Meter identification method, device and equipment

Abstract

The invention provides a meter identification method, device and equipment, and relates to the technical field of substation meter identification. The method comprises the following steps: acquiring a plurality of dial images of a to-be-identified meter; performing character extraction on each dial image to obtain a character extraction result; wherein the character extraction result comprises a meter category, a measuring range and a scale range, and the scale range comprises a minimum scale, a maximum scale and positions of the minimum scale and the maximum scale; performing scale contour identification on each dial plate image to obtain a scale identification result, and calculating a numerical value and a unit represented by each degree according to the corresponding scale range and the scale identification result; performing pointer identification on each dial plate image to obtain a pointer identification result; the character extraction result, the scale identification result, the numerical value and unit represented by each degree and the pointer identification result are packaged with the corresponding dial images and then classified and stored; and calculating the pointer reading according to the corresponding numerical value and unit represented by each degree and the pointer identification result. According to the invention, the recognition efficiency can be improved when multiple images are processed.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer substation meter identification, and in particular to a meter identification method, device and equipment. Background Art

[0002] Meters are crucial equipment in substation operation monitoring, indicating key parameters such as current, voltage, and power. They are crucial for ensuring the safe and stable operation of substations and power grids. Traditionally, meter readings rely on manual inspections, requiring personnel to visit sites and review records one by one. This approach is inefficient, prone to errors, and lacks real-time performance, making it difficult to meet the demands of automated and intelligent operation and management in modern substations.

[0003] Image recognition technology has made significant progress in recent years, making it possible to automatically identify meters. Traditional image recognition methods typically use a sequential processing approach, processing only one image at a time. This approach is inefficient when processing multiple images and fails to fully utilize computing resources. For example, traditional OpenCV-based image recognition methods typically use a loop structure to process each image sequentially, resulting in a long processing time when processing large numbers of images. Summary of the Invention

[0004] Embodiments of the present invention provide a meter recognition method, apparatus, and device to solve the problem of low efficiency in recognizing multiple meter images.

[0005] In a first aspect, an embodiment of the present invention provides a meter identification method, comprising:

[0006] Acquire multiple dial images of meters to be identified;

[0007] Performing text extraction on each dial image to obtain a text extraction result; wherein the text extraction result includes meter type, range, and scale range, and the scale range includes the minimum scale and the maximum scale and their positions;

[0008] Performing scale contour recognition on each dial image to obtain a scale recognition result, and calculating the value and unit represented by each degree based on the corresponding scale range and the scale recognition result; wherein the scale recognition result includes the angular range of the scale contour;

[0009] Perform pointer recognition on each dial image to obtain a pointer recognition result; calculate the pointer reading of the meter based on the corresponding numerical value and unit represented by each degree and the pointer recognition result; wherein the pointer recognition result includes the deflection angle of the pointer, and the text extraction result, the scale recognition result, the numerical value and unit represented by each degree, and the pointer recognition result are respectively packaged with the corresponding dial image, and the packaged data are classified and stored according to the text extraction result, the scale recognition result, the numerical value and unit represented by each degree, and the pointer recognition result.

[0010] In a possible implementation, before calculating the value and unit represented by each degree according to the corresponding scale range and the scale recognition result, the following steps are included:

[0011] Get the scale range and scale recognition results;

[0012] Based on the dial image pixels, the scale range and scale recognition results are matched and verified to obtain the corresponding scale range and scale recognition results;

[0013] Before calculating the meter pointer reading according to the corresponding value per degree and unit and pointer identification results, including:

[0014] Obtain the value and unit of each degree in the scale recognition result and the pointer recognition result;

[0015] Based on the dial image pixels, the numerical value represented by each degree, the unit and the pointer recognition result are matched and checked to obtain the corresponding numerical value represented by each degree, the unit and the pointer recognition result.

[0016] In a possible implementation, performing pointer recognition on each dial image to obtain a pointer recognition result includes:

[0017] Perform edge detection on each dial image;

[0018] Extracting the pointer outline from the dial image after edge detection; wherein the pointer outline includes the position of the pointer outline;

[0019] The deflection angle of the pointer is determined based on the extracted pointer profile and the minimum scale.

[0020] In one possible implementation, calculating the pointer reading of the meter according to the corresponding value and unit represented by each degree and the pointer identification result includes:

[0021] Calculate the pointer reading of the meter based on the corresponding pointer deflection angle, the value represented per degree, and the unit.

