Electricity charge management method and device based on electric meter data

Abstract

The present application relates to the field of electric meter monitoring management, and relates to an electricity charge management method and device based on electric meter data, the method comprising: acquiring electric meter image information, time-space stamp information corresponding to the electric meter image, and historical electricity consumption data of a user; identifying the electric meter image information to obtain a preliminary identification result; performing fusion processing according to the preliminary identification result, the historical electricity consumption data of the user, and the time-space stamp information corresponding to the electric meter image to obtain a reliable reading report attached with a confidence level and an abnormal type label; performing dynamic routing decision processing based on the confidence level and the abnormal type according to the reliable reading report to obtain electricity charge settlement processing decisions of different abnormal types; and managing electricity charges according to the electricity charge settlement processing decisions of different abnormal types, which effectively improves the intelligence and reliability of electricity charge management, reduces the electricity charge settlement error rate, and improves the electricity charge settlement efficiency and resource allocation rationality.

Description

A method and apparatus for electricity fee management based on electricity meter data Technical Field

[0001] This invention relates to the field of electricity meter monitoring and management, and more specifically, to a method and apparatus for electricity fee management based on electricity meter data. Background Technology

[0002] In the field of electricity meter monitoring and management, with the advancement of smart grid construction, traditional manual meter reading methods have been gradually replaced by automated data acquisition technologies. Existing technologies mainly acquire electricity billing data by deploying smart meters or using basic image recognition methods, such as recognizing meter readings through simple OCR technology or relying on regular manual inspections. However, these methods face many challenges in complex real-world application scenarios. For example, meter images are easily affected by lighting, angle, or occlusion, leading to recognition errors. Furthermore, anomaly handling often relies on fixed rules or manual intervention, lacking adaptive decision-making capabilities. This results in problems such as low electricity billing efficiency, high error rates, and suboptimal resource allocation, limiting the improvement of the intelligence and reliability of the electricity billing management system. Summary of the Invention

[0003] The purpose of this invention is to provide an electricity fee management method and apparatus based on electricity meter data to improve the above-mentioned problems.

[0004] To achieve the above objectives, the embodiments of this application provide the following technical solutions:

[0005] On the one hand, embodiments of this application provide an electricity fee management method based on electricity meter data, the method comprising:

[0006] Acquire meter image information, the corresponding spatiotemporal stamp information of the meter image, and the user's historical electricity consumption data;

[0007] The meter image information is identified to obtain preliminary identification results;

[0008] The preliminary identification results, the user's historical electricity consumption data, and the spatiotemporal stamp information corresponding to the meter image are fused together to obtain a reliable reading report with confidence level and anomaly type label.

[0009] Based on the trusted reading report, dynamic routing decision processing based on confidence level and anomaly type is performed to obtain electricity billing processing decisions for different anomaly types.

[0010] Electricity charges are managed based on the electricity billing processing decisions for the different types of anomalies.

[0011] Secondly, embodiments of this application provide an electricity fee management device based on electricity meter data, the device comprising:

[0012] The acquisition module is used to acquire meter image information, the time and space stamp information corresponding to the meter image, and the user's historical electricity consumption data;

[0013] The first processing module is used to identify the image information of the electricity meter and obtain preliminary identification results;

[0014] The second processing module is used to perform fusion processing based on the preliminary identification results, the user's historical electricity consumption data, and the spatiotemporal stamp information corresponding to the meter image to obtain a reliable reading report with confidence level and anomaly type label;

[0015] The decision module is used to perform dynamic routing decision processing based on the confidence level and anomaly type according to the trusted reading report, and to obtain electricity billing processing decisions for different anomaly types.

[0016] The management module is used to manage electricity charges based on the electricity charge settlement processing decisions for the different anomaly types.

[0017] Thirdly, embodiments of this application provide an electricity fee management device based on electricity meter data, the device including a memory and a processor. The memory is used to store a computer program; the processor is used to execute the computer program to implement the steps of the above-described electricity fee management method based on electricity meter data.

[0018] Fourthly, embodiments of this application provide a readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described electricity fee management method based on meter data.

[0019] The beneficial effects of this invention are as follows:

[0020] This invention obtains preliminary identification results by acquiring meter image information, its corresponding spatiotemporal stamp information, and users' historical electricity consumption data. Then, it combines the preliminary identification results, historical electricity consumption data, and spatiotemporal stamp information for fusion processing to generate a reliable reading report with confidence level and anomaly type labels. Subsequently, based on the reliable reading report, dynamic routing decisions are made to obtain electricity billing processing decisions for different anomaly types. Finally, electricity bills are managed according to these decisions, effectively improving the intelligence and reliability of electricity billing management, reducing the error rate of electricity billing, and improving the efficiency of electricity billing and the rationality of resource allocation.

[0021] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 is a schematic diagram of the electricity fee management method based on electricity meter data according to an embodiment of the present invention.

[0024] Figure 2 is a schematic diagram of the structure of the electricity fee management device based on electricity meter data according to an embodiment of the present invention.

[0025] Figure 3 is a schematic diagram of the structure of the electricity fee management device based on electricity meter data according to an embodiment of the present invention.

[0026] The diagram is labeled as follows: 800, Electricity fee management device based on electricity meter data; 801, Processor; 802, Memory; 803, Multimedia component; 804, I / O interface; 805, Communication component; 901, Acquisition module; 902, First processing module; 903, Second processing module; 904, Decision module; 905, Management module. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0028] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0029] Example 1:

[0030] This embodiment provides a method for electricity fee management based on electricity meter data. It can be understood that a scenario can be set up in this embodiment, such as: staff uploading electricity meter data to settle electricity fees.

