Power meter reading method, device and equipment based on key points and storage medium

Abstract

The invention relates to the technical field of electric power meter reading, in particular to an electric power meter reading method, device and equipment based on key points and a storage medium, and the electric power meter reading method based on the key points comprises the steps: obtaining a to-be-recognized dial plate image, and obtaining dial plate type information according to the to-be-recognized dial plate image; obtaining a corresponding meter key point detection model according to the dial plate type information, and inputting the to-be-identified dial plate image into the meter key point detection model to obtain dial plate key point information; calculating according to the dial key point information to obtain pointer angle data and dial scale coordinates; and performing calculation according to the pointer angle data and the dial scale coordinates, and outputting a dial reading result corresponding to the to-be-identified dial according to a calculation result. The method has the effect of improving the reading efficiency of the wattmeter.

Description

Technical Field

[0001] This application relates to the technical field of electricity meter readings, and in particular to a method, device, equipment, and storage medium for electricity meter readings based on key points. Background Art

[0002] Currently, the readings of equipment meters in substations usually need to be carried out manually, which not only consumes a large amount of human resources but also has low efficiency. With the continuous improvement of industrial automation levels, more and more enterprises are beginning to seek methods for automatic meter reading to reduce labor costs and improve work efficiency. The application of automatic meter reading technology can not only achieve real-time monitoring and data collection but also effectively improve the accuracy and reliability of data, thus better serving the management and maintenance of the power system.

[0003] Currently, in order to achieve automatic meter reading, the commonly used methods in the industry include but are not limited to the following: one is the traditional optical character recognition (OCR) technology, which takes pictures of the dial and uses the OCR algorithm to identify the digital information on the dial; the second is the pointer position recognition technology based on machine vision, which uses image processing technology to identify the pointer position on the dial and then calculates the reading; the third is the method based on template matching, which compares the currently captured dial image with various pre-stored dial templates to find the most matching template and thus determine the reading. Although these methods have achieved automatic meter reading to a certain extent, there are still some limitations in actual applications.

[0004] The main problems faced by the above existing automatic meter reading methods in actual applications are: it is difficult to ensure the accuracy and stability of meter reading under different lighting conditions, shooting angles, and complex background environments. For example, the traditional OCR technology is prone to misrecognition in case of insufficient light or complex background; the pointer position recognition technology based on machine vision is also difficult to accurately identify in case of pointer occlusion or deformation; the method based on template matching depends on the quantity and quality of the pre-stored templates, and once encountering a dial type that has not been seen before, the recognition effect will be greatly reduced. Therefore, how to achieve efficient and accurate electricity meter readings under various complex conditions has become an urgent technical problem to be solved. Summary of the Invention

[0005] In order to improve the efficiency of electricity meter readings, this application provides a method, device, equipment, and storage medium for electricity meter readings based on key points.

[0006] The above-mentioned first invention object of this application is achieved through the following technical solutions: A method for electricity meter readings based on key points, the method for electricity meter readings based on key points includes: Obtain an image of the dial to be recognized, and obtain the dial type information according to the image of the dial to be recognized; Obtain the corresponding meter key point detection model according to the dial type information, input the dial image to be recognized into the meter key point detection model, and obtain the dial key point information; Calculate according to the dial key point information to obtain the pointer angle data and the dial scale coordinates; Calculate according to the pointer angle data and the dial scale coordinates, and output the dial reading result corresponding to the dial to be recognized according to the calculation result.

[0007] By adopting the above technical solutions, the image of the dial to be recognized is collected by a camera. This process can not only cover different types of dials but also adapt to various environmental conditions, such as different lighting conditions, shooting angles, and background complexities. This step ensures that the system can obtain high-quality image data from the source, laying a solid foundation for subsequent processing; through the predefined reading inspection route and inspection progress information, the system can automatically identify the specific type of the dial to be recognized. This process avoids manual intervention and improves the automation degree of the system. Obtaining the dial type information not only helps to select the appropriate meter key point detection model but also provides necessary parameter support for subsequent key point detection and reading calculation; different dial types may require different processing methods. Therefore, the system selects the most suitable meter key point detection model according to the dial type information. For example, for a circular dial, a model based on YOLOv8 can be selected for positioning; for a non-standard shaped dial, other customized models may be required. This step ensures the accuracy and robustness of key point detection, thus improving the reliability and stability of the entire system; by inputting the image of the dial to be recognized into the selected meter key point detection model, the system can quickly and accurately identify the key point information of the dial. These key points include the start and end points of the pointer, scale marks, and other important feature points. This process relies on the high performance and high precision of the deep learning model, which can quickly lock the target area in a complex background and provide accurate data support for subsequent reading calculation; through further analysis and calculation of the dial key point information, the system can accurately calculate the angle data of the pointer and the coordinates of the dial scale. Specifically, the system will fit the straight line equation of the pointer according to the coordinate values of the key points and calculate the angle of the pointer relative to the center of the dial. At the same time, the system will also construct the scale curve data of the dial according to the coordinate values of the scale key points, providing a basis for the final reading calculation; finally, based on the calculated pointer angle data and dial scale coordinates, the system calculates the specific position of the pointer on the dial through mathematical models and algorithms and converts it into the corresponding reading result. This process takes into account the scale distribution characteristics of different dials, such as fan-shaped dials, circular dials, upper-notch-shaped, lower-notch-shaped, etc., as well as different reading schemes, such as clockwise reading and counterclockwise reading. In this way, the system can efficiently and accurately output the reading result of the dial to be recognized, meeting the requirements of different application scenarios.

