling timestamps, and determining whether the time length between the two sampling timestamps is greater than a preset length;

[0039] In the case where it is determined that the time length between the two sampling timestamps is greater than the preset length, randomly inserting a new sampling timestamp between the two sampling timestamps;

[0040] Three sampling images are captured at every two sampling timestamps.

[0041] Optionally, the image recognition model includes:

[0042] A first input layer, used to receive the first image of the three sampled images;

[0043] The second input layer is used to receive the second image of the three sampled images;

[0044] The third input layer is used to receive the third image of the three sampled images;

[0045] a first convolution module, one end of which is connected to the first input layer and is used for extracting features from the first image;

[0046] a second convolution module, one end of which is connected to the second input layer, and is used for extracting features from the second image;

[0047] a third convolution module, one end of which is connected to the third input layer, and is used for extracting features from a third image;

[0048] A first feature combination unit, connected to the feature map outputs of at least two convolution units in the first convolution module, and configured to combine the at least two feature map outputs to obtain a corresponding combined feature;

[0049] A second feature combination unit, connected to the feature map outputs of at least two convolution units in the second convolution module, and configured to combine the at least two feature map outputs to obtain a corresponding combined feature;

[0050] A third feature combination unit, connected to the feature map outputs of at least two convolution units in the third convolution module, and used to combine at least two feature map outputs to obtain a corresponding combined feature;

[0051] A first feature detection unit, connected to the output end of the first feature combination unit, and configured to perform a feature detection operation on the combined feature output by the first feature combination unit;

[0052] A second feature detection unit, connected to the output end of the second feature combination unit, and used to perform a feature detection operation on the combined feature output by the second feature combination unit;

[0053] A third feature detection unit is connected to the output end of the third feature combination unit and is used to perform a feature detection operation on the combined feature output by the third feature combination unit;

[0054] The fusion judgment unit is connected to the first feature detection unit, the second feature detection unit and the third feature detection unit, and is used to perform fusion judgment on the results of the feature detection operation to determine whether there is any violation of standard data in the current monitoring video.

[0055] Optionally, the first convolution module includes three convolution units connected in series, the second convolution module includes six convolution units connected in series, and the third convolution module includes nine convolution units connected in series, each of which includes:

[0056] The first residual convolution layer is used to perform the first residual convolution operation on the input image;

[0057] A second residual convolution layer, used to perform a second residual convolution operation on the result of the first residual convolution operation;

[0058] The two-dimensional convolution layer is used to perform a two-dimensional convolution operation according to the result of the first residual convolution operation and the result of the second residual convolution operation, and to fuse the result of the two-dimensional convolution operation and the result of the second residual convolution operation for output.

[0059] Optionally, performing a supervision operation according to the standard data includes:

[0060] Obtain the supervision log uploaded daily by the supervision engineer;

[0061] Obtaining the material model and size in the supervision log;

[0062] The on-site material situation is supervised according to the material model and size and the standard data.

[0063] On the other hand, the present invention also provides an artificial intelligence supervision system for the quality of electric power information engineering, the system comprising a processor, and the processor is configured to execute any of the methods described.

[0064] Through the above technical solution, the embodiment of the present invention provides an artificial intelligence supervision method and system for the quality of electric power information engineering. The method and system analyze the project engineering data of the electric power information engineering project by constructing it, and then combines the matching of the standard file database to achieve the acquisition of standard data for unmanned supervision operation. Compared with the prior art, the method and system provided by the present invention directly extracts standard data from project engineering data without human intervention, thereby improving the supervision efficiency and accuracy of the electric power information engineering project.

[0065] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0067] Figure 1 is a flow chart of an artificial intelligence supervision method for power information engineering quality according to an embodiment of the present invention;

[0068] Figure 2 is a flow chart of a method for determining engineering quality standards according to one embodiment of the present invention;

[0069] Figure 3 is a flow chart of a method for obtaining standard data according to one embodiment of the present invention;

[0070] Figure 4 is a flow chart of a method for extracting a sample image according to one embodiment of the present invention;

[0071] Figure 5 is a structural block diagram of an image recognition model according to one embodiment of the present invention;

[0072] Figure 6 is a block diagram of a convolution unit according to an embodiment of the present invention;

[0073] Figure 7 Detailed Description of the Invention The invention relates to a first feature detection unit, a second feature detection unit and a third feature detection unit according to an embodiment of the present invention. DETAILED DESCRIPTION

[0074] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0075] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. In the embodiments of this application, some existing solutions in the industry such as certain software, components, and models may be mentioned, which should be considered as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but it does not mean that the applicant has or will necessarily use the solution.