[0031] Referring to Figure 1, the figure shows that the method includes steps S1-S5.

[0032] Step S1: Obtain the meter image information, the time and space stamp information corresponding to the meter image, and the user's historical electricity consumption data;

[0033] Step S1 further includes steps S11-S15, which specifically include:

[0034] Step S11: Acquire data from various sensors built into the image acquisition device;

[0035] In this step, the image acquisition device can be an intelligent inspection terminal integrating multiple sensors. The built-in sensors specifically include a GPS sensor, an IMU unit, a gravity sensor, and a light sensor. The GPS sensor collects the device's current geographical coordinates, providing a basis for subsequent spatial positioning. The IMU unit, composed of an accelerometer and a gyroscope, can capture the device's angular velocity and linear acceleration in real time, with a frequency up to 100Hz, reflecting the device's motion status. The gravity sensor detects the device's tilt angle relative to the horizontal plane, and the light sensor collects the current ambient light intensity. In power inspection scenarios, when maintenance personnel use this terminal near meter boxes in residential buildings or industrial parks, the terminal synchronously and continuously collects data from these sensors. This provides raw data support for subsequent judgment of the device's shooting posture and ensures the quality of meter image acquisition, avoiding deviations in subsequent shooting posture judgments due to missing initial data.

[0036] Step S12: The sensor data is fused to obtain the spatial coordinates and attitude information of the image acquisition device;

[0037] In this step, the Kalman filter algorithm from multi-sensor data fusion technology is used to achieve data fusion. First, the geographical coordinates collected by the GPS sensor are filtered and denoised to eliminate coordinate fluctuations caused by signal drift. Then, the angular velocity and linear acceleration data collected by the IMU are synchronized with the tilt angle data of the gravity sensor. Through the prediction-update stage of the Kalman filter, the measurement error of a single sensor is corrected. For example, the IMU has high accuracy in the short term but is prone to drift in the long term, while the GPS is stable in the long term but has noise in the short term. After fusion, more accurate three-dimensional spatial coordinates and attitude information of the device can be obtained. Since meter boxes may be densely distributed in the same area, accurate spatial coordinates can distinguish the location of different meter boxes, and attitude information provides data basis for subsequent adjustment of shooting angle and ensuring that the lens is facing the meter dial, avoiding deviation of shooting direction due to blurry device attitude.

[0038] Step S13: Perform attitude analysis based on the spatial coordinates and attitude information to obtain the shooting pose;

[0039] Step S13 further includes steps S131, S132, S133, and S134, which specifically include:

[0040] Step S131: Calculate the geometric relationship between the current shooting plane of the image acquisition device and the plane where the meter dial is located based on the attitude data of the spatial coordinates and orientation of the image acquisition device, and obtain the actual spatial angle between the normal vector of the current shooting plane and the normal vector of the plane where the meter dial is located;

[0041] Step S132: Compare the actual spatial angle with a preset angle threshold to obtain a pose adjustment guidance command;

[0042] In this step, the angle threshold is set to 5°. If the actual spatial angle is less than or equal to the preset angle threshold, the current device pose does not need to be adjusted. If the actual spatial angle is greater than the preset angle threshold, the specific parameters that need to be adjusted are determined by inverse kinematics calculation based on the direction difference between the two plane normal vectors and the device attitude parameters. These parameters include the roll angle adjustment angle (e.g., +3° means a 3° roll to the right), the pitch angle adjustment angle (e.g., -2° means a 2° pitch downwards), and the distance adjustment value between the device and the meter (e.g., +0.3m means a 0.3m forward movement). These parameters are then integrated into a pose adjustment guidance command.

[0043] Step S133: Based on the pose adjustment guidance command, the guidance screen is obtained through real-time rendering and display processing of the augmented reality interface;

[0044] In this step, based on the obtained pose adjustment guidance instructions, the device's built-in AR rendering engine is invoked. This engine employs simultaneous localization and mapping (SLAM) technology to fuse and render the real-time captured environmental images (electricity meter and surrounding scene) with virtual guidance elements. A specific implementation method is as follows: based on the angle parameters in the adjustment instructions, a virtual arrow with angle annotations is generated, such as a red arrow pointing in the direction to be adjusted, labeled "tilt 3° to the left"; based on the distance adjustment value, a virtual distance scale is generated, such as a blue line segment marking the difference between the current distance and the target distance. These virtual elements are then overlaid on the corresponding positions in the real-time view, ensuring spatial alignment between the virtual elements and the real electricity meter scene, forming a guidance screen that is displayed in real-time on the device screen, improving the intuitiveness of pose adjustment.

[0045] Step S134: Adjust the device pose in real time according to the guide screen to obtain the shooting pose.

[0046] Step S14: Acquire an image of the electricity meter according to the shooting pose and evaluate the quality of the electricity meter image to obtain a qualified electricity meter image;

[0047] In this step, a convolutional neural network model is used to analyze the features of key areas in the image. Specifically, it detects the local sharpness of the digital display area to assess whether motion blur exists; analyzes the overall illumination uniformity of the dial to identify reflections or shadows; and combines prior knowledge of the meter (such as the digital area usually being located in the center of the dial) to locate the key evaluation areas. If the overall score is lower than the preset evaluation threshold, the image is retaken.

[0048] Step S15: Bind the qualified electricity meter image, the spatial coordinates of the image acquisition device, and the attitude information to the spatiotemporal stamp and data integrity to obtain the spatiotemporal stamp information corresponding to the electricity meter image.