[0022] In a possible implementation, text extraction is performed on each dial image to obtain a text extraction result, including:

[0023] Perform edge detection on each dial image and perform binarization;

[0024] Perform connected domain analysis on the binarized dial image and perform segmentation;

[0025] Perform text recognition on the segmented dial image and extract text features; the text features include the position, size and shape of the text;

[0026] Calculate the similarity between text features and characters in the template library, and determine the text extraction result based on the similarity and similarity threshold.

[0027] In a possible implementation, performing scale contour recognition on each dial image to obtain a scale recognition result includes:

[0028] Perform edge detection on each dial image;

[0029] Extract scale contours from the dial image after edge detection;

[0030] The angle between the minimum scale and the maximum scale of the dial is calculated according to the extracted scale profile to obtain the angle range of the scale profile.

[0031] In one possible implementation, the numerical value and unit represented by each degree are calculated according to the corresponding scale range and the scale recognition result, including:

[0032] The value represented by each degree is calculated based on the corresponding scale range and the angular range of the scale profile;

[0033] The unit of the value represented per degree is determined according to the unit of the range.

[0034] In a possible implementation, after calculating the pointer reading of the meter according to the corresponding value and unit represented by each degree and the pointer identification result, the method further includes:

[0035] Determine the intermediate scale and its position according to the positions of the minimum scale and the maximum scale;

[0036] The rotation angle of the pointer is calculated based on the intersection of the minimum scale line and the pointer and the circle. A coordinate system is established using the intersection of the minimum scale and the maximum scale and the middle scale, with the origin as the intersection point and the vertical axis as the line between the middle scale and the origin. A circle of a preset radius is constructed with the origin as the center.

[0037] The pointer recognition result is verified based on the pointer recognition result, the pointer rotation angle and the difference threshold.

[0038] In one possible implementation, obtaining multiple dial images of meters to be identified includes:

[0039] Obtain multiple substation monitoring images to be identified;

[0040] Grayscale processing is performed on the monitoring image of the substation to be identified;

[0041] Filtering the grayscale substation monitoring image to be identified to obtain a preprocessed image;

[0042] The pre-processed image is segmented using a threshold to obtain the initial image of the dial;

[0043] The initial dial image is enhanced to obtain the dial image.

[0044] In one possible implementation, the enhancement includes one or more of rotation, translation, scaling, and flipping.

[0045] In a second aspect, an embodiment of the present invention provides a meter identification device, including:

[0046] An image acquisition module, used to acquire multiple dial images of meters to be identified;

[0047] a text extraction module, configured to extract text from each dial image to obtain a text extraction result; wherein the text extraction result includes the meter type, range, and scale range, and the scale range includes the minimum scale and the maximum scale and their positions;

[0048] a calculation module for performing scale contour recognition on each dial image to obtain a scale recognition result, and calculating the value and unit represented by each degree based on the corresponding scale range and the scale recognition result; wherein the scale recognition result includes the angular range of the scale contour;

[0049] The pointer reading module is used to perform pointer recognition on each dial image to obtain a pointer recognition result; calculate the pointer reading based on the corresponding numerical value and unit represented by each degree and the pointer recognition result; wherein the pointer recognition result includes the deflection angle of the pointer, and the text extraction result, scale recognition result, numerical value and unit represented by each degree and pointer recognition result are respectively packaged with the corresponding dial image, and the packaged data are classified and stored according to the text extraction result, scale recognition result, numerical value and unit represented by each degree and pointer recognition result.

[0050] In a third aspect, an embodiment of the present invention provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method in the first aspect or any possible implementation of the first aspect is implemented.

[0051] In an embodiment of the present invention, by processing text extraction tasks for multiple images in parallel, the efficiency of text extraction can be improved. Similarly, processing scale recognition and calculation tasks for multiple images in parallel can increase the speed of scale recognition and calculation. In addition, processing pointer recognition and reading calculation tasks for multiple images in parallel can also improve the efficiency of pointer recognition and reading calculation. Compared with traditional methods, this method can process multiple images simultaneously, fully mobilize and utilize computing resources, reduce processing time, and avoid the problem of idle resources caused by processing images one by one in traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1This is a flow chart of an implementation of a meter identification method provided by an embodiment of the present invention;

[0053] Figure 2 1 is a schematic structural diagram of a meter identification device provided by an embodiment of the present invention;

[0054] Figure 3 is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0055] It should be noted that the present invention identifies single-pointer meters such as voltmeters and ammeters.