[0008] In a preferred example of the present application, it can be further configured that: the obtaining of the image of the dial to be recognized and the obtaining of the dial type information according to the image of the dial to be recognized specifically include: Obtaining the reading inspection route and the inspection progress information corresponding to the image of the dial to be recognized; Obtaining the dial type information according to the inspection progress information and the reading inspection route.

[0009] By adopting the above technical solution, it is possible to efficiently and accurately obtain the type information of the dial to be recognized. Specifically, first, the reading inspection route and the inspection progress information corresponding to the dial to be recognized are obtained. The reading inspection route refers to the pre-set inspection path of the electric meters, which usually includes the positions of multiple meters and their arrangement order. The inspection progress information refers to the progress of the current inspection work, such as which meters have been read and which meter needs to be read next. According to the inspection progress information and the reading inspection route, the system can accurately determine the specific position and type of the dial to be recognized, not only improving the speed of dial recognition but also reducing the need for manual intervention, thus greatly improving work efficiency. Traditional meter reading methods often rely on manual inspections, which are not only time-consuming and laborious but also prone to omissions or errors. In contrast, the present invention ensures that each reading can be carried out at the correct time and location through an automated method, avoiding the influence of human factors. Further, by obtaining the reading inspection route and the inspection progress information, the system can predict in advance the type of the upcoming dial, and thus select an appropriate meter key point detection model for image processing. The system can continuously optimize the inspection route and progress arrangement based on historical data and machine learning algorithms, making the entire reading process more intelligent and efficient. For example, the system can automatically adjust the inspection route by analyzing past inspection records to avoid peak hours or maintenance areas, reducing unnecessary waiting time. At the same time, the system can also dynamically adjust the inspection progress according to the real-time feedback information to ensure that each reading is carried out in the best state.

[0010] In a preferred example, the present application can be further configured as follows: Before obtaining the corresponding meter key point detection model according to the dial type information and inputting the image of the dial to be recognized into the meter key point detection model to obtain the dial key point information, the key point-based electric meter reading method further includes: Obtaining the reading inspection route and the electric meter image data in sequence along the reading inspection route, and performing pointer marking and scale marking on each type of the electric meter image data; Performing category recognition on the electric meter image data, classifying the electric meter image data according to the recognition result, and obtaining a dial training set corresponding to each category; After extracting the pointer key point features and the scale key point information from the pointer marking and the scale marking in each of the dial training sets, training an initial model to obtain the meter key point detection model.

[0011] By adopting the above technical solutions, first, the reading inspection route and the image data of the electricity meters along this route in sequence are obtained, and pointer markings and scale markings are made for the image data of each type of electricity meter. This process ensures that the system can collect a large amount of labeled data, providing a basis for subsequent model training. Then, the category recognition is performed on these electricity meter image data, and they are classified according to the recognition results to obtain the dial training set corresponding to each category. This step enables different types of dial images to be processed separately, avoiding interference between different types and improving the accuracy and robustness of the model. Subsequently, the pointer key-point features and scale key-point information are extracted from each dial training set. The pointer key-point features include the starting point and the ending point of the pointer, while the scale key-point information covers each scale position on the dial and its corresponding value. The accurate extraction of these key-point information is crucial for subsequent model training because they directly affect the accuracy of the model for reading the dial. Next, the initial model is trained using the extracted key-point features and scale key-point information to obtain the meter key-point detection model. This meter key-point detection model has powerful detection capabilities and can quickly and accurately identify the position and bounding box of the dial in a complex background. Especially in different lighting conditions, shooting angles, and complex background environments, the model can still maintain high precision and stability, greatly improving the overall performance and reliability of the automatic meter reading system.

