A Visualized Power Grid Fault Identification and Location Method and Device Based on BIM Modeling

Abstract

This invention relates to the field of power grid line monitoring technology, and provides a method and device for visual power grid fault identification and location based on BIM modeling. The method involves conducting on-site data collection on the power grid system and drawing a power grid engineering system diagram based on the data. Simultaneously, the power grid engineering system diagram and a monitoring network distribution map are loaded. For monitoring points located near power equipment, the association between the corresponding power equipment and one or more monitoring points is established. The power grid engineering system diagram, along with the video frame relationships of the one or more monitoring points associated with each power equipment obtained from the analysis, are imported into the BIM to generate a 3D model. This invention avoids manual inspection, promptly detects power grid faults, improves inspection efficiency, and saves labor costs.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid line monitoring, in particular to a visual power grid fault identification and positioning method and device based on BIM modeling. BACKGROUND

[0002] Although the research and planning of power grids in different countries are different, the fundamental requirements for power grids are consistent, that is, power grids should be "stronger and smarter". Strongness is the requirement for the safety of power grids, that is, the power grid can still maintain the power supply capacity to users when a large disturbance and fault occur in the power grid, and large-area power outage accidents do not occur; the safety of the power grid can be ensured under natural disasters, extreme weather conditions or external damage; and the power information security can be ensured. However, the current fault analysis of the power grid generally needs to be judged by manpower according to experience, which is not only low in efficiency, but also low in accuracy.

[0003] Therefore, it is urgent to overcome the defects of the prior art in the technical field. SUMMARY

[0004] The purpose of the present application is to provide a visual power grid fault identification and positioning system and method based on BIM modeling, which draws CAD drawings by actually measuring the power grid engineering, generates a power grid engineering GIM digital model by using the drawings, simultaneously monitors and processes video images of the power grid, finds out the fault points of the power grid by comparing the similarity of the images, and displays the fault point positions on the power grid engineering GIM digital model, thereby solving the problems of untimely discovery of power grid faults and high cost of manpower inspection.

[0005] The present application adopts the following technical solutions:

[0006] In a first aspect, the present application provides a visual power grid fault identification and positioning method based on BIM modeling, which comprises:

[0007] The power grid system is actually collected, and a power grid engineering system diagram is drawn according to the actual collection; wherein the power grid engineering system diagram contains one or more of power equipment, equipment coordinates, equipment connection relationships, and equipment information;

[0008] The power grid engineering system diagram and a monitoring network distribution diagram are simultaneously loaded, and for the corresponding monitoring points distributed near the power equipment, an association relationship between the corresponding power equipment and one or more monitoring points is established;

[0009] The power equipment recognition in the images collected by different monitoring points is identified to complete the video frame relationship of the one or more monitoring points under the optimal shooting angle combination of the power equipment; wherein, the optimal shooting angle refers to the video frames collected by the corresponding one or more monitoring points at the same time or asynchronous time, which can present the external and / or internal information of the power equipment to the greatest extent after completing the picture splicing of the video frames;

[0010] The power grid engineering system diagram and the video frame relationship of the one or more monitoring points associated with each power equipment obtained by corresponding analysis are imported into the BIM to generate a three-dimensional model; wherein, the BIM constructs a three-dimensional model of the power grid engineering system according to the power grid engineering system diagram and the video frames, and periodically acquires the corresponding video frame content to update the visual state of the power equipment in the three-dimensional model according to one or more information of the shooting angle of the monitoring point camera of the corresponding video frame associated monitoring point, the time difference between the video frames of different monitoring points, and the jurisdiction subject to which the different monitoring points belong.

[0011] Preferably, when the one or more monitoring points belong to different jurisdiction subjects, the completion of the video frame relationship of the one or more monitoring points under the optimal shooting angle combination of the power equipment specifically includes:

[0012] The video content in each monitoring period of the one or more monitoring points is acquired from the different jurisdiction subjects to which the one or more monitoring points belong;

[0013] From the one or more monitoring points, a main monitoring point is first selected, and the others are used as auxiliary monitoring points, wherein the main monitoring point has one or more of the largest amount of information, the longest subject identification time in a period, the completeness and timeliness of the acquired video frames are guaranteed;

[0014] The video frame with the largest amount of information from the main monitoring point is acquired as a base frame, and under the condition of the base frame, the video frames in each period of the one or more monitoring points are traversed to select a group of video frames that can best restore the visual information of the power equipment from the base frame, and record one or more information of the monitoring point camera shooting angle of the corresponding group of video frames in each monitoring point, the time point of the video frame of each monitoring point in the period, and the jurisdiction subject to which the different monitoring points belong; so as to be used for subsequent periodic updating of the visual state of the power equipment in the three-dimensional model.