[0049] In this step, the timestamp uses UTC time format to record the specific time of image acquisition; the spatial stamp represents the device's spatial coordinates. A cryptographic hash function is used to bind the qualified meter image, the spatial coordinates of the image acquisition device, its attitude information, and the spatial stamp together for data integrity. This data is integrated into a single data packet, and a unique hash value is generated using the SHA-256 hash algorithm. This hash value corresponds one-to-one with the data packet; if subsequent data is tampered with, the hash value will change, and data integrity can be determined by verifying the hash value. In electricity billing scenarios, the timestamp accurately traces the time and location of meter data acquisition, while hash binding prevents data tampering, providing traceable and tamper-proof original data evidence for subsequent electricity billing disputes, thus protecting the rights of both users and the power company.

[0050] Step S2: Recognize the meter image information to obtain preliminary recognition results;

[0051] Step S2 further includes steps S21-S24, which specifically include:

[0052] Step S21: Use the target detection model to detect the meter image information and obtain the detection results of the key areas of the meter.

[0053] In this step, YOLOv5 is selected as the object detection model. A clustering algorithm is used to re-cluster the bounding box sizes in the meter dataset, generating anchor box sizes more suitable for the meter components. It is understood that the detection results for key areas of the meter include the dial area, nameplate area, and brand logo area, etc.

[0054] Step S22: Classify the meter models according to the detection results of the key areas of the meter to obtain the meter model information;

[0055] In this step, a dual-branch network is used to process the visual and textual features included in the key region detection results: the visual branch uses a convolutional neural network to extract image features of the located logo region; the text branch uses a convolutional recurrent neural network to perform end-to-end text recognition on the nameplate region image. Then, a feature-level fusion strategy is used to concatenate the visual feature vector and the textual feature vector, and finally, a fully connected layer outputs the model classification result. This method effectively solves the recognition bottleneck of a single modality when the logo is worn or the nameplate is damaged, significantly improving the reliability of model classification.

[0056] Step S23: Based on the meter model information, query the preset meter model knowledge base to obtain the reading area and structured description of the reading area corresponding to the meter model.

[0057] In this step, precise location of the reading area is achieved through querying and parsing based on a pre-defined meter model knowledge base. This pre-defined knowledge base uses a graph database to store structured information about different meter models, including: the relative coordinate range of the reading area within the meter face image (e.g., top-left corner coordinates (x, y), width w, height h), digital display type (mechanical pointer / digital LCD), digit distribution, decimal point position, and other metadata. Once the meter model is obtained, it is used as an index to perform precise matching within the knowledge base. The relative coordinates of the reading area corresponding to that model are retrieved, and combined with the actual resolution of the meter image, the actual pixel coordinates of the reading area are calculated. Simultaneously, the corresponding structured description is retrieved to form a complete query result. This prior knowledge-based location method avoids the efficiency problem of blind searching in general OCR algorithms and is particularly suitable for handling scenarios with a wide variety of meter models.

[0058] Step S24: Perform model-adaptive reading extraction processing based on the reading area corresponding to the meter model and the structured description of the reading area to obtain the preliminary identification result of the meter reading.

[0059] Step S24 further includes steps S241-S244, which specifically include:

[0060] Step S241: Perform multi-dial center positioning processing based on the reading area corresponding to the meter model and the structured description of the reading area to obtain the center coordinates and radius information of each pointer dial;

[0061] In this step, the actual pixel coordinates of the reading area corresponding to the meter model are used as the range. Multi-dial image blocks of the mechanical meter are cropped out, and the Hough circle detection algorithm is used to locate the dial center. The specific implementation process includes: preprocessing the image blocks to obtain preprocessed image blocks; performing edge detection on the preprocessed image blocks; then identifying the circular contours through probabilistic Hough circle transform; and finally detecting the circular contours to output the center coordinates and radius information of each pointer dial. It should be noted that the key improvement in this step lies in introducing a concentric circle constraint: setting a threshold for the deviation of the center coordinates of adjacent dials. When the center coordinates of multiple detected circles fall within this threshold range, they are determined to be the same group of concentric circle dials. Simultaneously, prior knowledge (such as the main dial usually having the largest radius) is used to sort and filter the detection results. This application effectively solves the problem that traditional Hough transform is prone to false detection or missed detection when the dial edges are blurred or reflective, significantly improving the robustness of recognition of dense dial structures under complex lighting conditions.

[0062] Step S242: Measure the angle of each dial pointer based on the center coordinates and radius information of each dial to obtain the effective angle value of each dial pointer;

[0063] In this step, a polar coordinate transformation is first performed on each dial area to convert the circular dial image into a rectangular unfolded image for easier analysis. Then, the Radon transform is used to perform projection integration at different angular directions, and the coarse direction of the pointer is determined by finding the maximum projection value. Within a small angular range near the coarse direction, the gradient direction histogram of the local image is calculated, and the pointer angle is refined by statistically analyzing the mode of the gradient direction, thus obtaining the effective angle value of each dial pointer. This two-stage measurement strategy utilizes both the sensitivity of the Radon transform to the overall contour and the ability of local gradient statistics to capture edge details, effectively solving the edge blurring problem caused by long-term use of the meter, such as pointer wear or dial dirt, and improving reading accuracy.

[0064] Step S243: Perform multi-dial reading combination and carry-over calculation based on the effective angle value of each dial pointer to obtain the preliminary reading of the integer part of the meter.