[0076] like Figure 1 The figure is a flow chart of an artificial intelligence supervision method for power information engineering quality according to one embodiment of the present invention. Figure 1 In the method, the method may include the following steps:

[0077] In step S10, the project engineering data of the electric power information engineering project is obtained;

[0078] In step S11, the project engineering data is parsed to obtain the project detailed requirements;

[0079] In step S12, the engineering quality standard is determined according to the project details;

[0080] In step S13, standard data is read from a preset standard file database according to the engineering quality standard;

[0081] In step S14, the supervisory operation is performed according to the standard data.

[0082] In this Figure 1 In the method shown, step S10 can be used to obtain project engineering data of the power information engineering project. Since the project engineering data generally includes a construction plan and a procurement plan, step S10 can also be to first obtain the construction plan of the power information engineering project, search for construction types from the construction plan, and then obtain the procurement plan of the power information engineering project, and search for material types from the procurement plan.

[0083] Step S11 can be used to parse the project engineering data to obtain the project detailed requirements. Step S12 can be used to determine the engineering quality standards according to the project detailed requirements. The project detailed requirements can be used to represent the construction requirements of each construction type in the construction plan of the project and the material requirements of each material type in the procurement plan. However, since the default construction requirements and material requirements are sometimes used in the project engineering data, the corresponding project detailed requirements cannot be parsed from the project engineering data. Therefore, in one example of the present invention, the steps S11 and S12 can also include the following: Figure 2 In the steps shown in Figure 2 In the step S11 and the step S12, the step S11 and the step S12 may further include the following steps:

[0084] In step S20, it is determined whether there is a construction requirement matching the construction type recorded in the construction plan;

[0085] In step S21, when it is determined that the construction plan records construction requirements that match the construction type, the construction requirements are associated with the construction type;

[0086] In step S22, if it is determined that the construction plan does not record a construction requirement matching the construction type, the construction requirement associated with the construction type is set to blank;

[0087] In step S23, it is determined whether material requirements corresponding to the material type are recorded in the procurement plan;

[0088] In step S24, when it is determined that there is a material requirement matching the material type recorded in the procurement plan, the material requirement is associated with the material type;

[0089] In step S25, when it is determined that no material requirement matching the material type is recorded in the procurement plan, the material requirement associated with the material type is set to blank.

[0090] Step S13 is used to read standard data from a preset standard file database according to the engineering quality standard. Since the standard data from the standard file database may conflict with the project details obtained from the project engineering data, step S13 may also include: Figure 3 In the method shown in Figure 3 In the step S13, the step S13 may include the following steps:

[0091] In step S30, the corresponding standard file is read from the standard file database according to the construction type;

[0092] In step S31, it is determined whether the construction requirement corresponding to the construction type is empty;

[0093] In step S32, when it is determined that the construction requirements corresponding to the construction type are empty, the construction standards in the standard file are used as standard data;

[0094] In step S33, when it is determined that the construction requirement corresponding to the construction type is not empty, the higher of the construction standard and the construction requirement is selected as the standard data;

[0095] In step S34, the corresponding standard file is read from the standard file database according to the material type;

[0096] In step S35, it is determined whether the construction requirement corresponding to the material type is empty;

[0097] In step S36, when it is determined that the material requirement corresponding to the material type is empty, the material standard in the standard file is used as the standard data;

[0098] In step S37, when it is determined that the material requirement corresponding to the material type is not empty, the higher of the material standard and the material requirement is selected as the standard data.

[0099] Step S14 can be used to perform supervision operations according to standard data. Specifically, considering that the aforementioned standard data includes construction requirements and material requirements, in one example of the present invention, step S14 can be to obtain on-site surveillance videos daily for the construction requirements in the standard data, then extract sample images from the surveillance videos, and finally input the sample images into a preset image recognition model to complete the supervision operation of the on-site construction situation.

