Electric power operation behavior determination method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a power operation behavior determination method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring an electric power operation image, and performing portrait extraction on the electric power operation image to obtain an operator image; carrying out human body key point analysis on the operator image, and determining key point coordinate information of the electric power operator in the operator image; drawing a human body skeleton image of the power operator based on the key point coordinate information; and performing behavior feature extraction on the human body skeleton image, and determining the power operation behavior of the power operation personnel. By adopting the method, the accuracy can be improved.

Description

Technical Field

[0001] The present application relates to the field of electric power technology, and in particular to a method, device, computer equipment, computer-readable storage medium and computer program product for determining electric power operation behavior. Background Art

[0002] With the development of power technology, power operation behavior determination technology has emerged. It can ensure the safety of power workers by monitoring their behavior, especially for high-risk activities such as climbing and high-voltage equipment operation, thereby minimizing possible safety accidents during operations.

[0003] In traditional technology, visual recognition technology is generally used to identify people and their behaviors. However, once power workers are in a complex, obscured or long-distance working scene, they will be affected by the complex scene and can only identify the presence or absence of people, but cannot accurately analyze the specific action details of the workers, resulting in low accuracy. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for determining electric power operation behavior that can improve accuracy in order to solve the above technical problems.

[0005] In a first aspect, the present application provides a method for determining an electric power operation behavior, comprising:

[0006] Acquire an electric power operation image, and perform portrait extraction on the electric power operation image to obtain an operator image;

[0007] Performing a human body key point analysis on the operator image to determine key point coordinate information of the power operator in the operator image;

[0008] Drawing a human skeleton image of the power worker based on the key point coordinate information;

[0009] Behavior features are extracted from the human skeleton image to determine the power operation behavior of the power operation personnel.

[0010] In one embodiment, the method further comprises:

[0011] Acquire standard resolution and initial operational imagery for the power industry;

[0012] Based on the standard resolution, the image resolution of the initial operation image is adjusted to obtain an electric power operation image.

[0013] In one embodiment, extracting a human figure from the power operation image to obtain an image of the operator includes:

[0014] Obtain a human body detection model for extracting a human body frame;

[0015] Inputting the power operation image into the human body detection model to obtain human body frame information in the power operation image;

[0016] The power operation image is segmented based on the human body frame information to obtain an operator image.

[0017] In one embodiment, the performing of human body key point analysis on the worker image to determine key point coordinate information of the power worker in the worker image includes:

[0018] Inputting the worker image into a posture analysis model to perform human body key point analysis to obtain a plurality of predicted heat maps; the number of the plurality of predicted heat maps is the same as the number of key points in the worker image;

[0019] For each of the predicted heat maps, the coordinate information of the maximum brightness point in the predicted heat map is used as the coordinate information of the key point corresponding to the predicted heat map.

[0020] In one embodiment, the key point coordinate information is multiple; and drawing the human skeleton image of the power worker based on the key point coordinate information includes:

[0021] Determine the connection relationship corresponding to each key point coordinate information according to the association relationship between the key points corresponding to each key point coordinate information;

[0022] The key points are connected based on the connection relationship to obtain a human skeleton image of the power operator.

[0023] In one embodiment, extracting behavioral features from the human skeleton image to determine the power operation behavior of the power operator includes:

[0024] Obtaining candidate behavior features corresponding to each of the plurality of candidate power operation behaviors;

[0025] Extracting behavioral features from the human skeleton image to determine the behavioral features of the human skeleton image;

[0026] The behavior feature is matched with each of the candidate behavior features, and the target power operation behavior corresponding to the successfully matched target behavior feature is used as the power operation behavior of the power operation personnel.

[0027] In a second aspect, the present application also provides a device for determining electric power operation behavior, comprising:

[0028] A portrait extraction module is used to obtain an electric power operation image and perform portrait extraction on the electric power operation image to obtain an operator image;

[0029] A key point analysis module, used to perform human key point analysis on the operator image to determine key point coordinate information of the power operator in the operator image;

[0030] A skeleton image drawing module, used for drawing a human skeleton image of the power worker based on the key point coordinate information;

[0031] The behavior feature extraction module is used to extract the behavior features of the human skeleton image to determine the power operation behavior of the power operation personnel.