[0012] In a preferred example of the present application, it can be further configured that: calculating according to the dial key-point information to obtain the pointer angle data and the dial scale coordinates specifically includes: Obtaining at least two pointer key-points from the dial key-point information, and calculating the pointer angle data according to the pointer key-points; Obtaining scale key-points from the dial key-point information, and locating the dial scale coordinates according to the scale key-points.

[0013] By adopting the above technical solutions, the key-point-based electricity meter reading method can efficiently and accurately identify and calculate the pointer positions and scale coordinates on the dial, thus achieving accurate meter reading. Specifically: First, the system obtains the initial image of the dial to be recognized through a camera or other image acquisition devices. This process ensures the authenticity and accuracy of the original data, laying a foundation for subsequent key-point detection. Next, based on the obtained image of the dial to be recognized, the system automatically identifies the type information of the dial. This step is very important because different types of dials may have different structures and scale distributions, so it is necessary to select an appropriate key-point detection model for the electricity meter for subsequent processing. According to the dial type information, the system will select a pre-trained key-point detection model for the electricity meter, such as deep learning models like YOLOv8 or CenterNet. These models have been trained with a large number of samples, possess powerful detection capabilities and high precision, and can quickly and accurately locate the position and bounding box of the dial in a complex background. By inputting the image of the dial to be recognized into the selected key-point detection model for the electricity meter, the system can obtain the key-point information of the dial, including the starting and ending points of the pointer, scale marks, and other important feature points. From the key-point information of the dial, the system extracts at least two pointer key points, usually the starting point and the ending point of the pointer. The data of these two key points are crucial for calculating the angle of the pointer. At the same time, the system also extracts scale key points from the key-point information of the dial, that is, the positions of each scale on the dial. These scale key points are used to construct the scale curve data of the dial for subsequent calculation. Using the extracted pointer key points, the system calculates the actual angle of the pointer through geometric algorithms. A common calculation method is to first obtain the vector between the starting point and the ending point of the pointer, and then calculate the angle between this vector and the horizontal axis through trigonometric functions. This angle data reflects the true pointing of the pointer on the dial. Similarly, the system accurately locates the positions of each scale on the dial according to the extracted scale key points. These scale positions are usually represented in the form of pixel coordinates, which is convenient for subsequent numerical conversion and calculation. To improve the accuracy, the system may perform an average process on multiple scale points to reduce the influence of errors. Finally, based on the calculated pointer angle data and the scale coordinates of the dial, the system further calculates the specific reading of the dial.

[0014] In a preferred example of the present application, it can be further configured that: calculating according to the pointer angle data and the scale coordinates of the dial, and outputting the dial reading result corresponding to the dial to be recognized according to the calculation result, specifically including: Constructing scale curve data according to the scale coordinates of the dial, simulating the pointer angle data to the scale curve data, and obtaining the pointer coincidence position; Calculating according to the pointer coincidence position, and obtaining the dial reading result according to the calculation result.

[0015] By adopting the above technical solution, the key-point-based electricity meter reading method can achieve efficient and accurate dial reading. Specifically, the method first obtains the dial type information based on the dial image to be recognized, and then selects a suitable meter key-point detection model to detect and locate the key points of the dial. Next, by calculating the key-point information of the dial, the pointer angle data and the dial scale coordinates can be obtained. Finally, further calculations are performed using these data to output the accurate dial reading result. In the specific steps, after obtaining the pointer angle data and the dial scale coordinates, the system constructs one or more scale curve data according to the dial scale coordinates. These scale curve data reflect the relative position relationship between the scales on the dial. For example, for a circular dial, the scale curve may be a series of arc segments; for a fan-shaped or notched dial, the scale curve may be a combination of multiple broken lines. Subsequently, the pointer angle data is simulated into these scale curve data. This process involves mapping the actual angle of the pointer to the corresponding scale curve and finding the coincidence position of the pointer on the scale curve. To achieve this, the algorithm needs to consider the following factors: 1. **Pointer angle accuracy**: Since there may be slight errors in the movement of the pointer, a certain tolerance range needs to be introduced to ensure the accuracy of the pointer angle data. 2. **Scale interval**: The scale intervals of different types of dials may vary. For example, the scale spacing of some dials is larger, while that of others is smaller. The algorithm needs to configure the corresponding scale interval parameters according to different types of dials. 3. **Nonlinear influence**: The design of some dials may result in uneven scale distribution and nonlinear changes. At this time, the algorithm needs to have the ability to handle this nonlinear influence to ensure the one-to-one correspondence between the pointer angle and the scale position. Once the coincidence position of the pointer on the scale curve is found, the next step is to perform the final reading calculation based on this position. Specifically, the algorithm calculates the specific value pointed to by the pointer according to the coincidence position of the pointer and the scale values in its vicinity. It should be noted here that if the pointer is located between two scales, the final reading result needs to be determined according to the distance ratio of the pointer from these two scales. For example, assuming the pointer is closer to the left scale, the reading result will be closer to the value of the left scale. In addition, to improve the accuracy of the reading, some optimization measures can be introduced. For example, the dial images can be continuously collected at multiple time points, and the average value can be taken through multiple calculations to reduce the result deviation caused by single measurement errors. At the same time, combined with the historical reading records, the current reading is corrected to further improve the reliability and stability of the system.