[0100] Among them, since the image recognition model itself consumes a lot of computing power, if high-frequency real-time monitoring is directly adopted, it will obviously cause a large power consumption, and it will also put forward higher requirements on the monitoring equipment. Therefore, in this example, it is necessary to extract a sample image from the monitoring video. The specific method for extracting the sample image can be a variety of forms known to those skilled in the art. In one example of the present invention, the method for extracting the sample image may include: Figure 4 The steps shown in:

[0101] In step S40, it is detected whether there is a person in the preset image anchoring frame triggering an entry and exit event;

[0102] In step S41, when it is determined that someone in the image anchor frame triggers an entry or exit event, a sampling timestamp is generated;

[0103] In step S42, traverse every two sampling timestamps to determine whether the time length between the two sampling timestamps is greater than a preset length;

[0104] In step S43, when it is determined that the time length between two sampling timestamps is greater than a preset length, a new sampling timestamp is randomly inserted between the two sampling timestamps;

[0105] In step S44, three sampling images are captured at every two sampling time stamps.

[0106] For the image recognition model, it can be a variety of structures known to those skilled in the art. In one example of the present invention, the image recognition model may include a first input layer 10, a second input layer 20, a third input layer 30, a first convolution module 11, a second convolution module 21, a third convolution module 31, a first feature combination unit 12, a second feature combination unit 22, a third feature combination unit 32, a first feature detection unit 13, a second feature detection unit 23, a third feature detection unit 33 and a fusion judgment unit 40. Among them, the first input layer 10 can be used to receive the first image of the three sampled images. The second input layer 20 can be used to receive the second image of the three sampled images. The third input layer 30 can be used to receive the third image of the three sampled images. One end of the first convolution module 11 can be connected to the first input layer 10 for feature extraction from the first image. One end of the second convolution module 21 can be connected to the second input layer 20 for feature extraction from the second image. One end of the third convolution module 31 can be connected to the third input layer 20 for feature extraction from the third image. The first feature combination unit 12 may be connected to the feature map outputs of at least two convolution units in the first convolution module 11, and is used to combine the at least two feature map outputs to obtain a corresponding combined feature. The second feature combination unit 22 may be connected to the feature map outputs of at least two convolution units in the second convolution module 21, and is used to combine the at least two feature map outputs to obtain a corresponding combined feature. The third feature combination unit 32 may be connected to the feature map outputs of at least two convolution units in the third convolution module 31, and is used to combine the at least two feature map outputs to obtain a corresponding combined feature. The first feature detection unit 13 may be connected to the output end of the first feature combination unit 12, and is used to perform a feature detection operation on the combined feature output by the first feature combination unit 12. The second feature detection unit 23 may be connected to the output end of the second feature combination unit 22, and is used to perform a feature detection operation on the combined feature output by the second feature combination unit 22. The third feature detection unit 33 may be connected to the output end of the third feature combination unit 32, and is used to perform a feature detection operation on the combined feature output by the third feature combination unit 32. The fusion judgment unit 40 can be connected to the first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 to perform fusion judgment on the results of the feature detection operation to determine whether there is any violation of standard data in the current monitoring video.

[0107] In this Figure 5 In the image recognition model shown, the first input layer 10, the second input layer 20 and the third input layer 30 respectively obtain three sample images, and extract features based on different scales. Finally, through the fusion judgment of the fusion judgment unit 40, accurate recognition of the surveillance video is achieved, avoiding misrecognition of a single image. Figure 5 In the image recognition model shown, the first convolution module 11 may include three convolution units connected in series, the second convolution module 21 may include six convolution units connected in series, and the third convolution module 31 may include nine convolution units connected in series. The structural block diagram of each convolution unit may be as follows: Figure 6 As shown. Figure 6 In the embodiment, each convolution unit may include a first residual convolution layer 50, a second residual convolution layer 51 and a two-dimensional convolution layer 52. The first residual convolution layer 50 may be used to perform a first residual convolution operation on the input image. The second residual convolution layer 51 may be used to perform a second residual convolution operation on the result of the first residual convolution operation. The two-dimensional convolution layer 52 may be used to perform a two-dimensional convolution operation according to the result of the first residual convolution operation and the result of the second residual convolution operation, and fuse the result of the two-dimensional convolution operation and the result of the second residual convolution operation to output.