[0032] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above method are implemented.

[0033] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0034] In a fifth aspect, the present application further provides a computer program product, wherein the computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0035] The above-mentioned method, device, computer equipment, computer-readable storage medium and computer program product for determining electric power operation behavior obtain electric power operation images, and perform human figure extraction on the electric power operation images to obtain operator images, and can screen out irrelevant image features in the electric power operation images, so as to facilitate subsequent targeted analysis of the operator images, and then perform human body key point analysis on the operator images to determine the key point coordinate information of the electric power operator in the operator images, and draw the human skeleton image of the electric power operator based on the key point coordinate information, so as to accurately determine the key human information in the operator images, and draw the human skeleton image in a targeted manner, and finally, extract the behavior features of the human skeleton image, so as to determine the electric power operation behavior of the electric power operator. The above-mentioned method is used to analyze the electric power operation behavior, and the electric power operation behavior of the operator can be accurately analyzed without external interference, thereby ensuring the accuracy of the electric power operation behavior analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 An application environment diagram of a power operation behavior method in an embodiment;

[0038] Figure 2 A schematic diagram of a flow chart of a method for electric power operation in one embodiment;

[0039] Figure 3 A schematic diagram of a flow chart of a power operation behavior method in another embodiment;

[0040] Figure 4 A structural block diagram of an electric power operation behavior device in an embodiment;

[0041] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0043] The power operation behavior method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the image acquisition device 104 through the network. Among them, the terminal 102 can be but not limited to various personal computers, laptops, smart phones, tablets, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The image acquisition device 104 refers to a device with image acquisition capabilities, which can obtain digital images of objects and convert them into digital signals that can be processed by computers or related devices. Specifically, in the process of determining the power working behavior, the terminal 102 obtains the power working image from the image acquisition device 104, and performs portrait extraction on the power working image to obtain the image of the worker; performs human body key point analysis on the worker image to determine the key point coordinate information of the power working worker in the worker image; draws the human skeleton image of the power working worker based on the key point coordinate information; performs behavioral feature extraction on the human skeleton image to determine the power working behavior of the power working worker.

[0044] In an exemplary embodiment, Figure 2 As shown, a method for determining electric power operation behavior is provided, and the method is applied to Figure 1 The terminal 102 in the example is used as an example to illustrate, including the following steps S202 to S208. Among them:

[0045] Step S202, acquiring an electric power operation image, and performing portrait extraction on the electric power operation image to obtain an operator image.

[0046] Among them, the power operation image shows the scenes of power workers installing, maintaining, and repairing power facilities in various environments. For example, workers can be seen busy installing equipment in a substation or building a power tower in the sky. The worker image refers to an image that depicts or records workers in a work scene. In this embodiment, the worker image refers to an image that only includes workers.

[0047] Specifically, in the process of determining the electric power operation behavior, in order to avoid the electric power operation behavior determination process being affected by the complex scene when the electric power operation personnel are in a complex, obstructed or long-distance operation scene, the terminal can first obtain the electric power operation image containing the scene where the electric power operation personnel are located, and then perform portrait extraction on the electric power operation image, remove the scene, and obtain the operator image containing only the electric power operation personnel. In some specific embodiments, the corresponding model can be pre-trained, and the portrait is extracted through the model.

[0048] Step S204: perform human body key point analysis on the worker image to determine key point coordinate information of the power worker in the worker image.

[0049] Among them, the key point coordinate information refers to the coordinate information of the key points of the power operator. It can be understood that the key points refer to the key points of the portrait of the power operator. For example, the key points include the arm key points, leg key points, head key points, etc. of the power operator.

[0050] Specifically, after determining the operator image, the terminal can perform a human body key point analysis on the operator image, and then determine the key point coordinate information of the power operator in the operator image, and can determine the behavior position of the power operator at this time based on the key point coordinate information. In some specific embodiments, the terminal can train a corresponding model for key point determination, analyze the operator image based on the model, and finally determine the key point coordinate information of the power operator in the operator image.

[0051] Step S206: drawing a human skeleton image of the power worker based on the key point coordinate information.

[0052] The human skeleton image refers to the human skeletal structure image of the power operator, which can represent the skeleton shape of the power operator.