[0016] The second invention object of the present application is achieved by the following technical solution: A key-point-based electricity meter reading device, the key-point-based electricity meter reading device includes: A type acquisition module, configured to acquire a dial image to be recognized, and acquire dial type information according to the dial image to be recognized; A key point recognition module, configured to acquire a corresponding meter key point detection model according to the dial type information, input the dial image to be recognized into the meter key point detection model, and acquire dial key point information; A scale coordinate calculation module, configured to perform calculations according to the dial key point information to obtain pointer angle data and dial scale coordinates; A meter reading module, configured to perform calculations according to the pointer angle data and the dial scale coordinates, and output a dial reading result corresponding to the dial image to be recognized according to the calculation result.

[0017] By adopting the above technical solution, an image of the dial to be recognized is collected by a camera. This process can not only cover different types of dials but also adapt to various environmental conditions, such as different lighting conditions, shooting angles, and background complexities. This step ensures that the system can obtain high-quality image data from the source, laying a solid foundation for subsequent processing. Through the predefined reading inspection route and inspection progress information, the system can automatically identify the specific type of the current dial to be recognized. This process avoids manual intervention and improves the automation level of the system. Obtaining the dial type information not only helps to select the appropriate meter key point detection model but also provides necessary parameter support for subsequent key point detection and reading calculation. Different dial types may require different processing methods. Therefore, the system selects the most suitable meter key point detection model according to the dial type information. For example, for a circular dial, a model based on YOLOv8 can be selected for positioning, while for a non-standard shaped dial, other customized models may be needed. This step ensures the accuracy and robustness of key point detection, thus improving the reliability and stability of the entire system. By inputting the image of the dial to be recognized into the selected meter key point detection model, the system can quickly and accurately identify the key point information of the dial. These key points include the start and end points of the pointer, scale marks, and other important feature points. This process relies on the high performance and high precision of the deep learning model, which can quickly lock the target area in a complex background and provide accurate data support for subsequent reading calculation. Through further analysis and calculation of the dial key point information, the system can accurately calculate the angle data of the pointer and the coordinates of the dial scale. Specifically, the system will fit the straight line equation of the pointer through geometric algorithms based on the coordinate values of the key points and calculate the angle of the pointer relative to the center of the dial. At the same time, the system will also construct the scale curve data of the dial based on the coordinate values of the scale key points, providing a basis for the final reading calculation. Finally, based on the calculated pointer angle data and dial scale coordinates, the system calculates the specific position of the pointer on the dial through mathematical models and algorithms and converts it into the corresponding reading result. This process takes into account the scale distribution characteristics of different dials, such as fan-shaped dials, circular dials, upper-notch dials, lower-notch dials, etc., as well as different reading schemes, such as clockwise reading and counterclockwise reading. In this way, the system can efficiently and accurately output the reading result of the dial to be recognized, meeting the requirements of different application scenarios.

[0018] The above object three of this application is achieved by the following technical solution: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above method for reading the power meter based on key points.

[0019] The above-mentioned fourth object of the present application is achieved by the following technical solutions: A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for reading power meter readings based on key points are implemented.

[0020] In summary, the present application includes at least one of the following beneficial technical effects: 1. Quickly and accurately identify the position and bounding box of the dial under complex backgrounds, improving the overall performance and reliability of the automatic meter reading system, and solving the recognition limitations of the prior art under complex backgrounds and different lighting conditions; 2. Accurately capture the starting and ending points of the pointer, scale marks, and other important feature points on the dial, ensuring efficient and accurate recognition of key information on the dial under different shooting angles and lighting conditions, and improving the accuracy of reading; 3. Flexibly adapt to different types of dials, and calculate the final reading based on the pointer angle and scale coordinates, achieving high precision and high reliability in automatic meter reading. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a flowchart of a method for reading power meter readings based on key points in an embodiment of the present application; Figure 2 is a flowchart for implementing step S10 in the method for reading power meter readings based on key points in an embodiment of the present application; Figure 3 is another flowchart for implementing the method for reading power meter readings based on key points in an embodiment of the present application; Figure 4 is a flowchart for implementing step S30 in the method for reading power meter readings based on key points in an embodiment of the present application; Figure 5 is a flowchart for implementing step S40 in the method for reading power meter readings based on key points in an embodiment of the present application; Figure 6 is a schematic block diagram of a system for reading power meter readings based on key points in an embodiment of the present application; Figure 7 is a schematic diagram of a device in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The present application will be further described in detail below with reference to the accompanying drawings.