[0065] This step employs a multi-dial reading synthesis algorithm based on a carry state machine. The algorithm processes each dial sequentially from the least significant digit to the most significant digit. Specifically, it reads the pointer angle of the units digit dial and converts it to a value between 0 and 9. Then, it processes the tens digit dial, considering the potential carry from the units digit dial—a carry occurs when the pointer approaches the zero mark and the value changes from 9 to 0. The state machine model explicitly defines the numerical range, carry conditions, and carry propagation rules for each dial. For each dial, the algorithm not only reads the current pointer value but also considers the states of adjacent dials to determine if a carry has occurred, ensuring the accuracy of the synthesized readings. This solves the carry error problem that might arise from simply identifying the values ​​of each dial independently.

[0066] Step S244: Based on the preliminary reading of the integer part of the meter, perform decimal place recognition to obtain the preliminary recognition result of the meter reading.

[0067] In this step, the meter model knowledge base is first queried to confirm whether the model includes a separate decimal dial and its location information. If a decimal dial exists, its reading is obtained using the same detection method as for integer dials. Then, a weighting coefficient is determined based on the type of the decimal dial (tenths or hundredths). The final reading is obtained by a weighted sum of the integer and decimal parts. It should be noted that for some meter models, the decimal part may be represented by the pointer's subdivisions on the integer dial. In this case, interpolation calculations are required based on the scale division rules stored in the knowledge base. This flexible decimal processing mechanism adapts to the design differences of different types of meters, ensuring the completeness and accuracy of the reading extraction.

[0068] Step S3: Based on the preliminary identification results, the user's historical electricity consumption data, and the spatiotemporal stamp information corresponding to the meter image, perform fusion processing to obtain a reliable reading report with confidence level and anomaly type label;

[0069] Step S3 further includes steps S31-S34, which specifically include:

[0070] Step S31: Perform a rationality verification process based on the preliminary identification results and the user's historical electricity consumption data to obtain a rationality verification result;

[0071] In this step, the user's historical electricity consumption data for the past 6 months is divided into time periods: 8-12 AM, 12-6 PM, 6-12 AM, and 0-8 AM. The mean and standard deviation of the electricity consumption data for each time period are calculated, and a 95% confidence interval is set. The electricity consumption periods and data corresponding to the preliminary identification results are extracted, and it is determined whether the electricity consumption data falls within the corresponding 95% confidence interval. Simultaneously, if the electricity consumption in the current time period increases by more than 300% compared to the previous time period, it is marked as an abrupt change anomaly. The judgment results of "normal within the interval," "abnormal outside the interval," and "abnormal change" are used as the reasonableness verification results. This step can distinguish the historical electricity consumption characteristics of weekdays and weekends, avoiding misjudging normal electricity consumption as abnormal due to not considering time period differences, and ensuring that the verification results are consistent with actual electricity consumption habits. It can be understood that the 95% confidence interval means that in multiple samplings, 95% of the sample data will fall within this interval; in this scenario, it is used to define the range of normal electricity consumption.

[0072] Step S32: Perform a group electricity consumption pattern consistency analysis based on the spatiotemporal stamp information corresponding to the meter image and the meter model information to obtain the group consistency analysis result;

[0073] This step utilizes a spatiotemporally constrained density clustering algorithm to analyze the consistency of electricity consumption in a group. First, based on spatiotemporal stamp information, proximity relationships are constructed in the spatiotemporal dimension. Spatially, meters within a 500-meter radius are selected; temporally, readings within one hour before and after the collection time are chosen, resulting in a spatiotemporally correlated meter sample set. The electricity consumption change of each meter is calculated as a feature vector based on this spatiotemporally correlated sample set, yielding a sample dataset with electricity consumption characteristics. Based on this sample dataset with electricity consumption characteristics, an improved DBSCAN clustering process is applied to obtain preliminary clustering results. Finally, based on these preliminary clustering results, pattern recognition and outlier detection are performed to obtain the final group electricity consumption consistency analysis results. It should be noted that the improved DBSCAN clustering process specifically incorporates spatial distance constraints into the density calculation; the similarity between two meter samples depends not only on the numerical closeness of their values ​​but also on their spatial proximity.

[0074] Step S33: Perform anomaly detection based on the meter image information and the corresponding spatiotemporal stamp information to obtain anomaly detection results;

[0075] In terms of image physical anomaly detection, this step uses an image anomaly recognition model to process the meter image information. The image anomaly recognition model is a two-branch design with a shared backbone: the backbone network uses ResNet-50 to extract general image features; the first branch is a target detection head, responsible for detecting specific targets and outputting their bounding boxes and state classifications (intact / damaged / missing); the second branch is a semantic segmentation head, responsible for segmenting the meter glass area and calculating the proportion of occlusion. The loss functions of the two branches are dynamically weighted. For meter anomaly features, positive samples such as lead cutting, meter cover opening, and illegal wiring are added to the training dataset. In terms of spatiotemporal logic anomaly detection, the historical collection records of the current meter are first queried. If multiple different meter images correspond to the same spatiotemporal stamp, the spatiotemporal stamp is marked as duplicated. If the current meter reading shows a negative increase compared to the previous collection reading (e.g., the previous reading was 1200 kWh, the current reading is 1180 kWh, excluding legitimate operations such as meter zeroing), the reading logic is marked as abnormal. This step can simultaneously identify both physical damage to the meter and data acquisition logic problems, avoiding the omission of anomalies caused by only looking at the readings without considering the image / spatiotemporal logic.