[0108] In addition, for this Figure 5 The first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 in the image recognition model shown. The specific forms of the first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 can be various forms known to those skilled in the art. In one example of the present invention, Figure 7 As shown, the first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 may include a convolution and pooling layer 60, a first matrix multiplication layer 61, a global average pooling layer 62, a one-dimensional convolution layer 63, a second matrix multiplication layer 64 and an output layer 65. Among them, the convolution and pooling layer 60 is used to perform convolution and pooling operations on the result of feature extraction, and the first matrix multiplication layer 61 can perform matrix multiplication on the result of feature extraction and the result of convolution and pooling operations, thereby corresponding to the multiplication feature. The global average pooling layer 62 can be used to perform a global average pooling operation on the multiplication feature to obtain a pooling feature. The one-dimensional convolution layer 63 can be used to perform a one-dimensional convolution operation on the pooling feature, and the second matrix multiplication layer 64 can be used to perform a matrix multiplication operation on the result of the one-dimensional convolution operation and the multiplication feature. The output layer 65 can be used to perform a classification operation based on the result of the matrix multiplication operation to obtain the result of feature detection.

[0109] In addition, for the material requirements in the standard data, step S14 can be to obtain the supervision log uploaded daily by the supervision engineer, then obtain the material model and size in the supervision log, and finally perform supervision operations on the on-site material conditions based on the material model and size and the standard data.

[0110] On the other hand, the present invention also provides an artificial intelligence supervision system for power information engineering quality, the system comprising a processor, the processor being configured to execute any of the methods described above. Specifically, the method may include the following steps:

[0111] In step S10, the project engineering data of the electric power information engineering project is obtained;

[0112] In step S11, the project engineering data is parsed to obtain the project detailed requirements;

[0113] In step S12, the engineering quality standard is determined according to the project details;

[0114] In step S13, standard data is read from a preset standard file database according to the engineering quality standard;

[0115] In step S14, the supervisory operation is performed according to the standard data.

[0116] In this Figure 1 In the method shown, step S10 can be used to obtain project engineering data of the power information engineering project. Since the project engineering data generally includes a construction plan and a procurement plan, step S10 can also be to first obtain the construction plan of the power information engineering project, search for construction types from the construction plan, and then obtain the procurement plan of the power information engineering project, and search for material types from the procurement plan.

[0117] Step S11 can be used to parse the project engineering data to obtain the project detailed requirements. Step S12 can be used to determine the engineering quality standards according to the project detailed requirements. The project detailed requirements can be used to represent the construction requirements of each construction type in the construction plan of the project and the material requirements of each material type in the procurement plan. However, since the default construction requirements and material requirements are sometimes used in the project engineering data, the corresponding project detailed requirements cannot be parsed from the project engineering data. Therefore, in one example of the present invention, the steps S11 and S12 can also include the following: Figure 2 In the steps shown in Figure 2 In the step S11 and the step S12, the step S11 and the step S12 may further include the following steps:

[0118] In step S20, it is determined whether there is a construction requirement matching the construction type recorded in the construction plan;

[0119] In step S21, when it is determined that the construction plan records construction requirements that match the construction type, the construction requirements are associated with the construction type;

[0120] In step S22, if it is determined that the construction plan does not record a construction requirement matching the construction type, the construction requirement associated with the construction type is set to blank;

[0121] In step S23, it is determined whether material requirements corresponding to the material type are recorded in the procurement plan;

[0122] In step S24, when it is determined that there is a material requirement matching the material type recorded in the procurement plan, the material requirement is associated with the material type;

[0123] In step S25, when it is determined that no material requirement matching the material type is recorded in the procurement plan, the material requirement associated with the material type is set to blank.

[0124] Step S13 is used to read standard data from a preset standard file database according to the engineering quality standard. Since the standard data from the standard file database may conflict with the project details obtained from the project engineering data, step S13 may also include: Figure 3 In the method shown in Figure 3 In the step S13, the step S13 may include the following steps:

[0125] In step S30, the corresponding standard file is read from the standard file database according to the construction type;

[0126] In step S31, it is determined whether the construction requirement corresponding to the construction type is empty;

[0127] In step S32, when it is determined that the construction requirements corresponding to the construction type are empty, the construction standards in the standard file are used as standard data;

[0128] In step S33, when it is determined that the construction requirement corresponding to the construction type is not empty, the higher of the construction standard and the construction requirement is selected as the standard data;

[0129] In step S34, the corresponding standard file is read from the standard file database according to the material type;

[0130] In step S35, it is determined whether the construction requirement corresponding to the material type is empty;

[0131] In step S36, when it is determined that the material requirement corresponding to the material type is empty, the material standard in the standard file is used as the standard data;

[0132] In step S37, when it is determined that the material requirement corresponding to the material type is not empty, the higher of the material standard and the material requirement is selected as the standard data.