[0023] In one embodiment, as Figure 1 shown, the present application discloses a method for reading power meter readings based on key points, which specifically includes the following steps: S10: Obtain the dial image to be recognized, and obtain the dial type information according to the dial image to be recognized.

[0024] In this embodiment, the dial image to be recognized refers to the image that needs to automatically detect and recognize the readings in the pointer-type electric meter. The dial type information refers to the data of the specific type of this electric meter.

[0025] Specifically, when reading the electric meter for a fixed area, by controlling a tracing cart for reading the electric meter or a device such as an unmanned aerial vehicle (UAV) remote sensing that has remote control or automatic driving and is equipped with a shooting function, conduct inspections in this area according to the position of the pointer-type electric meter. During the inspection process, control the camera device of the device to face the electric meter and take pictures, so as to obtain the dial image to be recognized.

[0026] Furthermore, after obtaining the dial image to be recognized, recognize the dial image to be recognized, so as to recognize the specific type of the dial as the dial type information. Among them, the dial type information includes the appearance of the dial, the pointer style, and the pointer rotation direction, etc.

[0027] S20: Obtain the corresponding meter key point detection model according to the dial type information, input the dial image to be recognized into the meter key point detection model, and obtain the dial key point information.

[0028] In this embodiment, the pointer positioning model refers to a model that has been pre-trained and is used to locate the dial, the scales on the dial, and the position of the pointer in the image.

[0029] Specifically, obtain the corresponding meter key point detection model through the dial type information, and input the dial image to be recognized into the marker positioning model for recognition, so as to recognize the key points of the scales on the dial and the key point information of the pointer as the dial key point information.

[0030] S30: Calculate according to the dial key point information to obtain the pointer angle data and the dial scale coordinates.

[0031] In this embodiment, the dial scale coordinates refer to the position of the scales on the dial represented in the form of coordinates in the dial image to be recognized.

[0032] Specifically, after obtaining the key point information of the dial, the key point information of the pointer and the key point information of the dial scale are obtained from the key point information of the dial. Further, according to the dial type information, the scale arc and the rotation arc of the pointer of this type of dial are obtained. According to the key point information of the dial and the scale arc of the dial, the position of the scale of the dial in the image to be recognized is recognized from the image of the dial to be recognized and is marked in the form of coordinates, so as to obtain the dial scale coordinates. And, according to the recognized key points of the pointer, the position of the pointer in the image is recognized, and the pointer angle data is calculated according to this position.

[0033] S40: Calculate according to the pointer angle data and the dial scale coordinates, and output the dial reading result corresponding to the dial to be recognized according to the calculation result.

[0034] Specifically, according to the pointer angle data, the relative position between the current pointer and the scale of the dial is calculated, and according to this relative position and the dial scale coordinates, the position coordinates of this relative position in the coordinate system where the dial scale coordinates are located are calculated, and then the dial reading result is obtained.

[0035] In this embodiment, the image of the dial to be recognized is collected by a camera. This process can not only cover different types of dials but also adapt to various environmental conditions, such as different lighting conditions, shooting angles, and background complexities. This step ensures that the system can obtain high-quality image data from the source, laying a solid foundation for subsequent processing. Through the predefined reading inspection route and inspection progress information, the system can automatically identify the specific type of the dial to be recognized currently. This process avoids manual intervention and improves the automation level of the system. Obtaining the dial type information not only helps to select the appropriate meter key point detection model but also provides necessary parameter support for subsequent key point detection and reading calculation. Different dial types may require different processing methods. Therefore, the system selects the most suitable meter key point detection model according to the dial type information. For example, for a circular dial, a model based on YOLOv8 can be selected for positioning, while for a non-standard shaped dial, other customized models may be needed. This step ensures the accuracy and robustness of key point detection, thus improving the reliability and stability of the entire system. By inputting the image of the dial to be recognized into the selected meter key point detection model, the system can quickly and accurately identify the key point information of the dial. These key points include the start and end points of the pointer, scale marks, and other important feature points. This process relies on the high performance and high precision of the deep learning model, which can quickly lock the target area in a complex background and provide accurate data support for subsequent reading calculation. Through further analysis and calculation of the dial key point information, the system can accurately calculate the angle data of the pointer and the coordinates of the dial scale. Specifically, the system will fit the straight line equation of the pointer according to the coordinate values of the key points and calculate the angle of the pointer relative to the center of the dial. At the same time, the system will also construct the scale curve data of the dial according to the coordinate values of the scale key points, providing a basis for the final reading calculation. Finally, based on the calculated pointer angle data and dial scale coordinates, the system calculates the specific position of the pointer on the dial through mathematical models and algorithms and converts it into the corresponding reading result. This process takes into account the scale distribution characteristics of different dials, such as fan-shaped dials, circular dials, upper-notch dials, lower-notch dials, etc., as well as different reading schemes, such as clockwise reading and counterclockwise reading. In this way, the system can efficiently and accurately output the reading result of the dial to be recognized, meeting the requirements of different application scenarios.