[0076] Step S34: Generate a reliable reading report based on the rationality verification results, the group consistency analysis results, and the anomaly detection results.

[0077] In this step, the rationality verification result is first assigned a weight of 0.3, the group consistency analysis result a weight of 0.3, and the anomaly detection result a weight of 0.4. A weighted scoring method is then used to calculate the total confidence score, and confidence levels are determined based on the total score. It should be noted that this invention does not limit the threshold range for determining confidence levels based on the total score.

[0078] Step S4: Perform dynamic routing decision processing based on confidence level and anomaly type according to the trusted reading report to obtain electricity billing processing decisions for different anomaly types;

[0079] Step S4 further includes steps S41-S43, which specifically include:

[0080] Step S41: Perform preliminary screening of routing channels based on fuzzy rules according to the confidence level of the trusted reading report to obtain a preliminary set of routing channel suggestions. The preliminary set of routing channel suggestions includes a set of several routing channel options suitable for processing trusted reading reports.

[0081] In this step, based on the confidence level (high, medium, low) of the reliable reading report, discrete levels are mapped to continuous membership degrees. For example, using a triangular membership function, the membership degree of high confidence increases linearly from 0 to 1 in the 0.8-1.0 range, medium confidence forms a triangular distribution with a peak of 1 in the 0.6-0.9 range, and low confidence decreases linearly from 1 to 0 in the 0-0.7 range, thus obtaining the fuzzy membership degree value of the confidence level. Based on the fuzzy membership degree value of the confidence level and the preset fuzzy rule base, the input membership degree is matched with the rule antecedent. The rule trigger strength is obtained by taking the smaller value, and then the membership degree of the consequent is pruned. Finally, the fuzzy output set of each channel is synthesized by taking the larger value, resulting in the fuzzy output of the automatic channel, manual channel, and abnormal channel. Based on the fuzzy output of each routing channel, the centroid coordinates of the fuzzy output set are calculated, that is, the weighted average of the area enclosed by the channel membership function and the horizontal axis is calculated, with the weight being the corresponding membership degree value, thus obtaining the membership degree weight of each channel and forming the initial screening result of the routing channel.

[0082] Step S42: Perform weighted priority calculation based on the set of routing channel suggestions and the anomaly type labels in the trusted reading report to obtain the final routing channel instruction;

[0083] This step first establishes a hierarchical structure: the target layer determines the optimal routing decision, while the criteria layer includes three dimensions: processing efficiency, risk control, and resource consumption. Considering the characteristics of electricity fee management, priority weights are assigned to different anomaly types (e.g., suspected electricity theft > equipment failure > identification error), and a judgment matrix is ​​constructed using expert scoring to calculate subjective weights. Simultaneously, based on historical work order data, the objective weights of each criterion are calculated using the entropy weight method, and a comprehensive weight is obtained through weighted combination. The comprehensive weight is then combined with the membership weights of each channel to calculate the priority score for each channel, thus obtaining the final routing channel instruction. This step effectively balances business rule constraints with real-time status adaptation, ensuring that high-risk anomalies are prioritized for manual review channels.

[0084] Step S43: Determine the electricity billing processing decision for different anomaly types based on the final routing channel instruction and the anomaly type label in the trusted reading report.

[0085] In this step, the automatic settlement channel is used when the anomaly type is "no anomaly"; the manual review channel is used when the anomaly type is "suspected identification error"; and the anomaly handling channel is used when the anomaly type is "suspected electricity theft". It should be noted that when the channel is set to automatic processing but the historical anomaly frequency is high, a random sampling rule is added.

[0086] Step S5: Manage electricity fees according to the electricity fee settlement processing decisions for the different anomaly types.

[0087] Step S5 further includes steps S51-S54, which specifically include:

[0088] Step S51: Determine whether manual intervention is needed based on the electricity billing processing decision for the different types of anomalies, and obtain the determination result;

[0089] Step S52: When the judgment result indicates that manual intervention is required, obtain the information of the verification work order to be dispatched, the real-time location and skill data of available maintenance personnel, and perform multi-objective work order dispatch optimization model construction processing to obtain the work order dispatch optimization model.

[0090] In this step, the objective function of the work order assignment optimization model is specifically:

[0091] ;

[0092] In the above formula, , , These represent distance cost weight, load balancing weight, and skill matching weight, respectively. This represents a binary variable; it is 1 when work order i is assigned to personnel j, and 0 otherwise. n represents the number of work orders to be processed; m represents the number of available maintenance personnel. A function representing the distance between work order position pi and personnel position lj; Indicates the estimated time for personnel j to process work order i; This represents the maximum working time of person j; This represents the skill matching degree between person j and work order type i.

[0093] The constraints are:

[0094] ;

[0095] In the above formula, This represents the skill level vector of person j; This indicates the minimum skill level required for work order i; This indicates the maximum permissible service distance. It should be noted that the skill level vector is obtained through the fusion and quantification of multi-source data. This multi-source data includes: personnel qualification certificates; a historical work order database: analyzing all work orders previously handled by maintenance personnel, extracting the distribution of successfully handled anomaly types, the types of meters handled, average resolution time, and re-dispatch rate; and assessment results from training courses.

[0096] A standardized skill level vector S=[S1,S2,S3,S4] is defined, where S1 represents anti-electricity theft investigation capability, weighted by the success rate of handling suspected electricity theft work orders, the number of historical cases detected, and relevant qualifications; S2 represents equipment fault diagnosis capability, scored based on the efficiency and model coverage of handling various hardware fault work orders; S3 represents communication module debugging capability, assessing the skill in resolving communication problems for smart meters with communication functions; and S4 represents comprehensive business familiarity, considering factors such as years of service, the diversity of work orders handled, and user satisfaction. The features in the skill level vector are weighted and calculated based on multi-source data to obtain the skill level vector for each maintenance personnel.