[0015] Preferably, after confirming one or more information of the monitoring point camera shooting angle of each monitoring point, the time difference between the video frames of different monitoring points, and the jurisdiction subject to which the different monitoring points belong, the method further includes:

[0016] Send a message containing the camera shooting angle and corresponding video frame of the monitoring point within the period to the respective jurisdiction of the monitoring point;

[0017] The respective governing bodies of the monitoring points will send a preset number of video frames collected under the corresponding conditions to the 3D modeling system when the video streams collected at their respective monitoring points meet the message conditions that the camera shooting angle and the corresponding video frames are located at time points within the cycle.

[0018] Preferably, during the previous round of image stitching, the outline information and key points of the power equipment are extracted as the basis for the stitching process. After updating the power equipment in the 3D model for one round, the outline information and key point information generated for each video frame are recorded separately. In the next update cycle, when a new round of video frames is obtained for stitching, after the outline information and key point information are generated, they are matched with the stored outline information and key point information corresponding to the previous round of video frames. If the corresponding deviation exceeds a preset threshold, an adjustment message for the camera shooting angle of the monitoring point and the time point of the corresponding video frame within the cycle is sent to the respective jurisdiction of the monitoring point.

[0019] Preferably, the power grid engineering system diagram is a CAD two-dimensional measured drawing, then the method includes:

[0020] The actual situation of power grid projects is simulated by conducting field measurements on power grid projects from multiple locations, angles, and directions.

[0021] Record the electrical equipment required for the power grid project and draw CAD two-dimensional measured drawings of the power grid project;

[0022] Import the CAD 2D measured sketch into the BIM 3D modeling tool, and obtain the BIM 3D model based on the control point coordinates.

[0023] Preferably, the step of completing the image stitching under the video frame specifically includes:

[0024] The original power grid image is segmented using an image segmentation algorithm to generate different initialized regions of the image. After initialization, the default similarity of all images is Y=100%;

[0025] The original power grid image data of different initialization areas were processed sequentially by mean, sample variance and standardization;

[0026] Feature extraction of power grid images is performed using an image extreme point detection method, and local features of the power grid images are represented by scale-invariant feature transformation descriptors. This method is used to detect key points in power grid images, accurately locate key points of power grid faults, assign principal directions to key points, and calculate scale-invariant feature transformation descriptors.

[0027] Based on the feature extraction results of the feature fusion algorithm, images of the same power grid are aggregated while images of different power grids are separated.

[0028] Select adjacent regions of the power grid image and match two adjacent regions. And calculate the similarity of the feature fusion results of two adjacent regions. , ;

[0029] like ,make Merge two adjacent regions Remove about and similarity and The iterative calculation continues until Y=0, thus determining the splicing point.

[0030] Preferably, the step of segmenting the original power grid image using an image segmentation algorithm to generate different initialized regions of the image specifically includes:

[0031] The original overhead power line image is modeled using a Gaussian Markov random field model. The model expression is as follows:

[0032] ;

[0033] In the formula, B represents the original power grid image, d represents the location of a portion of the overhead lines in the original power grid image, and γ represents the adjacent location, the location of the fault point. The probability is expressed as C represents a custom plane;

[0034] The original power grid image B was simulated using a Gaussian Markov random field model. The expression is:

[0035] ;

[0036] In the formula, Let represent the model parameters of the Gaussian Markov random field, d represent the location image of any point in the power grid, and a represent the size of the image. Represents discrete regions in a power grid image. Zero-mean Gaussian noise is represented by the energy function of the energy function of the dissected normal level set, which represents the interaction between pixels in the texture of the original power grid image.

[0037] ;

[0038] In the formula, These represent the average gray levels of the foreground and background regions of the original power grid image, respectively. Represents the entire set of regions in the image. This represents the mean of the energy function. Let n represent the integral over the power grid image, where n represents a single image region. Represents the observed image;

[0039] The expression for the average gray level of the foreground region is:

[0040] ;

[0041] In the formula, G represents the Heaviside function. Represents the observed image;

[0042] The expression for the average grayscale value of the background region is:

[0043] ;

[0044] In the formula, Represent the observed image and calculate the iteration. Continue until convergence, then output the segmented image to complete image segmentation.

[0045] Preferably, the step of performing mean, sample variance, and standardization processing on the raw power grid image data of different initialization regions specifically includes:

[0046] The mean expression is:

[0047] ;

[0048] In the formula, This represents the mean value of the overhead power line image, where x represents any data point in the overhead power line image. denoted by , g represents the minimum batch processing value of the overhead line image, h represents the number of data points in the overhead line image, and h represents the sum of the gray values ​​of all data points in the overhead line image.