[0053] Specifically, after determining the key point coordinate information, the terminal can draw the human skeleton image of the power operator based on the key point coordinate information, so as to more clearly determine the behavioral characteristics of the power operator. In some specific embodiments, the terminal can use the model to draw the human skeleton image, or determine the connection relationship corresponding to each key point coordinate information based on the association relationship between the key points corresponding to each key point coordinate information, and connect each key point based on the connection relationship to obtain the human skeleton image of the power operator.

[0054] Step S208, extracting behavioral features from the human skeleton image to determine the power operation behavior of the power operation personnel.

[0055] Among them, power operation behavior refers to the behavior of power operators when performing power operations. For example, power operation behavior can include dangerous behaviors and safe behaviors. Among them, dangerous behaviors include climbing behaviors, crossing behaviors, and drilling behaviors.

[0056] Specifically, after drawing the human skeleton image of the power worker based on the key point coordinate information, the terminal can extract behavioral features of the human skeleton image, determine the power worker's power working behavior, and then determine whether the power working behavior is safe, so as to facilitate subsequent safety warnings based on the power working behavior.

[0057] In some specific embodiments, the terminal can obtain candidate behavior features corresponding to each of the multiple candidate power operation behaviors, extract behavior features from the human skeleton image, determine the behavior features of the human skeleton image, match the behavior features with the candidate behavior features, and use the target power operation behavior corresponding to the successfully matched target behavior features as the power operation behavior of the power operator. In other specific embodiments, the terminal can also train a corresponding neural network model, extract behavior features from the human skeleton image through the neural network model, and determine the power operation behavior of the power operator.

[0058] The above-mentioned method for determining electric power operation behavior obtains an electric power operation image, and performs portrait extraction on the electric power operation image to obtain an image of a worker. The irrelevant image features in the electric power operation image can be screened out to facilitate the subsequent targeted analysis of the worker image. After that, the worker image is subjected to a human key point analysis to determine the key point coordinate information of the electric power operator in the worker image. The human skeleton image of the electric power operator is drawn based on the key point coordinate information. The key human information in the worker image can be accurately determined, and a human skeleton image can be drawn in a targeted manner. Finally, the behavior feature extraction of the human skeleton image can determine the electric power operation behavior of the electric power operator. The above-mentioned method is used to analyze the electric power operation behavior, and the electric power operation behavior of the worker can be accurately analyzed without external interference, thereby ensuring the accuracy of the electric power operation behavior analysis.

[0059] In an exemplary embodiment, the method further includes: acquiring a standard resolution and an initial operation image of the electric power industry; and adjusting the image resolution of the initial operation image based on the standard resolution to obtain the electric power operation image.

[0060] Resolution refers to the number of pixels that a screen or image can display, and is an important parameter for measuring the fineness of an image or screen. Standard resolution refers to a pre-set resolution used for standard calibration. Initial operation images refer to power operation images in their initial state that have not been processed.

[0061] Specifically, in order to improve the accuracy of determining the power operation behavior, the power operation image can be enhanced, that is, the terminal can obtain the standard resolution and the initial operation image of the power industry, and adjust the image resolution of the initial operation image based on the standard resolution, thereby enhancing the initial operation image to obtain the power operation image.

[0062] In this embodiment, by adjusting the resolution of the initial operation image, image enhancement can be achieved, thereby improving the accuracy of subsequent power operation behavior determination.

[0063] In an exemplary embodiment, portrait extraction is performed on an electric power work image to obtain an image of a worker, including: obtaining a human body detection model for extracting a human body frame; inputting the electric power work image into the human body detection model to obtain human body frame information in the electric power work image; and performing image segmentation on the electric power work image based on the human body frame information to obtain an image of a worker.

[0064] The human body detection model refers to a model used to detect the human body frame. In this embodiment, the human body detection model is a YOLOV5 target detection model. The human body frame information refers to information used to describe the overall frame of the power worker. In this embodiment, the human body frame information refers to the upper left corner coordinate information and the lower right corner coordinate information of the human body target frame.