[0056] See also Figure 1 , which shows a flow chart for implementing the meter identification method provided by an embodiment of the present invention, and is described in detail as follows:

[0057] Step 101: Acquire multiple dial images of meters to be identified.

[0058] In this embodiment, image information of the dial parts of multiple meters to be identified is obtained from an actual scene through specific equipment or tools, such as a camera, an image acquisition card, etc. These images will be used as objects for subsequent identification processing.

[0059] Step 102: extract text from each dial image to obtain a text extraction result; wherein the text extraction result includes the meter type, range and scale range, and the scale range includes the minimum scale and the maximum scale and their positions.

[0060] In this embodiment, a text extraction operation is performed on each enhanced meter dial image to identify text information within the image, including: meter type (e.g., voltmeter, ammeter, or other meter types); range (e.g., 0-100 amperes or 0-250 volts); and scale range (e.g., minimum and maximum scale marks and their locations on the dial).

[0061] Step 103: Perform scale contour recognition on each dial image to obtain a scale recognition result, and calculate the value and unit represented by each degree based on the corresponding scale range and the scale recognition result; wherein the scale recognition result includes the angular range of the scale contour.

[0062] In this embodiment, a scale outline recognition operation is performed on each dial image to identify the outline information of the scale on the dial. The scale recognition result includes the angular range of the scale outline, such as the starting angle and ending angle of the scale on the dial.

[0063] Then, based on the corresponding scale range (such as the minimum and maximum scale values) and the angular range of the identified scale outline, the specific value and unit represented by each degree is calculated (for example, each degree represents 1 ampere or each degree represents 10 volts). This step identifies the dial scale outline and calculates the value and unit represented by each degree.

[0064] Step 104: Perform pointer recognition on each dial image to obtain a pointer recognition result; calculate the pointer reading based on the corresponding numerical value and unit represented by each degree and the pointer recognition result; wherein the pointer recognition result includes the deflection angle of the pointer, and the text extraction result, scale recognition result, numerical value and unit represented by each degree, and pointer recognition result are respectively packaged with the corresponding dial image, and the packaged data are classified and stored according to the text extraction result, scale recognition result, numerical value and unit represented by each degree, and pointer recognition result.

[0065] In this embodiment, a pointer recognition operation is performed on each dial image to identify the deflection angle of the pointer on the dial. The specific reading indicated by the meter pointer is then calculated by combining the calculated value and unit per degree (for example, 1 ampere per degree or 10 volts per degree) with the corresponding pointer deflection angle. This step identifies the pointer deflection angle, and the specific meter reading is calculated based on the value information per degree.

[0066] In the meter recognition method of the present invention, text extraction results, scale recognition results, the numerical value and unit represented by each degree, and pointer recognition results are obtained separately. These different types of result data are associated with the corresponding dial image to form complete data packets. This organizational approach ensures that all relevant information for each dial image can be fully associated and traced during subsequent processing. For example, all text extraction results and their corresponding dial images are stored in a data packet; all scale recognition results and their corresponding dial images are stored in a data packet; all numerical values ​​and units represented by each degree and their corresponding dial images are stored in a data packet; and all pointer recognition results and their corresponding different dial images are stored in a data packet. The data in the data packet is stored sequentially in a stack, with the stacking and popping rules being first-in, first-out.

[0067] In an embodiment of the present invention, by processing text extraction tasks for multiple images in parallel, the efficiency of text extraction can be improved. Similarly, processing scale recognition and calculation tasks for multiple images in parallel can increase the speed of scale recognition and calculation. In addition, processing pointer recognition and reading calculation tasks for multiple images in parallel can also improve the efficiency of pointer recognition and reading calculation. Compared with traditional methods, this method can process multiple images simultaneously, fully mobilize and utilize computing resources, reduce processing time, and avoid the problem of idle resources caused by processing images one by one in traditional methods.

[0068] In a possible implementation, before calculating the value and unit represented by each degree according to the corresponding scale range and the scale recognition result, the following steps are included:

[0069] Get the scale range and scale recognition results;

[0070] Based on the dial image pixels, the scale range and scale recognition results are matched and verified to obtain the corresponding scale range and scale recognition results;

[0071] Before calculating the meter pointer reading according to the corresponding value per degree and unit and pointer identification results, including:

[0072] Obtain the value and unit of each degree in the scale recognition result and the pointer recognition result;

[0073] Based on the dial image pixels, the numerical value represented by each degree, the unit and the pointer recognition result are matched and checked to obtain the corresponding numerical value represented by each degree, the unit and the pointer recognition result.