[0049] The expression for the sample variance is:

[0050] ;

[0051] In the formula, This represents the sample variance value of the power grid image;

[0052] The standardized processing expression is:

[0053] ;

[0054] In the formula, This represents the standardized value of the power grid image.

[0055] Preferably, the computational scale-invariant feature transformation descriptor specifically includes:

[0056] Rotate the X-axis of the power grid image to the principal direction of the scale-invariant feature transformation keypoint, centering on each keypoint;

[0057] Divide the region into sub-regions centered on each key point, calculate the Gaussian weighted gradient direction angle histogram in each sub-region, and plot the cumulative value of each gradient direction to form a seed point.

[0058] The expression for establishing the similarity function is:

[0059] ;

[0060] In the formula, F and G represent two adjacent power grid images, and E represents the set of all power grid images fused by edge features. The similarity coefficient of image x is represented. This represents the set of all feature points in an image.

[0061] Secondly, the present invention also provides a BIM-based visual power grid fault identification and location device for implementing the BIM-based visual power grid fault identification and location method described in the first aspect, the device comprising:

[0062] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the processor for performing the BIM-based visualized power grid fault identification and location method described in the first aspect.

[0063] Thirdly, the present invention also provides a non-volatile computer storage medium storing computer-executable instructions, which are executed by one or more processors to perform the BIM-based visualized power grid fault identification and location method described in the first aspect.

[0064] This invention acquires real-time video images of the power grid, uses an image extreme point detection method to extract features from the power grid images, quickly identifies and extracts image features, and uses a feature fusion algorithm to separate the images. Finally, it judges the similarity of adjacent image areas to quickly find the fault point of the overhead line, avoiding manual inspection, timely detection of power grid fault points, improving inspection efficiency, and saving labor costs.

[0065] This invention involves conducting on-site measurements and drawing CAD drawings of power grid projects, importing the CAD drawings into BIM 3D modeling tools to create BIM 3D models, using BIM 3D geometric image calculation tools to generate GIM digital models of power grid projects, and establishing real-time monitoring, early warning, and evaluation functions for the status and operating environment of power grid equipment, forming an information sharing platform, thereby improving the intelligence level and monitoring capabilities of power grid projects.

[0066] This invention displays the discovered power grid fault points on the GIM digital model of the power grid project, intuitively showing the location of the fault, enabling administrators to promptly assign maintenance personnel to repair it, improving maintenance efficiency and minimizing losses. Attached Figure Description

[0067] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0068] Figure 1 This is a schematic diagram of a visual power grid fault identification and location method based on BIM modeling provided in an embodiment of the present invention;

[0069] Figure 2 This is a schematic diagram of a visual power grid fault identification and location method based on BIM modeling provided in an embodiment of the present invention;

[0070] Figure 3 This is a schematic diagram of a visual power grid fault identification and location method based on BIM modeling provided in an embodiment of the present invention;

[0071] Figure 4 A structural block diagram of a BIM-based visualized power grid fault identification and location system provided in an embodiment of the present invention;

[0072] Figure 5 A flowchart of a GIM digital model for power grid engineering provided in an embodiment of the present invention;

[0073] Figure 6 This is a flowchart of a power grid fault identification and location method provided in an embodiment of the present invention;

[0074] Figure 7 A flowchart for generating a GIM digital model of a power grid project is provided in this embodiment of the invention.

[0075] Figure 8 A flowchart of a computational scale-invariant feature transformation descriptor provided in an embodiment of the present invention;

[0076] Figure 9 This is a schematic diagram of a BIM-based visualized power grid fault identification and location device provided in an embodiment of the present invention. Detailed Implementation

[0077] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0078] In the description of this invention, the terms "inner", "outer", "longitudinal", "lateral", "upper", "lower", "top", "bottom", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and do not require that this invention must be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0079] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0080] Example 1:

[0081] Embodiment 1 of the present invention provides a visual power grid fault identification and location method based on BIM modeling, such as... Figure 1 As shown, the method includes:

[0082] In step 201, the power grid system is collected on-site, and a power grid engineering system diagram is drawn based on the on-site data.

[0083] The power grid engineering system diagram includes one or more of the following: power equipment, equipment coordinates, equipment connection relationships, and equipment information.

[0084] In step 202, the power grid engineering system diagram and the monitoring network distribution diagram are loaded simultaneously. For the corresponding monitoring points distributed near the power equipment, the association relationship between the corresponding power equipment and one or more monitoring points is established.