[0065] Specifically, the terminal can first obtain a human body detection model for extracting a human body frame, input the power operation image into the human body detection model, and the output of the human body detection model is the human body frame information in the power operation image. After that, the terminal can perform image segmentation on the power operation image based on the human body frame information to obtain an image of the operator. In some specific embodiments, the terminal inputs the collected power operation image into the YOLOV5 target detection model for human target detection, wherein the YOLOV5 target detection model is divided into four modules, an input end, a backbone network, a feature enhancement module, and a prediction module. The input image is data enhanced at the input end, and feature extraction is performed in the backbone network after adaptive anchor frame calculation and adaptive image scaling. After that, feature enhancement and fusion are performed in the feature enhancement module, and finally prediction is performed in the prediction module through a loss function. The output information of YOLOV5 is the upper left corner coordinates and the lower right corner coordinates of the human body target frame. The image in the area is cropped out according to the coordinate information of the human body target frame to form an image containing only a single target of the power operator.

[0066] In this embodiment, the power operation image can be segmented through the human body detection model, thereby preventing the subsequent behavior analysis process from being affected by complex scenes and improving the accuracy of behavior analysis.

[0067] In an exemplary embodiment, a human body key point analysis is performed on a worker image to determine key point coordinate information of the power worker in the worker image, including: inputting the worker image into a posture analysis model to perform human body key point analysis to obtain multiple predicted heat maps; the number of the multiple predicted heat maps is the same as the number of key points in the worker image; for each predicted heat map, the coordinate information of the maximum brightness point in the predicted heat map is used as the key point coordinate information corresponding to the predicted heat map.

[0068] Among them, the prediction heat map is a graphical tool that indicates the degree of concentration of prediction data through color changes. Darker colors usually indicate larger values, and lighter colors indicate smaller values. The posture analysis model is used to analyze human posture. In this embodiment, the posture analysis model takes the HRNet network model as an example. The number of multiple prediction heat maps is the same as the number of key points in the operator image.

[0069] Specifically, after determining the image of the operator, the terminal can input the image of the operator into the posture analysis model, and the posture analysis model performs human key point analysis on the image of the operator to obtain multiple predicted heat maps. After that, for each predicted heat map, the coordinate information of the maximum brightness point in the predicted heat map is used as the key point coordinate information corresponding to the predicted heat map. In a specific embodiment, the terminal can input the image of the operator into the HRNet network model for human key point prediction. HRNet is a key point detection network proposed for two-dimensional human posture estimation tasks. The image is first processed through some initial convolution layers in the HRNet network. After the initial layer, there is a multi-resolution subnet, in which the highest resolution path maintains the original resolution of the image. Then, features of multiple resolutions are processed in parallel. The network starts with a single high-resolution stream and gradually adds lower and lower resolution streams to form a parallel multi-resolution structure. Streams of different resolutions exchange information through cross-scale connections, so that each resolution stream can obtain useful context information from other streams. The HRNet network ensures that features between different resolutions can be frequently fused by inserting convolutions in each resolution stream at each stage and then merging the results back to other resolutions. The final stage of the network goes through a series of fusion operations to ensure that the final output retains high-resolution features, which are used to generate the final output, which is the coordinate information of the key points of the human body.

[0070] In this embodiment, the operator image is input into the HRNet network model for key point analysis, which can improve the accuracy of the key point analysis.

[0071] In an exemplary embodiment, there are multiple key point coordinate information; drawing the human skeleton image of the power worker based on the key point coordinate information includes: determining the connection relationship corresponding to each key point coordinate information according to the association relationship between the key points corresponding to each key point coordinate information; connecting each key point based on the connection relationship to obtain the human skeleton image of the power worker.

[0072] Specifically, after determining the coordinate information of each key point, in order to draw the human skeleton image of the power operator, the terminal can first determine the association relationship between the key points corresponding to each key point coordinate information, and then determine the connection relationship corresponding to each key point coordinate information based on the association relationship between the key points, and connect each key point based on the connection relationship to obtain the human skeleton image of the power operator. For example, after the terminal determines the key point coordinate information of the key points "leg", "head", and "arm", it can first determine the association relationship between the key points "leg", "head", and "arm", and then determine the connection relationship corresponding to each key point coordinate information based on the association relationship between the key points, and connect each key point based on the connection relationship to obtain the human skeleton image of the power operator.

[0073] In this embodiment, based on the association relationship between the key points corresponding to the coordinate information of each key point, the connection relationship corresponding to the coordinate information of each key point is determined, so as to ensure the accuracy of drawing the human skeleton image.