[0133] Step S14 can be used to perform supervision operations according to standard data. Specifically, considering that the aforementioned standard data includes construction requirements and material requirements, in one example of the present invention, step S14 can be to obtain on-site surveillance videos daily for the construction requirements in the standard data, then extract sample images from the surveillance videos, and finally input the sample images into a preset image recognition model to complete the supervision operation of the on-site construction situation.

[0134] Among them, since the image recognition model itself consumes a lot of computing power, if high-frequency real-time monitoring is directly adopted, it will obviously cause a large power consumption, and it will also put forward higher requirements on the monitoring equipment. Therefore, in this example, it is necessary to extract a sample image from the monitoring video. The specific method for extracting the sample image can be a variety of forms known to those skilled in the art. In one example of the present invention, the method for extracting the sample image may include: Figure 4 The steps shown in:

[0135] In step S40, it is detected whether there is a person in the preset image anchoring frame triggering an entry and exit event;

[0136] In step S41, when it is determined that someone in the image anchor frame triggers an entry or exit event, a sampling timestamp is generated;

[0137] In step S42, traverse every two sampling timestamps to determine whether the time length between the two sampling timestamps is greater than a preset length;

[0138] In step S43, when it is determined that the time length between two sampling timestamps is greater than a preset length, a new sampling timestamp is randomly inserted between the two sampling timestamps;

[0139] In step S44, three sampling images are captured at every two sampling time stamps.

[0140] For the image recognition model, it can be a variety of structures known to those skilled in the art. In one example of the present invention, the image recognition model may include a first input layer 10, a second input layer 20, a third input layer 30, a first convolution module 11, a second convolution module 21, a third convolution module 31, a first feature combination unit 12, a second feature combination unit 22, a third feature combination unit 32, a first feature detection unit 13, a second feature detection unit 23, a third feature detection unit 33 and a fusion judgment unit 40. Among them, the first input layer 10 can be used to receive the first image of the three sampled images. The second input layer 20 can be used to receive the second image of the three sampled images. The third input layer 30 can be used to receive the third image of the three sampled images. One end of the first convolution module 11 can be connected to the first input layer 10 for feature extraction from the first image. One end of the second convolution module 21 can be connected to the second input layer 20 for feature extraction from the second image. One end of the third convolution module 31 can be connected to the third input layer 20 for feature extraction from the third image. The first feature combination unit 12 may be connected to the feature map outputs of at least two convolution units in the first convolution module 11, and is used to combine the at least two feature map outputs to obtain a corresponding combined feature. The second feature combination unit 22 may be connected to the feature map outputs of at least two convolution units in the second convolution module 21, and is used to combine the at least two feature map outputs to obtain a corresponding combined feature. The third feature combination unit 32 may be connected to the feature map outputs of at least two convolution units in the third convolution module 31, and is used to combine the at least two feature map outputs to obtain a corresponding combined feature. The first feature detection unit 13 may be connected to the output end of the first feature combination unit 12, and is used to perform a feature detection operation on the combined feature output by the first feature combination unit 12. The second feature detection unit 23 may be connected to the output end of the second feature combination unit 22, and is used to perform a feature detection operation on the combined feature output by the second feature combination unit 22. The third feature detection unit 33 may be connected to the output end of the third feature combination unit 32, and is used to perform a feature detection operation on the combined feature output by the third feature combination unit 32. The fusion judgment unit 40 can be connected to the first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 to perform fusion judgment on the results of the feature detection operation to determine whether there is any violation of standard data in the current monitoring video.