[0085] In step 203, by identifying the power equipment in the images collected from different monitoring points, the video frame relationship under the optimal shooting angle combination of one or more monitoring points on the power equipment is completed.

[0086] The optimal shooting angle refers to the video frames collected simultaneously or asynchronously from one or more monitoring points, which, after image stitching of the video frames, can present the external and / or internal information of the power equipment to the greatest extent.

[0087] In step 204, the power grid engineering system diagram and the video frame relationships of one or more monitoring points associated with each power equipment obtained from the corresponding analysis are imported into the BIM to generate a 3D model.

[0088] In this process, BIM constructs a three-dimensional model of the power grid engineering system based on the power grid engineering system diagram and video frames. It also periodically acquires the content of the corresponding video frames based on one or more of the following information: the shooting angle of the camera at the monitoring point associated with the corresponding video frame, the time difference between video frames at different monitoring points, and the jurisdiction of the different monitoring points. This information is used to update the visual state of the power equipment in the three-dimensional model.

[0089] This invention acquires real-time video images of the power grid, uses an image extreme point detection method to extract features from the power grid images, quickly identifies and extracts image features, and uses a feature fusion algorithm to separate the images. Finally, it judges the similarity of adjacent image areas to quickly find the fault point of the overhead line, avoiding manual inspection, timely detection of power grid fault points, improving inspection efficiency, and saving labor costs.

[0090] In this embodiment of the invention, when one or more monitoring points belong to different governing bodies, the video frame relationship of the one or more monitoring points under the optimal shooting angle combination on the power equipment is as follows: Figure 2 As shown, it specifically includes:

[0091] In step 301, video content within a monitoring cycle is obtained from the different governing bodies to which the one or more monitoring points belong.

[0092] In step 302, from the one or more monitoring points, one primary monitoring point is selected first, and the others are used as secondary monitoring points.

[0093] Among them, the main monitoring point has one or more of the following characteristics: the shooting angle with the most information, the longest duration of main equipment identification within the cycle, and the integrity and timeliness of the acquired video frames are guaranteed.

[0094] In step 303, the video frame with the most information is obtained from the main monitoring point as the base frame. With the base frame set, the video frames within their respective cycles are traversed from one or more other monitoring points. The best set of video frames that can be stitched together with the base frame to restore the visual information of the power equipment is selected. One or more pieces of information are recorded in the corresponding set of video frames, including the shooting angle of the monitoring point camera at each monitoring point, the time point of each monitoring point video frame within the cycle, and the jurisdiction of different monitoring points. This information is used for subsequent periodic updates of the visual status of the power equipment in the 3D model.

[0095] In this embodiment of the invention, after confirming one or more of the following information: the shooting angle of the camera at each monitoring point, the time difference between video frames at different monitoring points, and the jurisdiction to which different monitoring points belong, such as Figure 3 As shown, the method also includes:

[0096] In step 401, a message carrying the camera shooting angle and corresponding video frame of the monitoring point within the period is sent to the respective jurisdiction of the monitoring point.

[0097] In step 402, the respective monitoring point's governing body will send a preset number of video frames collected under the corresponding conditions to the 3D modeling system when the video stream collected by each monitoring point meets the message conditions that the camera's shooting angle and the corresponding video frame are located within the time period.

[0098] In the specific implementation process, it is also involved that during the previous round of image stitching, the outline information and key points of the power equipment are extracted as the basis for the stitching process. After the power equipment in the 3D model is updated, the outline information and key point information generated for each video frame are recorded separately. So, in the next update cycle, when a new round of video frames is obtained for stitching, after the outline information and key point information are generated again, they are matched with the stored outline information and key point information corresponding to the previous round of video frames. So, when the corresponding deviation exceeds the preset threshold, an adjustment message for the shooting angle of the camera at the monitoring point and the time point of the corresponding video frame within the cycle is sent to the respective jurisdiction of the monitoring point.

[0099] Another innovative aspect of this invention is that it provides an effective system for centrally utilizing monitoring points belonging to different governing bodies. These governing bodies can be understood as monitoring points in specific residential areas, schools, public institutions, etc. Furthermore, this invention creatively proposes that monitoring points under different governing bodies only collect the optimal number of frames. This reduces the amount of data transmitted over the network between the system and the corresponding governing body, saving bandwidth. It also improves the data reading efficiency and the corresponding video frame stitching processing efficiency. Importantly, this invention also reuses the image processing results during the stitching process, effectively correcting and adjusting for potential delays caused by the camera pan-tilt control. This further ensures that, in actual processing, after statistically analyzing the frequency of adjustments needed, the stitching process can use the outline and key points from the previous round as initial matching points, thereby greatly improving stitching efficiency.