[0074] In an exemplary embodiment, behavioral features are extracted from a human skeleton image to determine the power working behavior of a power worker, including: obtaining candidate behavioral features corresponding to each of a plurality of candidate power working behaviors; extracting behavioral features from a human skeleton image to determine the behavioral features of the human skeleton image; matching the behavioral features with each candidate behavioral feature, and using the target power working behavior corresponding to the successfully matched target behavioral feature as the power working behavior of the power worker.

[0075] The candidate power operation behavior refers to the power operation behavior used as an alternative, and the candidate behavior feature refers to the behavior feature used as an alternative. The behavior feature can reflect the characteristics of the power operation behavior.

[0076] Specifically, after determining the human skeleton image, the terminal can first obtain multiple candidate power operation behaviors, and determine the candidate behavior features corresponding to each of the multiple candidate power operation behaviors, and then extract the behavior features of the human skeleton image, determine the behavior features of the human skeleton image, match the behavior features with the candidate behavior features, and use the target power operation behavior corresponding to the successfully matched target behavior features as the power operation behavior of the power operation personnel.

[0077] In a specific embodiment, a support vector machine algorithm can be used to perform regression classification on human skeleton images. Support vector machine is a powerful supervised learning algorithm for classification and regression tasks. Its basic principle is to find a hyperplane (a line in two-dimensional space) that can best separate different categories. In the classification task, the goal of the support vector machine is to find a hyperplane that maximizes the interval (also called margin) between data points of different categories. This maximized interval provides a certain degree of fault tolerance for future data point classification. The support vector machine uses kernel technology to process the relationship between nonlinear features and finds a hyperplane that can effectively separate data by mapping the original data to a higher-dimensional space. The support vector machine classifies power workers, including the three most likely dangerous behaviors of power workers, climbing behavior, crossing behavior, and bottom drilling behavior, and normal behavior that is not the above three dangerous behaviors, a total of four types of behavior results.

[0078] In this embodiment, the human skeleton image is regressed and classified by using a support vector machine algorithm, so as to ensure the accuracy of determining the power operation behavior.

[0079] In a specific embodiment, a method for determining electric power operation behavior in an actual application scenario is also provided, including:

[0080] Step S1: Acquire power operation images through monitoring cameras.

[0081] Step S2: The terminal inputs the collected power operation image into the YOLOV5 target detection model for human target detection, wherein the YOLOV5 target detection model is divided into four modules, input end, backbone network, feature enhancement module and prediction module. The input image is subjected to data enhancement at the input end, feature extraction is performed in the backbone network after adaptive anchor frame calculation and adaptive image scaling, and then feature enhancement and fusion are performed in the feature enhancement module, and finally prediction is performed in the prediction module through the loss function. The output information of YOLOV5 is the coordinates of the upper left corner and the lower right corner of the human target frame. The image in the area is cropped out according to the coordinate information of the human target frame to form an image containing only the single target of the power operator.

[0082] Step S3: The terminal can input the worker image into the HRNet network model for human key point prediction. HRNet is a key point detection network proposed for the task of 2D human pose estimation. The image is first processed through some initial convolutional layers in the HRNet network, followed by a multi-resolution subnet, where the highest resolution path maintains the original resolution of the image. Then, features of multiple resolutions are processed in parallel. The network starts with a single high-resolution stream and gradually adds lower and lower resolution streams to form a parallel multi-resolution structure. Streams of different resolutions exchange information through cross-scale connections, so that each resolution stream can obtain useful contextual information from other streams. The HRNet network ensures that features between different resolutions can be frequently fused by inserting convolutions in each resolution stream at each stage and then merging the results back to other resolutions. The final stage of the network ensures that the final output retains high-resolution features through a series of fusion operations, which are used to generate the final output, that is, the coordinate information of the human key points.

[0083] Step S4: Draw a human skeleton image using the coordinate information of key points of the human body.