[0097] Step S53: Solve the multi-objective work order dispatch optimization model using a genetic algorithm to obtain a set of work order dispatch schemes;

[0098] In this step, based on the work order information (including location, required skills, and processing time limit) and maintenance personnel data (including location, skill tags, and current load), heuristic initialization prioritizes assigning each work order to the nearest personnel with a skill match rate ≥80%. Work orders without matching personnel are randomly assigned to personnel who meet the basic skill requirements, resulting in an initial population. Based on the initial population, the objective function value of each solution is calculated, and then the solutions are categorized according to constraint satisfaction: solutions that fully satisfy the constraints enter higher levels, while solutions that violate the processing time limit constraint enter lower levels. Within the same level, solutions are sorted according to the dominance relationship of the objective function, resulting in a hierarchically sorted population. Based on the sorting... After the initial population is processed, the selection operation uses tournament selection to retain high-quality individuals; the crossover operation uses simulated binary crossover with a crossover probability of 0.9 to recombine the gene relationship between work orders and personnel; the mutation operation combines polynomial mutation and exchange mutation operators with a mutation probability of 0.1, where the exchange mutation randomly selects two work orders to exchange their assigned personnel, resulting in a new generation of population; based on the new generation of population, the non-dominated sorting and genetic operations are repeated through iterative optimization with an early termination condition, and the optimal solution is recorded in each generation. The iteration terminates when the optimal solution has not improved for 50 consecutive generations, resulting in a uniformly distributed Pareto optimal solution set, which is the work order assignment scheme set.

[0099] Step S54: Use the fuzzy comprehensive evaluation method to select the assignment scheme with the highest overall satisfaction from the set of work order assignment schemes.

[0100] In this step, the estimated completion time, skill matching degree, and travel distance of each work order assignment scheme in the Pareto optimal solution set are extracted as evaluation indicators. Positive ideal solutions consist of the optimal values ​​for each dimension (shortest time, highest matching degree, shortest distance), while negative ideal solutions consist of the worst values ​​for each dimension. The Euclidean distance between each scheme and the positive and negative ideal solutions is calculated to obtain the relative proximity. The scheme with the highest relative proximity is selected as the assignment scheme with the highest overall satisfaction in the work order assignment scheme set. This ensures that the decision-making process conforms to both multi-objective optimization results and business priority requirements.

[0101] Example 2:

[0102] As shown in Figure 2, this embodiment provides an electricity fee management device based on electricity meter data. The device includes an acquisition module 901, a first processing module 902, a second processing module 903, a decision module 904, and a management module 905, specifically including:

[0103] The acquisition module 901 is used to acquire electricity meter image information, the time and space stamp information corresponding to the electricity meter image, and the user's historical electricity consumption data.

[0104] The first processing module 902 is used to identify the meter image information and obtain preliminary identification results;

[0105] The second processing module 903 is used to perform fusion processing based on the preliminary identification results, the user's historical electricity consumption data and the spatiotemporal stamp information corresponding to the meter image, to obtain a reliable reading report with confidence level and anomaly type label;

[0106] Decision module 904 is used to perform dynamic routing decision processing based on confidence level and anomaly type according to the trusted reading report, and obtain electricity billing processing decisions for different anomaly types;

[0107] The management module 905 is used to manage electricity fees based on the electricity fee settlement processing decisions for the different anomaly types.

[0108] In one specific embodiment of this disclosure, the acquisition module includes an acquisition unit, a first processing unit, a second processing unit, a third processing unit, and a fourth processing unit, specifically comprising:

[0109] The acquisition unit is used to acquire data from various sensors built into the image acquisition device;

[0110] The first processing unit is used to fuse the sensor data to obtain the spatial coordinates and attitude information of the image acquisition device.

[0111] The second processing unit is used to perform attitude analysis based on the spatial coordinates and attitude information to obtain the shooting pose.

[0112] The third processing unit is used to acquire the meter image according to the shooting posture and to evaluate the quality of the meter image to obtain a meter image with qualified quality.

[0113] The fourth processing unit is used to bind the qualified electricity meter image, the spatial coordinates of the image acquisition device, and the attitude information with spatiotemporal stamps and data integrity to obtain the spatiotemporal stamp information corresponding to the electricity meter image.

[0114] In one specific embodiment of this disclosure, the second processing unit includes a fifth processing unit, a sixth processing unit, a seventh processing unit, and an eighth processing unit, specifically including:

[0115] The fifth processing unit is used to calculate the geometric relationship between the current shooting plane of the image acquisition device and the plane where the meter dial is located based on the attitude data of the spatial coordinates and orientation of the image acquisition device, and to obtain the actual spatial angle between the normal vector of the current shooting plane and the normal vector of the plane where the meter dial is located;

[0116] The sixth processing unit is used to compare the actual spatial angle with a preset angle threshold to obtain a pose adjustment guidance command;

[0117] The seventh processing unit is used to adjust the guidance instructions according to the pose, and then render and display the guidance screen in real time through the augmented reality interface.

[0118] The eighth processing unit is used to adjust the device pose in real time according to the guide screen to obtain the shooting pose.