[0100] Example 2:

[0101] See Figure 4 As shown, this invention is a BIM-based visual power grid fault identification and location system, which can be used to execute the methods described in Embodiment 1 and subsequent Embodiment 3 of this invention. It includes a model creation unit and a power grid fault identification and location unit.

[0102] The model creation unit includes a power grid field measurement module, a power grid CAD drawing module, a BIM 3D modeling module, a GIM digital model module, and a human-machine interface visualization module. The power grid field measurement module is used to conduct field measurements on power grid projects to obtain their actual conditions. The power grid CAD drawing module is used to record the electrical equipment of the power grid project and draw CAD 2D measured drawings. The BIM 3D modeling module is used to import the CAD 2D measured drawings into BIM 3D modeling tools to construct a BIM 3D model. The GIM digital model module is used to generate a GIM digital model of the power grid project based on the BIM 3D model. The human-machine interface visualization module is used to display the GIM digital model of the power grid project and show the fault locations in real time.

[0103] The power grid fault identification and location unit includes a power grid video acquisition module, an image processing module, a feature extraction module, a similarity judgment module, and a fault point location module. The power grid video acquisition module is used to acquire power grid video images in real time. The image processing module is used to process the acquired video images. The feature extraction module is used to extract features from the processed images according to the feature fusion algorithm. The similarity judgment module is used to select power grid images of adjacent areas for matching and calculating the similarity of the feature fusion results of two adjacent areas. The fault point location module is used to calculate the fault points of the power grid in all areas.

[0104] The establishment of a power grid project includes the feasibility study stage, the preliminary design stage, the construction drawing stage, and the as-built drawing stage.

[0105] Before importing CAD 2D measured drawings into BIM 3D modeling tools, the sectional views of the measured drawings need to be processed into layers to generate complete CAD measured power grid equipment drawings; the connection positions of overhead lines are determined according to the as-built drawings of the power grid equipment, and the overhead lines are connected with the same number for the same row to complete the connection processing of the conductors; and the electrical equipment is saved one by one according to the number, and the base point of the model is determined according to the geographical coordinates.

[0106] The connection location of overhead lines is determined by generating the planar projection of the conductors using a three-dimensional geometric algorithm. The positioning points between each device are then determined based on the planar projection lines between electrical equipment. Finally, a geometric intersection algorithm is used to determine the interface between the devices to which each conductor belongs.

[0107] The image processing module performs mean, sample variance, and standardization processing sequentially; the feature extraction module performs key point detection of power grid images, precise location of key points of power grid faults, key point principal orientation assignment, and calculation of scale-invariant feature transformation descriptors sequentially.

[0108] Example 3:

[0109] This invention provides a visual power grid fault identification and location method based on BIM modeling. The corresponding method is the embodiment of the technical solution in Embodiment 1 combined with a specific platform, tools, and algorithm. Furthermore, this invention describes the method from the perspective of fundamental theoretical principles; the optimal correlation scheme in Embodiment 1 is not elaborated upon in detail here. Figures 5-6 As shown, it includes the following steps:

[0110] Step S1: Conduct on-site measurements of the power grid project, using multi-location, multi-angle, and multi-directional methods to simulate the actual situation of the power grid project.

[0111] Step S2: Record the electrical equipment required for the power grid project and draw the CAD two-dimensional measured drawings of the power grid project.

[0112] Step S3: Import the CAD 2D measured sketch into the BIM 3D modeling tool, and obtain the BIM 3D model based on the control point coordinates.

[0113] Step S4: Using BIM visualization programming tools, based on BIM 3D graphics tools, calculate and generate overhead line curves according to 3D geometric graphics algorithms.

[0114] Step S5: Create a GIM digital model of the power grid project. Using BIM 3D geometric image calculation tools and BIM visualization programming tools, a GIM digital model of the power grid project is generated based on the GIM 3D graphics tools.

[0115] Step S6: The video surveillance equipment collects image data of the power grid in real time.

[0116] Step S7: Segment the original power grid image using an image segmentation algorithm to generate different initialized regions of the image. After initialization, the default similarity of all images is Y=100%.

[0117] Step S8: Perform mean, sample variance, and standardization processing on the original power grid image data of different initialization areas in sequence.

[0118] Step S9: Use the image extreme point detection method to extract features from the power grid image, and use scale-invariant feature transformation descriptors to represent the local features of the power grid image. This is divided into four steps: detecting key points in the power grid image, accurately locating key points of power grid faults, assigning principal directions to key points, and calculating scale-invariant feature transformation descriptors.