[0036] In one embodiment, as Figure 2 shown, in step S10, that is, obtaining the image of the dial to be recognized and obtaining the dial type information according to the image of the dial to be recognized, specifically including: S11: Obtain the reading inspection route and the inspection progress information corresponding to the image of the dial to be recognized.

[0037] In this embodiment, the reading inspection route refers to the formal route when using a reading inspection device equipped with a camera device to conduct inspection readings.

[0038] Specifically, according to the installation positions of each pointer-type electricity meter, a route for the inspection device to conduct inspections is pre-planned as the reading inspection route, so as to control the inspection device to conduct inspections along this inspection route.

[0039] Further, after obtaining the reading inspection route of the inspection device, according to the inspection task corresponding to the set reading inspection route, the inspection progress information is obtained according to this inspection task.

[0040] S12: Obtain the dial type information according to the inspection progress information and the reading inspection route.

[0041] Specifically, according to this inspection progress information, judge the dial type corresponding to the specific dial image to be recognized at the current progress as the dial type information.

[0042] Optionally, when obtaining the dial type information, the dial image to be recognized can be combined and input into a preset target detection model for detection. According to the detection result and the inspection progress, the dial type information is output. If the detection result of the target detection model is inconsistent with the result judged by the inspection progress information, the result detected by the target detection model is used as the dial type information.

[0043] In one embodiment, as Figure 3 shown, before step S20, the key-point electricity meter reading method further includes: S201: Obtain the reading inspection route and the electricity meter image data in sequence along the reading inspection route, and perform pointer marking and scale marking on each type of electricity meter image data.

[0044] Specifically, in the manner of step S11, obtain the pre-determined reading inspection route of this area. Further, in the order of this reading inspection route, take images of each pointer-type electricity meter as the electricity meter image data (such as electricity meters, water meters, and gas meters, etc.). After obtaining the electricity meter image data, mark the key points of the pointer in the image, such as the starting end and the ending end of the pointer, and the scale points, to obtain the pointer marking and the scale marking.

[0045] S202: Perform category recognition on the electricity meter image data, classify the electricity meter image data according to the recognition result, and obtain the corresponding dial training set for each category.

[0046] Specifically, classify the circuit meter image data of the same type with pointer marks and scale marks. First, classify according to the functions of the power meters. Further, perform image recognition for each category, and classify the images of power meters with the same model or style into one category as the corresponding dial training set.

[0047] S203: After extracting the pointer key point features and scale key point information from the pointer marks and scale marks in each dial training set, train the initial model to obtain a meter key point detection model.

[0048] Specifically, extract the marks in the images of each dial training set as pointer key point information and scale key point information, and use the YOLOv8 (You Only Look Once version 8) deep learning model for training. Combining with the CenterNet deep learning framework, through real-time detection and key point positioning of the dials in the images, it is possible to accurately capture the starting and ending points of the pointers, scale marks, and other important feature points on the dials, and obtain a meter key point detection model.

[0049] In one embodiment, as Figure 4 shown, in step S30, that is, calculate according to the dial key point information to obtain the pointer angle data and the dial scale coordinates, specifically including: S31: Obtain at least two pointer key points from the dial key point information, and calculate the pointer angle data according to the pointer key points.

[0050] Specifically, when obtaining the pointer angle data, obtain at least two pointer key points from the image of the pointer, including the pivot point and the tip of the pointer, and simulate the straight line and position where the pointer is located by the method of two points forming a line.

[0051] Further, according to the rotation arc at the end of the pointer of this type of power meter, obtain the position of the pointer in this rotation arc, so as to calculate the pointer angle data.

[0052] S32: Obtain the scale key points from the dial key point information, and locate the dial scale coordinates according to the scale key points.

[0053] Specifically, from the image of the dial to be recognized, infer the picture through the meter key point detection model to obtain the coordinate values of the meter in the picture. According to the coordinate values, the camera will move and magnify the magnification to re-take the picture of the meter. After the meter key point detection model infers the new picture and intercepts the dial, obtain the recognized scale key points from the dial key points, and obtain the dial scale coordinates according to the coordinate values of the meter in the picture.