[0074] Specifically, first, obtain the first scale range and its corresponding dial image, and obtain the first scale recognition result and its corresponding dial image, match them according to the pixels of the corresponding dial images, and determine whether the two dial images are the same image. When the two dial images are the same image, determine that the set of scale ranges and the scale recognition results correspond, and calculate the value and unit represented by each degree based on the set of scale ranges and the scale recognition results.

[0075] When the judgment result is that the two dial images are not the same image, it is determined that the group of scale ranges does not correspond to the scale recognition result, the first scale range is stored in the spare space (and the dial images corresponding to the first scale range are labeled in sequence), and the next scale range and its corresponding dial image are re-acquired, and the first scale recognition result and the dial image corresponding to the next scale range are matched according to the pixels of the dial image until it is determined that the two dial images are the same, and then the group of scale ranges is determined to correspond to the scale recognition result, and then the value and unit represented by each degree are calculated according to the scale range and the scale recognition result.

[0076] Among them, the priority of the spare space is higher than the priority of the results of the classified storage. This priority design can optimize the matching process, reduce repeated matching, and improve matching efficiency. When the previous scale range does not match the scale recognition result, first obtain the scale range of the spare space and its corresponding dial image in turn, and match the scale range of the spare space with the current scale recognition result. If the match is successful, perform subsequent calculations and delete the scale range and its dial image from the spare space; if the match fails, the scale range and its corresponding dial image will continue to be stored in the spare space. When all the images corresponding to the scale ranges in the spare space fail to match the dial image corresponding to the current scale recognition, obtain the next scale range and its corresponding dial image from the classified storage results.

[0077] Exemplarily, the method for matching and checking the numerical value and unit represented by each degree with the pointer identification result is the same as the above principle and will not be repeated here.

[0078] In this embodiment, pixels in the dial image corresponding to the scale range and the scale recognition result are compared to determine whether they are the same image. This pixel-based matching method ensures consistency between the recognition result and the original image, preventing mismatches caused by errors in image processing. If the two dial images match successfully, the scale range and the scale recognition result correspond, and the numerical value and unit represented by each degree can be calculated based on this information. If the match fails, the scale range is stored in a backup space, and the next scale range is retrieved for matching. This "backup space" design ensures that all possible matches are attempted, preventing the failure of the entire recognition process. Furthermore, the backup space has a higher priority than the results stored in the categorized storage, further optimizing the matching process and ensuring both efficiency and accuracy. The backup space is designed to handle match failures, ensuring that all possible matches are attempted. When a scale range in the backup space successfully matches the current scale recognition result, subsequent calculations are performed, and the successfully matched data is deleted from the backup space. If the match fails, the data is stored in the backup space. This backup space design can handle match failures, preventing the entire recognition process from failing due to individual match failures. By giving priority to the use of spare space, the number of repeated matches is reduced and the matching efficiency is improved.

[0079] The present invention provides accurate and reliable correspondence data for the parallel processing of text extraction, scale recognition, and pointer recognition tasks across multiple images through accurate matching and timely correction. This ensures that each task is based on correct information during parallel processing, fully leveraging the parallel advantages of computing resources and improving overall processing speed. Clear correspondence and timely matching verification avoid subsequent processing of incorrectly matched data, reducing unnecessary calculations and resource usage, focusing computing resources on valid data, and improving resource utilization efficiency.

[0080] In a possible implementation, performing pointer recognition on each dial image to obtain a pointer recognition result includes:

[0081] Perform edge detection on each dial image;

[0082] Extracting the pointer outline from the dial image after edge detection; wherein the pointer outline includes the position of the pointer outline;

[0083] The deflection angle of the pointer is determined based on the extracted pointer profile and the minimum scale.

[0084] Specifically, edge detection is first performed on the dial image to highlight the boundaries of different areas on the dial, including the edge of the pointer, so that subsequent processing can focus more on the pointer part and remove some irrelevant details. The outline of the pointer is extracted from the image after edge detection, including its shape and position information in the image, so that the specific position of the pointer on the dial can be accurately located. Using the extracted pointer outline position and the position of the minimum scale, the deflection angle of the pointer relative to the starting position of the dial (generally the minimum scale position) is calculated through geometric or trigonometric functions. This deflection angle is the key result of pointer recognition, and it is combined with other information to calculate the actual reading of the meter.