[0141] In this Figure 5In the image recognition model shown, the first input layer 10, the second input layer 20 and the third input layer 30 respectively obtain three sample images, and extract features based on different scales. Finally, through the fusion judgment of the fusion judgment unit 40, accurate recognition of the surveillance video is achieved, avoiding misrecognition of a single image. Figure 5 In the image recognition model shown, the first convolution module 11 may include three convolution units connected in series, the second convolution module 21 may include six convolution units connected in series, and the third convolution module 31 may include nine convolution units connected in series. The structural block diagram of each convolution unit may be as follows: Figure 6 As shown. Figure 6 In the embodiment, each convolution unit may include a first residual convolution layer 50, a second residual convolution layer 51 and a two-dimensional convolution layer 52. The first residual convolution layer 50 may be used to perform a first residual convolution operation on the input image. The second residual convolution layer 51 may be used to perform a second residual convolution operation on the result of the first residual convolution operation. The two-dimensional convolution layer 52 may be used to perform a two-dimensional convolution operation according to the result of the first residual convolution operation and the result of the second residual convolution operation, and fuse the result of the two-dimensional convolution operation and the result of the second residual convolution operation to output.

[0142] In addition, for this Figure 5 The first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 in the image recognition model shown. Among them, the specific forms of the first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 can be various forms known to those skilled in the art. In one example of the present invention, the first feature detection unit 13, the second feature detection unit 23 and the third feature detection unit 33 may include a convolution and pooling layer 60, a first matrix multiplication layer 61, a global average pooling layer 62, a one-dimensional convolution layer 63, a second matrix multiplication layer 64 and an output layer 65. Among them, the convolution and pooling layer 60 is used to perform convolution and pooling operations on the result of feature extraction, and the first matrix multiplication layer 61 can perform matrix multiplication on the result of feature extraction and the result of convolution and pooling operations, thereby corresponding to the multiplication feature. The global average pooling layer 62 can be used to perform a global average pooling operation on the multiplication feature to obtain a pooled feature. The one-dimensional convolution layer 63 can be used to perform a one-dimensional convolution operation on the pooled features, and the second matrix multiplication layer 64 can be used to perform a matrix multiplication operation on the result of the one-dimensional convolution operation and the multiplication feature. The output layer 65 can perform a classification operation based on the result of the matrix multiplication operation to obtain a feature detection result.

[0143] In addition, for the material requirements in the standard data, step S14 can be to obtain the supervision log uploaded daily by the supervision engineer, then obtain the material model and size in the supervision log, and finally perform supervision operations on the on-site material conditions based on the material model and size and the standard data.

[0144] Through the above technical solution, the embodiment of the present invention provides an artificial intelligence supervision method and system for the quality of electric power information engineering. The method and system analyze the project engineering data of the electric power information engineering project by constructing it, and then combines the matching of the standard file database to achieve the acquisition of standard data for unmanned supervision operation. Compared with the prior art, the method and system provided by the present invention directly extracts standard data from project engineering data without human intervention, thereby improving the supervision efficiency and accuracy of the electric power information engineering project.

[0145] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0146] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0147] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0148] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0149] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0150] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0151] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules 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, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape 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. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0152] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0153] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. An artificial intelligence supervision method for power information engineering quality, characterized in that: The method comprises: Obtain project engineering data for power information engineering projects; Analyze the project engineering data to obtain project detailed requirements; Determine the engineering quality standards according to the requirements of the project details; Reading standard data from a preset standard file database according to the engineering quality standard; The supervision operation is performed according to the standard data.

2. The method according to claim 1, characterized in that Obtain project engineering data for power information engineering projects, including: Obtaining a construction plan for the electric power information engineering project; Obtaining a procurement plan for the electric power information engineering project; Searching for construction types from the construction plan; Search for material categories from the procurement plan.

3. The method according to claim 2, characterized in that Determine the engineering quality standards according to the project details, including: Determining whether construction requirements matching the construction type are recorded in the construction plan; When it is determined that the construction plan records a construction requirement that matches the construction type, associating the construction requirement with the construction type; If it is determined that the construction plan does not record a construction requirement matching the construction type, setting the construction requirement associated with the construction type to empty; Determine whether material requirements for the material type are recorded in the procurement plan; When it is determined that a material requirement matching the material type is recorded in the procurement plan, associating the material requirement with the material type; When it is determined that the material requirement matching the material type is not recorded in the procurement plan, the material requirement associated with the material type is set to blank.