[0084] Step S5: The support vector machine algorithm can be used to perform regression classification on the human skeleton image. Support vector machine is a powerful supervised learning algorithm for classification and regression tasks. Its basic principle is to find a hyperplane (a line in two-dimensional space) that can best separate different categories. In the classification task, the goal of the support vector machine is to find a hyperplane that maximizes the interval (also called margin) between data points of different categories. This maximized interval provides a certain degree of fault tolerance for future data point classification. The support vector machine uses kernel technology to process the relationship between nonlinear features and finds a hyperplane that can effectively separate the data by mapping the original data to a higher-dimensional space. The support vector machine classifies power workers, including the three most likely dangerous behaviors of power workers, climbing behavior, crossing behavior, and bottom drilling behavior, and normal behavior that is not one of the above three dangerous behaviors, a total of four types of behavior results.

[0085] Step S6: Output the judgment result of the behavior of the power operator according to the classification result.

[0086] Through the above process, the behavior information of power workers can be obtained more accurately, providing a reliable technical solution to prevent power workers from causing serious consequences due to their own dangerous behaviors.

[0087] In a specific embodiment, Figure 3 As shown, a method for determining electric power operation behavior is also provided, comprising:

[0088] Step S301, obtaining a standard resolution and an initial operation image of the power industry;

[0089] Step S302, adjusting the image resolution of the initial operation image based on the standard resolution to obtain the power operation image;

[0090] Step S303, obtaining an electric power operation image and a human body detection model for extracting a human body frame, inputting the electric power operation image into the human body detection model, and obtaining human body frame information in the electric power operation image;

[0091] Step S304, performing image segmentation on the power operation image based on the human frame information to obtain an image of the operator;

[0092] Step S305, inputting the worker image into a posture analysis model to perform human body key point analysis to obtain a plurality of predicted heat maps;

[0093] The number of the multiple predicted heat maps is the same as the number of key points in the worker image;

[0094] Step S306, for each predicted heat map, using the coordinate information of the maximum brightness point in the predicted heat map as the coordinate information of the key point corresponding to the predicted heat map;

[0095] Among them, there are multiple key point coordinate information;

[0096] Step S307, determining the connection relationship corresponding to each key point coordinate information according to the association relationship between the key points corresponding to each key point coordinate information;

[0097] Step S308, connecting the key points based on the connection relationship to obtain a human skeleton image of the power worker;

[0098] Step S309, obtaining candidate behavior features corresponding to each of the plurality of candidate power operation behaviors, extracting behavior features from the human skeleton image, and determining the behavior features of the human skeleton image;

[0099] Step S310, matching the behavior feature with each candidate behavior feature, and taking the target power operation behavior corresponding to the successfully matched target behavior feature as the power operation behavior of the power operation personnel.

[0100] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0101] Based on the same inventive concept, the embodiment of the present application also provides a device for determining electric power operation behavior for implementing the method for determining electric power operation behavior involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the method above, so the specific limitations in one or more embodiments of the device for determining electric power operation behavior provided below can refer to the limitations of the method for determining electric power operation behavior above, and will not be repeated here.

[0102] In an exemplary embodiment, Figure 4 As shown, a device 400 for determining electric power operation behavior is provided, comprising: a portrait extraction module 402, a key point analysis module 404, a skeleton image drawing module 406 and a behavior feature extraction module 408, wherein:

[0103] The portrait extraction module 402 is used to obtain the power operation image and perform portrait extraction on the power operation image to obtain the operator image;

[0104] The key point analysis module 404 is used to perform human key point analysis on the operator image to determine key point coordinate information of the power operator in the operator image;

[0105] The skeleton image drawing module 406 is used to draw the human skeleton image of the power worker based on the key point coordinate information;

[0106] The behavior feature extraction module 408 is used to extract behavior features from the human skeleton image to determine the power operation behavior of the power operation personnel.

[0107] In an exemplary embodiment, the power operation behavior determination device 400 further includes a power operation image determination module, which is specifically configured to:

[0108] Acquire standard resolution and initial operational imagery for the power industry;

[0109] Based on the standard resolution, the image resolution of the initial operation image is adjusted to obtain the power operation image.

[0110] In an exemplary embodiment, the portrait extraction module 402 is specifically used to: obtain a human body detection model for extracting a human body frame;

[0111] Input the power operation image into the human body detection model to obtain the human body frame information in the power operation image;

[0112] The power operation image is segmented based on the human body frame information to obtain the operator image.