[0054] In one embodiment, asFigure 5 As shown, in step S40, that is, calculating according to the pointer angle data and the dial scale coordinates, and outputting the dial reading result corresponding to the dial to be recognized according to the calculation result, specifically including: S41: Construct scale curve data according to the dial scale coordinates, simulate the pointer angle data to the scale curve data, and obtain the pointer coincidence position.

[0055] Specifically, first obtain the sizes of the corresponding digital scale points and the corresponding reading schemes from the configuration file according to the detected dial type information. The reading schemes can be divided into fan-shaped dials, circular dials, upper-notch-shaped, lower-notch-shaped and other reading schemes according to the dial shape, and can be divided into clockwise reading schemes and counterclockwise reading schemes according to the scale size order, so as to construct the scale curve data. Calculate according to the configuration and obtain the key point coordinates, and simulate the pointer angle data into the scale curve data, so as to obtain the coincidence position of the pointer and the curve where the scale is located, that is, the pointer coincidence position.

[0056] S42: Calculate according to the pointer coincidence position, and obtain the dial reading result according to the calculation result.

[0057] Specifically, the angle calculation scheme can be obtained according to the configuration attributes corresponding to the meter type, and then according to the obtained coordinate values and the configured scale values, the angle size corresponding to each scale value can be obtained; and the pointer straight line is fitted according to the key points on the pointer to obtain the angle corresponding to the pointer. According to these angles, according to the proportion of the pointer angle size in the middle of two adjacent scales, the reading can be calculated according to the proportion size and the scale value size, that is, the dial reading result.

[0058] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0059] In one embodiment, a key-point-based electric meter reading device is provided, and the key-point-based electric meter reading device corresponds one-to-one with the key-point-based electric meter reading method in the above embodiment. As Figure 6 shown, the key-point-based electric meter reading device includes a type acquisition module, a key point recognition module, a scale coordinate calculation module and a meter reading module. The detailed descriptions of each functional module are as follows: The type acquisition module is used to acquire the image of the dial to be recognized, and acquire the dial type information according to the image of the dial to be recognized; The key point recognition module is used to obtain the corresponding meter key point detection model according to the dial type information, input the image of the dial to be recognized into the meter key point detection model, and obtain the dial key point information; A scale coordinate calculation module, which is used to calculate according to the key point information of the dial to obtain pointer angle data and dial scale coordinates; A meter reading module, which is used to calculate according to the pointer angle data and the dial scale coordinates, and output the dial reading result corresponding to the dial to be recognized according to the calculation result.

[0060] Optionally, the type acquisition module includes: An inspection progress acquisition module, which is used to acquire the inspection progress information corresponding to the reading inspection route and the image of the dial to be recognized; A dial type acquisition sub-module, which is used to acquire the dial type information according to the inspection progress information and the reading inspection route.

[0061] Optionally, the key-point-based electric meter reading device further includes: An image marking module, which is used to acquire the reading inspection route and the electric meter image data in the order along the reading inspection route, and perform pointer marking and scale marking on each type of electric meter image data; An image classification module, which is used to perform category recognition on the electric meter image data, classify the electric meter image data according to the recognition result, and obtain the corresponding dial training set for each category; A model training module, which is used to extract the pointer key-point features and scale key-point information from the pointer marking and scale marking in each dial training set, and then train the initial model to obtain a meter key-point detection model.

[0062] Optionally, the scale coordinate calculation module includes: A pointer angle calculation sub-module, which is used to obtain at least two pointer key points from the dial key point information, and calculate the pointer angle data according to the pointer key points; A dial coordinate calculation sub-module, which is used to obtain the scale key points from the dial key point information, and locate the dial scale coordinates according to the scale key points.

[0063] Optionally, the meter reading module includes: A coincidence position acquisition sub-module, which is used to construct scale curve data according to the dial scale coordinates, simulate the pointer angle data to the scale curve data, and obtain the pointer coincidence position; A dial reading sub-module, which is used to calculate according to the pointer coincidence position, and obtain the dial reading result according to the calculation result.

[0064] For the specific limitations of the power meter reading device based on key points, reference can be made to the limitations of the power meter reading method based on key points in the above text, which will not be elaborated here. Each module in the above power meter reading device based on key points can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0065] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 7 shown. The computer device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a power meter reading method based on key points.

[0066] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: Obtain the dial image to be recognized, and obtain the dial type information according to the dial image to be recognized; Obtain the corresponding meter key point detection model according to the dial type information, input the dial image to be recognized into the meter key point detection model, and obtain the dial key point information; Calculate according to the dial key point information to obtain the pointer angle data and the dial scale coordinates; Calculate according to the pointer angle data and the dial scale coordinates, and output the dial reading result corresponding to the dial to be recognized according to the calculation result.