[0119] In one specific embodiment of this disclosure, the second processing module includes a ninth processing unit, a tenth processing unit, an eleventh processing unit, and a twelfth processing unit, specifically comprising:

[0120] The ninth processing unit is used to perform a rationality verification process based on the preliminary identification result and the user's historical electricity consumption data to obtain a rationality verification result;

[0121] The tenth processing unit is used to perform a group electricity consumption pattern consistency analysis based on the spatiotemporal stamp information corresponding to the meter image and the meter model information, and to obtain the group consistency analysis result.

[0122] The eleventh processing unit is used to perform anomaly detection based on the electricity meter image information and the corresponding spatiotemporal stamp information of the electricity meter image, and obtain anomaly detection results.

[0123] The twelfth processing unit is used to generate a reliable reading report based on the rationality verification result, the group consistency analysis result, and the anomaly detection result.

[0124] It should be noted that the specific manner in which each module performs its operation in the apparatus described in the above embodiments has been described in detail in the embodiments of the method, and will not be elaborated here.

[0125] Example 3:

[0126] Corresponding to the above method embodiments, this embodiment also provides an electricity fee management device based on electricity meter data. The electricity fee management device based on electricity meter data described below and the electricity fee management method based on electricity meter data described above can be referred to each other.

[0127] Figure 3 is a block diagram illustrating an electricity fee management device 800 based on electricity meter data according to an exemplary embodiment. As shown in Figure 3, the electricity fee management device 800 based on electricity meter data may include a processor 801 and a memory 802. The electricity fee management device 800 based on electricity meter data may also include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0128] The processor 801 controls the overall operation of the electricity billing management device 800 based on electricity meter data to complete all or part of the steps in the electricity billing management method based on electricity meter data described above. The memory 802 stores various types of data to support the operation of the electricity billing management device 800 based on electricity meter data. This data may include, for example, instructions for any application or method operating on the electricity billing management device 800 based on electricity meter data, as well as application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the electricity meter-based billing management device 800 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0129] In an exemplary embodiment, the electricity billing management device 800 based on electricity meter data may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the electricity billing management method based on electricity meter data described above.

[0130] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the electricity fee management method based on meter data described above. For example, the computer-readable storage medium may be the memory 802 including program instructions described above, which may be executed by the processor 801 of the electricity fee management device 800 based on meter data to complete the electricity fee management method based on meter data described above.

[0131] Example 4:

[0132] Corresponding to the above method embodiments, this embodiment also provides a readable storage medium. The readable storage medium described below can be referred to in conjunction with the electricity fee management method based on electricity meter data described above.

[0133] A readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the electricity fee management method based on electricity meter data described in the above method embodiments.

[0134] Specifically, the readable storage medium can be a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, or any other readable storage medium capable of storing program code.

[0135] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0136] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for electricity fee management based on electricity meter data, characterized in that, include: Acquire meter image information, the corresponding spatiotemporal stamp information of the meter image, and the user's historical electricity consumption data; The meter image information is identified to obtain preliminary identification results; The preliminary identification results, the user's historical electricity consumption data, and the spatiotemporal stamp information corresponding to the meter image are fused together to obtain a reliable reading report with confidence level and anomaly type labels. Dynamic routing decision processing based on confidence level and anomaly type is performed on the reliable reading report to obtain electricity billing decisions for different anomaly types. Electricity billing is managed based on the electricity billing decisions for different anomaly types. The fusion processing based on the preliminary identification results, the user's historical electricity consumption data, and the spatiotemporal stamp information corresponding to the meter image includes: performing a rationality verification process based on the preliminary identification results and the user's historical electricity consumption data to obtain a rationality verification result; performing a group electricity consumption pattern consistency analysis process based on the spatiotemporal stamp information corresponding to the meter image and the meter model information to obtain a group consistency analysis result, including: based on... The spatiotemporal stamp information is used to construct proximity relationships in the spatiotemporal dimension, that is, selecting electricity meters within a radius of 500 meters in space and selecting readings within one hour before and after the collection time in time, to obtain a spatiotemporally associated electricity meter sample set; the electricity consumption change of each electricity meter is calculated as a feature vector based on the spatiotemporally associated electricity meter sample set, to obtain a sample dataset with electricity consumption characteristics; based on the sample dataset with electricity consumption characteristics, an improved DBSCAN clustering process is used to obtain preliminary clustering results; based on the preliminary clustering results, pattern recognition and outlier identification processing are used to obtain the group electricity consumption consistency analysis results, and the improved DBSCAN clustering process incorporates spatial distance constraints into density calculation; anomaly detection is performed based on the electricity meter image information and the spatiotemporal stamp information corresponding to the electricity meter image to obtain anomaly detection results; a reliable reading report is generated based on the rationality verification results, the group consistency analysis results, and the anomaly detection results.

2. The electricity fee management method based on meter data according to claim 1, characterized in that, Acquiring the meter image information and the corresponding spatiotemporal stamp information includes: acquiring data from multiple sensors built into the image acquisition device; fusing the sensor data to obtain the spatial coordinates and attitude information of the image acquisition device; performing attitude analysis based on the spatial coordinates and attitude information to obtain the shooting pose; acquiring a meter image based on the shooting pose and evaluating the quality of the meter image to obtain a qualified meter image; binding the qualified meter image, the spatial coordinates of the image acquisition device, and the attitude information with spatiotemporal stamps and data integrity to obtain the spatiotemporal stamp information corresponding to the meter image.