[0113] In an exemplary embodiment, the key point analysis module 404 is specifically used to: input the worker image into the posture analysis model to perform human key point analysis to obtain a plurality of predicted heat maps; the number of the plurality of predicted heat maps is the same as the number of key points in the worker image;

[0114] For each predicted heat map, the coordinate information of the maximum brightness point in the predicted heat map is used as the coordinate information of the key point corresponding to the predicted heat map.

[0115] In an exemplary embodiment, there are multiple key point coordinate information. In this embodiment, the skeleton image drawing module 406 is specifically used to: determine the connection relationship corresponding to each key point coordinate information according to the association relationship between the key points corresponding to each key point coordinate information;

[0116] The key points are connected based on the connection relationship to obtain the human skeleton image of the power worker.

[0117] In an exemplary embodiment, the behavior feature extraction module 408 is specifically used to: obtain candidate behavior features corresponding to each of the plurality of candidate power operation behaviors;

[0118] Extracting behavioral features from the human skeleton image to determine the behavioral features of the human skeleton image;

[0119] The behavior feature is matched with each candidate behavior feature, and the target power operation behavior corresponding to the successfully matched target behavior feature is used as the power operation behavior of the power operation personnel.

[0120] Each module in the above-mentioned power operation behavior determination device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0121] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a method for determining an electric power operation behavior is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0122] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0123] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps of the above method when executing the computer program.

[0124] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0125] In one embodiment, a computer program product is provided, comprising a computer program, which implements the steps of the above method when executed by a processor.

[0126] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0127] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0128] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0129] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for determining electric power operation behavior, characterized in that: The method comprises: Acquire an electric power operation image, and perform portrait extraction on the electric power operation image to obtain an operator image; Performing a human body key point analysis on the operator image to determine key point coordinate information of the power operator in the operator image; Drawing a human skeleton image of the power worker based on the key point coordinate information; Behavior features are extracted from the human skeleton image to determine the power operation behavior of the power operation personnel.

2. The method according to claim 1, characterized in that The method further comprises: Acquire standard resolution and initial operational imagery for the power industry; Based on the standard resolution, the image resolution of the initial operation image is adjusted to obtain an electric power operation image.

3. The method according to claim 2, characterized in that The extracting the portrait from the power operation image to obtain the operator image includes: Obtain a human body detection model for extracting a human body frame; Inputting the power operation image into the human body detection model to obtain human body frame information in the power operation image; The power operation image is segmented based on the human body frame information to obtain an operator image.

4. The method according to claim 1, characterized in that The performing of human body key point analysis on the operator image to determine key point coordinate information of the electric power operator in the operator image includes: Inputting the worker image into a posture analysis model to perform human body key point analysis to obtain a plurality of predicted heat maps; the number of the plurality of predicted heat maps is the same as the number of key points in the worker image; For each of the predicted heat maps, the coordinate information of the maximum brightness point in the predicted heat map is used as the coordinate information of the key point corresponding to the predicted heat map.

5. The method according to claim 1, characterized in that The key point coordinate information is multiple; the human skeleton image of the power worker is drawn based on the key point coordinate information, including: Determine the connection relationship corresponding to each key point coordinate information according to the association relationship between the key points corresponding to each key point coordinate information; The key points are connected based on the connection relationship to obtain a human skeleton image of the power operator.

6. The method according to claim 1, characterized in that The extracting behavior features from the human skeleton image to determine the power operation behavior of the power operation personnel includes: Obtaining candidate behavior features corresponding to each of the plurality of candidate power operation behaviors; Extracting behavioral features from the human skeleton image to determine the behavioral features of the human skeleton image; The behavior feature is matched with each of the candidate behavior features, and the target power operation behavior corresponding to the successfully matched target behavior feature is used as the power operation behavior of the power operation personnel.

7. A device for determining electric power operation behavior, characterized in that: The device comprises: A portrait extraction module is used to obtain an electric power operation image and perform portrait extraction on the electric power operation image to obtain an operator image; A key point analysis module, used to perform human key point analysis on the operator image to determine key point coordinate information of the power operator in the operator image; A skeleton image drawing module, used for drawing a human skeleton image of the power worker based on the key point coordinate information; The behavior feature extraction module is used to extract the behavior features of the human skeleton image to determine the power operation behavior of the power operation personnel.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images