[0067] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the following steps are implemented: Obtain the dial image to be recognized, and obtain the dial type information according to the dial image to be recognized; Obtain the corresponding meter key point detection model according to the dial type information, input the dial image to be recognized into the meter key point detection model, and obtain the dial key point information; Calculate according to the dial key point information to obtain the pointer angle data and the dial scale coordinates; Calculate based on the pointer angle data and the dial scale coordinates, and output the dial reading result corresponding to the dial to be recognized according to the calculation result.

[0068] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0069] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0070] The above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for reading electricity meter readings based on key points, characterized in that The method for reading the power meter based on key points includes: Obtain the dial image to be recognized, and obtain the dial type information according to the dial image to be recognized; Obtain the corresponding meter key point detection model according to the dial type information, input the dial image to be recognized into the meter key point detection model, and obtain the dial key point information; Calculate according to the dial key point information to obtain the pointer angle data and the dial scale coordinates; Calculate according to the pointer angle data and the dial scale coordinates, and output the dial reading result corresponding to the dial to be recognized according to the calculation result.

2. The method for reading power meter readings based on key points according to claim 1, wherein The obtaining of the dial image to be recognized and the obtaining of the dial type information according to the dial image to be recognized specifically include: Obtain the reading inspection route and the inspection progress information corresponding to the dial image to be recognized; Obtain the dial type information according to the inspection progress information and the reading inspection route.

3. The method for reading the power meter reading based on key points according to claim 1, wherein Before obtaining the corresponding meter key point detection model according to the dial type information, inputting the dial image to be recognized into the meter key point detection model, and obtaining the dial key point information, the method for reading the power meter based on key points further includes: Obtain the reading inspection route and the power meter image data in sequence along the reading inspection route, and perform pointer marking and scale marking on each type of power meter image data; Perform category recognition on the power meter image data, classify the power meter image data according to the recognition result, and obtain the corresponding dial training set for each category; After extracting the pointer key point features and the scale key point information from the pointer marking and the scale marking in each dial training set, train the initial model to obtain the meter key point detection model.

4. The method for reading power meter readings based on key points according to claim 3, characterized in that, The calculating according to the dial key point information to obtain the pointer angle data and the dial scale coordinates specifically includes: Obtain at least two pointer key points from the dial key point information, and calculate the pointer angle data according to the pointer key points; Obtain the scale key points from the dial key point information, and locate the dial scale coordinates according to the scale key points.

5. The method for reading the power meter reading based on key points according to claim 4, wherein The calculating according to the pointer angle data and the dial scale coordinates, and outputting the dial reading result corresponding to the dial to be recognized according to the calculation result specifically includes: Construct scale curve data according to the dial scale coordinates, simulate the pointer angle data to the scale curve data, and obtain the pointer coincidence position; Calculate according to the pointer coincidence position, and obtain the dial reading result according to the calculation result.

6. A key-point-based electric meter reading device, characterized in that The device for reading the power meter based on key points includes: A type acquisition module, configured to obtain a dial image to be recognized, and obtain dial type information according to the dial image to be recognized; A key point recognition module, configured to obtain the corresponding meter key point detection model according to the dial type information, input the dial image to be recognized into the meter key point detection model, and obtain dial key point information; A scale coordinate calculation module, configured to calculate according to the dial key point information to obtain pointer angle data and dial scale coordinates; The meter reading module is used to calculate according to the pointer angle data and the dial scale coordinates, and output the dial reading result corresponding to the dial to be recognized according to the calculation result.

7. The power meter reading device based on key points according to claim 6, wherein The type acquisition module includes: The inspection progress acquisition module is used to acquire the reading inspection route and the inspection progress information corresponding to the dial image to be recognized; The dial type acquisition sub-module is used to acquire the dial type information according to the inspection progress information and the reading inspection route.

8. The power meter reading device based on key points according to claim 6, wherein, The key-point-based electric meter reading device further includes: The image marking module is used to acquire the reading inspection route and the electric meter image data in sequence along the reading inspection route, and perform pointer marking and scale marking on each type of the electric meter image data; The image classification module is used to perform category recognition on the electric meter image data, classify the electric meter image data according to the recognition result, and obtain the corresponding dial training set for each category; The model training module is used to train the initial model after extracting the pointer key-point features and scale key-point information from the pointer marking and the scale marking in each of the dial training sets, so as to obtain the meter key-point detection model.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the key-point-based electric meter reading method according to any one of claims 1 to 5.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the key-point-based electric meter reading method according to any one of claims 1 to 5.

Images