3. The electricity fee management method based on meter data according to claim 2, characterized in that, The acquisition guidance process based on the spatial coordinates and attitude information includes: calculating the geometric relationship between the current shooting plane of the image acquisition device and the plane where the meter dial is located based on the attitude data of the spatial coordinates and orientation of the image acquisition device, and obtaining the actual spatial angle between the normal vector of the current shooting plane and the normal vector of the plane where the meter dial is located; comparing the actual spatial angle with a preset angle threshold to obtain a pose adjustment guidance command; obtaining a guidance screen through real-time rendering and display processing of the pose adjustment guidance command via an augmented reality interface; and adjusting the device pose in real-time based on the guidance screen to obtain the shooting pose.

4. The electricity fee management method based on meter data according to claim 1, characterized in that, The dynamic routing decision processing based on confidence level and anomaly type according to the trusted reading report includes: performing preliminary routing channel screening based on fuzzy rules according to the confidence level of the trusted reading report to obtain a preliminary set of routing channel suggestions, the preliminary set of routing channel suggestions including a set of several routing channel options suitable for processing the trusted reading report; performing weighted priority calculation based on the set of routing channel suggestions and the anomaly type labels in the trusted reading report to obtain the final routing channel instruction; and determining the electricity billing processing decision for different anomaly types based on the final routing channel instruction and the anomaly type labels in the trusted reading report.

5. The electricity fee management method based on meter data according to claim 1, characterized in that, Electricity billing is managed based on the electricity billing settlement processing decisions for different anomaly types, including: determining whether manual intervention is required based on the electricity billing settlement processing decisions for different anomaly types, and obtaining the determination result; when the determination result indicates that manual intervention is required, obtaining the information of the verification work orders to be dispatched, the real-time location and skill data of available maintenance personnel, and constructing a multi-objective work order dispatch optimization model to obtain the work order dispatch optimization model; solving the multi-objective work order dispatch optimization model using a genetic algorithm to obtain a set of work order dispatch schemes; and selecting the dispatch scheme with the highest overall satisfaction from the set of work order dispatch schemes using a fuzzy comprehensive evaluation method.

6. An electricity fee management device based on electricity meter data, characterized in that, include: The acquisition module is used to acquire meter image information, the time and space stamp information corresponding to the meter image, and the user's historical electricity consumption data; The first processing module is used to identify the image information of the electricity meter and obtain preliminary identification results; The second processing module is used to perform fusion processing based on the preliminary identification results, the user's historical electricity consumption data, and the spatiotemporal stamp information corresponding to the meter image to obtain a reliable reading report with confidence level and anomaly type label; The decision module is used to perform dynamic routing decision processing based on the confidence level and anomaly type according to the trusted reading report, and to obtain electricity billing processing decisions for different anomaly types. The management module is used to manage electricity fees according to the electricity fee settlement processing decisions based on the different anomaly types; wherein, the second processing module includes: a ninth processing unit, used to perform reasonableness verification processing based on the preliminary identification results and the user's historical electricity consumption data to obtain reasonableness verification results; a tenth processing unit, used to perform group electricity consumption pattern consistency analysis processing based on the spatiotemporal stamp information corresponding to the meter image and the meter model information to obtain group consistency analysis results, including: constructing proximity relationships in the spatiotemporal dimension based on the spatiotemporal stamp information, i.e., selecting meters within a radius of 500 meters in space and selecting readings within 1 hour before and after the collection time in time, to obtain a spatiotemporally associated meter sample set; based on the spatiotemporally associated meters The sample set calculates the electricity consumption change of each meter as a feature vector to obtain a sample dataset with electricity consumption features. Based on the sample dataset with electricity consumption features, an improved DBSCAN clustering process is used to obtain preliminary clustering results. Based on the preliminary clustering results, pattern recognition and outlier detection processing are performed to obtain the group electricity consumption consistency analysis results. The improved DBSCAN clustering process incorporates spatial distance constraints into density calculation. The eleventh processing unit is used to perform anomaly detection based on the meter image information and the corresponding spatiotemporal stamp information to obtain anomaly detection results. The twelfth processing unit is used to generate a reliable reading report based on the rationality verification results, the group consistency analysis results, and the anomaly detection results.

7. The electricity fee management device based on meter data according to claim 6, characterized in that, The acquisition module includes: an acquisition unit for acquiring data from multiple sensors built into the image acquisition device; a first processing unit for fusing the sensor data to obtain the spatial coordinates and attitude information of the image acquisition device; a second processing unit for performing attitude analysis based on the spatial coordinates and attitude information to obtain the shooting pose; a third processing unit for acquiring a meter image based on the shooting pose and evaluating the quality of the meter image to obtain a qualified meter image; and a fourth processing unit for binding the qualified meter image, the spatial coordinates of the image acquisition device, and the attitude information with spatiotemporal stamps and data integrity to obtain spatiotemporal stamp information corresponding to the meter image.

8. The electricity fee management device based on meter data according to claim 7, characterized in that, The second processing unit includes: a fifth processing unit, used to calculate the geometric relationship between the current shooting plane of the image acquisition device and the plane where the meter dial is located based on the posture data of the spatial coordinates and orientation of the image acquisition device, and to obtain the actual spatial angle between the normal vector of the current shooting plane and the normal vector of the plane where the meter dial is located; a sixth processing unit, used to compare the actual spatial angle with a preset angle threshold to obtain a pose adjustment guidance command; a seventh processing unit, used to obtain a guidance screen through real-time rendering and display processing of the pose adjustment guidance command via an augmented reality interface; and an eighth processing unit, used to adjust the device pose in real time based on the guidance screen to obtain the shooting pose.

Images