[0119] Step S10: Based on the feature extraction results of the feature fusion algorithm, aggregate images of the same power grid and separate images of different power grids.

[0120] Step S11: Select adjacent regions of the power grid image and match two adjacent regions. And calculate the similarity of the feature fusion results of two adjacent regions. , .

[0121] Step S12: If ,make Merge two adjacent regions Remove about and similarity and .

[0122] Step S13: Iterate through steps S11 and S12 until Y=0.

[0123] Step S14: Locate the power grid fault point from all areas.

[0124] like Figure 7 As shown, the basic steps for generating the GIM digital model of the power grid project in step S5 are as follows.

[0125] Step S51: Open the GIM modeling tool, create a new Symbol and assign it a device name and description.

[0126] Step S52: Define the starting insertion point of the device, and define the subordinate relationship between the device component and the sub-device component and the insertion position of the sub-component.

[0127] Step S53: Select the Basic Elements tab, select Basic Elements to create the device outline, and define the placement position when placing it.

[0128] Step S54: When the device is added, add basic camera and GIM parameter attributes to the device.

[0129] In step S7, the original overhead line image is modeled using a Gaussian Markov random field model. The model expression is as follows:

[0130] ;

[0131] In the formula, B represents the original power grid image, d represents the location of a portion of the overhead lines in the original power grid image, and γ represents the adjacent location, the location of the fault point. The probability is expressed as C represents a custom plane;

[0132] The original power grid image B was simulated using a Gaussian Markov random field model. The expression is:

[0133] ;

[0134] In the formula, Let represent the model parameters of the Gaussian Markov random field, d represent the location image of any point in the power grid, and a represent the size of the image. Represents discrete regions in a power grid image. Zero-mean Gaussian noise is represented by the energy function of the energy function of the dissected normal level set, which represents the interaction between pixels in the texture of the original power grid image.

[0135] ;

[0136] In the formula, These represent the average gray levels of the foreground and background regions of the original power grid image, respectively. Represents the entire set of regions in the image. This represents the mean of the energy function. Let n represent the integral over the power grid image, where n represents a single image region. Represents the observed image;

[0137] The expression for the mean gray level of the foreground region:

[0138] ;

[0139] In the formula, G represents the Heaviside function. Represents the observed image;

[0140] The expression for the average gray level of the background area is:

[0141] ;

[0142] In the formula, Represent the observed image and calculate the iteration. Continue until convergence, then output the segmented image to complete image segmentation.

[0143] In step S8, the original power grid image data is sequentially processed by mean, sample variance, and standardization, wherein:

[0144] The expression for the mean is:

[0145] ;

[0146] In the formula, This represents the mean value of the overhead power line image, where x represents any data point in the overhead power line image. denoted by , g represents the minimum batch processing value of the overhead line image, h represents the number of data points in the overhead line image, and h represents the sum of the gray values ​​of all data points in the overhead line image.

[0147] The expression for the sample variance is:

[0148] ;

[0149] In the formula, This represents the sample variance value of the power grid image;

[0150] The standardized processing expression is:

[0151] ;

[0152] In the formula, This represents the standardized value of the power grid image.

[0153] like Figure 8 As shown, in step S9, the steps for calculating the scale-invariant feature transformation descriptor are as follows:

[0154] Step S91: Using each keypoint as the center, rotate the X-axis of the power grid image coordinate system to the principal direction of the scale-invariant feature transformation keypoint to ensure rotation invariance;

[0155] Step S92: Divide the region into sub-regions centered on each key point, calculate the Gaussian weighted gradient direction angle histogram in each sub-region, and plot the cumulative value of each gradient direction to form a seed point;

[0156] A 16×16 pixel window is taken centered on each keypoint and divided into 4×4 sub-regions. A Gaussian weighted gradient direction angle histogram is calculated in each sub-region in 8 directions. The gradient direction information of pixels closer to the keypoint contributes more. The accumulated value of each gradient direction is plotted to form a seed point. Therefore, a total of 16 seed points can be generated. This way, a data of length 128 can be generated for each keypoint, which is to finally obtain a 128-dimensional scale-invariant feature transformation descriptor.

[0157] In step S11, the expression for the similarity function is established as follows:

[0158] ;

[0159] In the formula, F and G represent two adjacent power grid images, and E represents the set of all power grid images fused by edge features. The similarity coefficient of image x is represented. This represents the set of all feature points in an image.