4. The method according to claim 3, characterized in that: Reading standard data from a preset standard file database according to the engineering quality standard includes: Reading a corresponding standard file in the standard file database according to the construction type; Determine whether the construction requirement corresponding to the construction type is empty; When it is determined that the construction requirement corresponding to the construction type is empty, using the construction standard in the standard file as the standard data; When it is determined that the construction requirement corresponding to the construction type is not empty, selecting the higher one of the construction standard and the construction requirement as the standard data; Reading a corresponding standard file in the standard file database according to the material type; Determine whether the construction requirement corresponding to the material type is empty; In the case where it is determined that the material requirement corresponding to the material type is empty, the material standard in the standard file is used as the standard data; When it is determined that the material requirement corresponding to the material type is not empty, a higher one of the material standard and the material requirement is selected as the standard data.

5. The method according to claim 4, characterized in that Perform supervision operations according to the standard data, including: According to the construction requirements in the standard data, obtain on-site surveillance videos every day; Extracting a sample image from the surveillance video; The sampled image is input into a preset image recognition model to complete the supervision operation of the on-site construction situation.

6. The method according to claim 5, characterized in that Extracting a sample image from the surveillance video includes: Detect whether someone triggers an entry or exit event in the preset image anchor frame; When it is determined that a person triggers an entry or exit event in the image anchor frame, a sampling timestamp is generated; Traversing every two sampling timestamps, and determining whether the time length between the two sampling timestamps is greater than a preset length; In the case where it is determined that the time length between the two sampling timestamps is greater than the preset length, randomly inserting a new sampling timestamp between the two sampling timestamps; Three sampling images are captured at every two sampling timestamps.

7. The method according to claim 6, characterized in that The image recognition model includes: A first input layer, used to receive the first image of the three sampled images; The second input layer is used to receive the second image of the three sampled images; The third input layer is used to receive the third image of the three sampled images; a first convolution module, one end of which is connected to the first input layer and is used to extract features from the first image; a second convolution module, one end of which is connected to the second input layer, and is used for extracting features from the second image; a third convolution module, one end of which is connected to the third input layer, and is used for extracting features from a third image; A first feature combination unit, connected to the feature map outputs of at least two convolution units in the first convolution module, and configured to combine the at least two feature map outputs to obtain a corresponding combined feature; A second feature combination unit, connected to the feature map outputs of at least two convolution units in the second convolution module, and configured to combine the at least two feature map outputs to obtain a corresponding combined feature; A third feature combination unit, connected to the feature map outputs of at least two convolution units in the third convolution module, and used to combine at least two feature map outputs to obtain a corresponding combined feature; A first feature detection unit, connected to the output end of the first feature combination unit, and configured to perform a feature detection operation on the combined feature output by the first feature combination unit; A second feature detection unit, connected to the output end of the second feature combination unit, and used to perform a feature detection operation on the combined feature output by the second feature combination unit; A third feature detection unit is connected to the output end of the third feature combination unit and is used to perform a feature detection operation on the combined feature output by the third feature combination unit; The fusion judgment unit is connected to the first feature detection unit, the second feature detection unit and the third feature detection unit, and is used to perform fusion judgment on the results of the feature detection operation to determine whether there is any violation of standard data in the current monitoring video.

8. The method according to claim 7, characterized in that The first convolution module includes three convolution units connected in series, the second convolution module includes six convolution units connected in series, and the third convolution module includes nine convolution units connected in series, each of which includes: The first residual convolution layer is used to perform the first residual convolution operation on the input image; A second residual convolution layer, used to perform a second residual convolution operation on the result of the first residual convolution operation; The two-dimensional convolution layer is used to perform a two-dimensional convolution operation according to the result of the first residual convolution operation and the result of the second residual convolution operation, and to fuse the result of the two-dimensional convolution operation and the result of the second residual convolution operation for output.

9. The method according to claim 4, characterized in that Perform supervision operations according to the standard data, including: Obtain the supervision log uploaded daily by the supervision engineer; Obtaining the material model and size in the supervision log; The on-site material situation is supervised according to the material model and size and the standard data.

10. An artificial intelligence supervision system for power information engineering quality, characterized in that: The system comprises a processor configured to execute the method according to any one of claims 1 to 9.

Images