[0160] Example 4:

[0161] like Figure 9 The diagram shown is an architectural schematic of a BIM-based visualized power grid fault identification and location device according to an embodiment of the present invention. This embodiment of the BIM-based visualized power grid fault identification and location device includes one or more processors 21 and a memory 22. Figure 9 Take a processor 21 as an example.

[0162] Processor 21 and memory 22 can be connected via a bus or other means. Figure 9 Taking the example of a connection between China and Israel via a bus.

[0163] The memory 22, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs and non-volatile computer-executable programs, such as the BIM-based modeling-based visualized power grid fault identification and location method in Embodiment 1. The processor 21 executes the BIM-based modeling-based visualized power grid fault identification and location method by running the non-volatile software programs and instructions stored in the memory 22.

[0164] Memory 22 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 22 may optionally include memory remotely located relative to processor 21, which can be connected to processor 21 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0165] The program instructions / modules are stored in the memory 22. When executed by one or more processors 21, they execute the BIM-based visualized power grid fault identification and location method described in Embodiment 1 above. For example, they execute the method described above. Figures 1-3 ,as well as Figures 5-8 The steps shown.

[0166] It is worth noting that the information interaction and execution process between the modules and units in the above-mentioned device and system are based on the same concept as the processing method embodiment of the present invention. For details, please refer to the description in the method embodiment of the present invention, and will not be repeated here.

[0167] Those skilled in the art will understand that all or part of the steps in the various methods of the embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0168] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A visual power grid fault identification and location method based on BIM modeling, characterized in that the method... The method comprises the following steps: Collecting the power grid system on site, and drawing the power grid engineering system diagram according to the on-site collection; wherein the power grid engineering system diagram contains one or more of the power equipment, equipment coordinates, equipment connection relationship, and equipment information; Loading the power grid engineering system diagram and the monitoring network distribution diagram at the same time, and establishing the association relationship between the corresponding power equipment and one or more monitoring points for the corresponding monitoring points distributed near the power equipment; Completing the video frame relationship of the one or more monitoring points on the power equipment under the best shooting angle combination through the power equipment recognition in the images collected by different monitoring points; wherein the best shooting angle refers to the video frames collected by the corresponding one or more monitoring points at the same time or asynchronous time, which can present the external and / or internal information of the power equipment to the greatest extent after completing the picture splicing of the video frames; Importing the power grid engineering system diagram and the video frame relationship of the one or more monitoring points associated with each power equipment obtained through analysis into the BIM to generate a three-dimensional model; wherein the BIM constructs a three-dimensional model of the power grid engineering system according to the power grid engineering system diagram and the video frames, and periodically acquires the corresponding video frame content to update the visual state of the power equipment in the three-dimensional model according to one or more of the shooting angle of the monitoring point camera of the corresponding video frame, the time difference between the video frames of different monitoring points, and the jurisdiction subject to which the different monitoring points belong; When the one or more monitoring points belong to different jurisdiction subjects, the completion of the video frame relationship of the one or more monitoring points on the power equipment under the best shooting angle combination specifically comprises: Acquiring the video content within a monitoring period of each monitoring point from the different jurisdiction subjects to which the one or more monitoring points belong; Selecting a main monitoring point from the one or more monitoring points first, and the others as auxiliary monitoring points, wherein the main monitoring point has one or more of the largest amount of information, the longest time of the main subject equipment within a period, and the guarantee of the completeness and timeliness of the video frames; Selecting a group of video frames that can best restore the visual information of the power equipment by splicing the basic frame and the video frames of the one or more monitoring points from the main monitoring point, and recording one or more of the shooting angle of the monitoring point camera in the corresponding group of video frames, the time point of the video frame of each monitoring point within the period, and the jurisdiction subject to which the different monitoring points belong; Sending a message carrying the shooting angle of the monitoring point camera and the time point of the corresponding video frame within the period to the jurisdiction subjects of the different monitoring points; The jurisdiction subjects of the different monitoring points will send a preset number of video frames collected under the corresponding conditions to the three-dimensional modeling system when the video stream collected by each monitoring point meets the message conditions of the shooting angle of the monitoring point camera and the time point of the corresponding video frame within the period.

2. The BIM modeling based visualized power grid fault identification and localization method according to claim 1, characterized in that, When the image stitching in the previous round is performed, the contour information and the key points of the power equipment are extracted as the basis for the corresponding stitching process; after the power equipment in the three-dimensional model is updated, the contour information and the key point information generated for each video frame are recorded separately, so that when a new round of video frames is obtained for stitching in the next update cycle, after the contour information and the key point information are generated, the contour information and the key point information corresponding to the previous round of video frames are matched, so that when the deviation exceeds the preset threshold, the adjustment message of the shooting angle of the monitoring point camera and the time point in the cycle corresponding to the video frame is sent to the jurisdiction subject of the different monitoring points.

3. The BIM modeling based visualized power grid fault identification and localization method according to claim 1 or 2, characterized in that, The power grid engineering system diagram is specifically a CAD two-dimensional measured drawing, and the method comprises the following steps: The actual situation of the power grid engineering is simulated by means of multi-position, multi-angle and multi-direction; The electrical equipment required by the power grid engineering is recorded, and the CAD two-dimensional measured drawing of the power grid engineering is drawn; The CAD two-dimensional measured sketch is imported into a BIM three-dimensional modeling tool, and the BIM three-dimensional model is obtained according to the control point coordinates.

4. The BIM modeling based visualized power grid fault identification and localization method according to claim 1 or 2, characterized in that, The image stitching under the video frame specifically comprises the following steps: The original power grid image is segmented by using an image segmentation algorithm to generate different initial regions of the image , and after initialization, the similarity of all images is Y=100% by default The original power grid image data of different initialization regions are subjected to mean value, sample variance and standardization processing in sequence; The feature extraction of the power grid image is performed by using an image extreme point detection method, and the local features of the power grid image are represented by using a scale invariant feature transformation descriptor, the key points of the power grid image are detected, the key points are accurately positioned, the main direction of the key points is allocated and the scale invariant feature transformation descriptor is calculated; According to the feature extraction result, the same power grid images are aggregated and the different power grid images are separated; Selecting adjacent regions of the power grid image and matching the two adjacent regions and calculating the similarity of the feature fusion result of the two adjacent regions , ; like ,make Merge two adjacent regions Remove about and similarity and The iterative calculation continues until Y=0, thus determining the splicing point.

5. The BIM modeling based visualized power grid fault identification and localization method of claim 4, wherein, The original power grid image is segmented by using an image segmentation algorithm to generate different initialization regions of the image, and the specific process comprises the following steps: The original overhead line image is modeled by using a Gaussian Markov random field model, and the model expression is as follows: ； In the formula, B represents an original power grid image, d represents a position of a part of overhead lines in the original power grid image, γ represents a position of an adjacent position, and a position of a fault point is expressed as C represents a self-defined plane. The original power grid image B is simulated by a Gaussian Markov random field model, The expression is: ； wherein, denote the model parameters of the Gaussian Markov Random Field, d denotes the location image of any point in the power grid, a denotes the size of the image, denote the discrete region of the power grid image, denote the zero-mean Gaussian noise, the interaction between the pixel points in the texture of the original power grid image is represented by using the disjunctive normal level set, and the energy function formula of the disjunctive normal level set is: ； wherein represent the mean gray value of the foreground region, the background region of the original power grid image, respectively, represent the set of all regions of the image, represent the mean value of the energy function, represent the integral over the power grid image, n represents a single image region, represent the observed image; The gray mean value expression of the foreground region is as follows: ； where G is the Heaviside function, represents the observed image; The gray mean value expression of the background region is as follows: ； In the formula, denotes the observed image, the iteration is calculated until convergence, the segmented image is output, and the image segmentation is completed.

6. The BIM modeling based visualized power grid fault identification and localization method of claim 4, wherein, The specific process that the original power grid image data of different initialization regions are subjected to mean value, sample variance and standardization processing in sequence comprises the following steps: The mean value expression is as follows: ； wherein, represents the overhead line image mean result, x represents any overhead line image data point, represents the overhead line image minimum batch value, g represents the overhead line image data point number, h represents the sum of all overhead line image data point grayscale values; The sample variance expression is as follows: ； In the formula, represents the sample variance value of the power grid image; The standardization processing expression is as follows: ； In the formula, denotes the normalized value of the power grid image.

7. The BIM modeling based visualized power grid fault identification and localization method of claim 4, wherein, The scale invariant feature transformation descriptor is calculated, and the specific process comprises the following steps: The power grid image coordinate X-axis is rotated to the main direction of the scale invariant feature transformation key point with each key point as the center; Each key point is segmented into a sub-region, the Gaussian weighted gradient direction angle histogram is calculated on each sub-region, the cumulative value of each gradient direction is drawn, and a seed point is formed, The expression of the similarity function is as follows: ； In the formula, F, G represent two adjacent power grid images, E represents all power grid image collections after edge feature fusion, represents the similarity coefficient of image x, represents all feature point collections of the image.

8. A device for visualizing and identifying and locating power grid faults based on BIM modeling, characterized by The device comprises: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the processor to execute the BIM modeling-based visual power grid fault identification and positioning method in any one of claims 1-7.

Images