[0085] In this embodiment, the position of the pointer can be accurately located through edge detection and pointer contour extraction, and the deflection angle of the pointer can be calculated in combination with the position of the minimum scale.

[0086] In one possible implementation, calculating the pointer reading of the meter according to the corresponding value and unit represented by each degree and the pointer identification result includes:

[0087] Calculate the pointer reading of the meter based on the corresponding pointer deflection angle, the value represented per degree, and the unit.

[0088] Specifically, the value and unit represented per degree reflects the magnitude and unit of the actual physical value (such as current, voltage, etc.) represented by each unit angle on the dial. For example, the value represented per degree is 2 amperes, and the unit is amperes. This value is calculated based on the meter's scale range (such as the values ​​corresponding to the minimum and maximum scales) and the angular range of the scale outline. The pointer deflection angle is the angle the pointer rotates relative to the starting position of the dial (usually the minimum scale position). This angle value is obtained through the pointer recognition process (such as edge detection, pointer outline extraction, and other steps).

[0089] In this embodiment, the corresponding pointer deflection angle is multiplied by the numerical value represented per degree, and the pointer reading is obtained based on the product. This method of calculating the meter reading using the pointer deflection angle, the numerical value represented per degree, and the unit is a key link in the entire meter recognition process, realizing the conversion from image recognition to actual physical quantity reading.

[0090] In a possible implementation, text extraction is performed on each dial image to obtain a text extraction result, including:

[0091] Perform edge detection on each dial image and perform binarization;

[0092] Perform connected domain analysis on the binarized dial image and perform segmentation;

[0093] Perform text recognition on the segmented dial image and extract text features; the text features include the position, size and shape of the text;

[0094] Calculate the similarity between text features and characters in the template library, and determine the text extraction result based on the similarity and similarity threshold.

[0095] Specifically, edge detection is performed on the dial image to highlight the boundaries of different dial areas (such as text, scales, and hands), removing irrelevant details and allowing subsequent processing to focus more on the text. The edge-detected image is converted into a binary image (containing only black and white), further simplifying the image and making the text and background more distinct, facilitating subsequent text segmentation and recognition. Connected domain analysis is performed on the binarized image to identify interconnected pixel regions and distinguish between different text characters.

[0096] Specifically, based on the results of the connected domain analysis, the text in the image is segmented one by one. Each segmented text character is an independent image area, providing a separate processing unit for subsequent text recognition. Text recognition is performed on the segmented text image to determine the content of each text character. This step determines the specific character it represents by analyzing the features of the text image (such as shape, strokes, etc.). Feature information such as the position, size, and shape of the text is extracted. The position of the text can determine its specific orientation in the dial image; the size and shape information helps in subsequent text matching and recognition. The similarity between the extracted text features and the characters in the template library is calculated. The template library contains various possible characters and their feature information. By comparison, the characters most similar to the extracted text can be found. The final text extraction result is determined based on the similarity and the similarity threshold. The similarity threshold is a set standard. Only characters with a similarity higher than the threshold will be considered as matching characters, thereby ensuring the accuracy of text extraction.

[0097] In this embodiment, edge detection and binarization effectively remove noise from the image, highlighting the text and providing a clear image for subsequent text segmentation and recognition. Connected component analysis and segmentation can isolate the individual characters on the dial, providing independent text units for subsequent text recognition. Character recognition and feature extraction can determine the content of each character on the dial and its specific location in the image, providing key information for subsequent meter recognition. Similarity calculation and result verification ensure the accuracy of recognized characters, improving the reliability of character extraction.

[0098] In a possible implementation, performing scale contour recognition on each dial image to obtain a scale recognition result includes:

[0099] Perform edge detection on each dial image;

[0100] Extract scale contours from the dial image after edge detection;

[0101] The angle between the minimum scale and the maximum scale of the dial is calculated according to the extracted scale profile to obtain the angle range of the scale profile.

[0102] In this embodiment, the scale outline is further extracted from the edge-detected image, separating the scale shape and position information from the complex dial image for subsequent processing. By calculating the angle between the minimum and maximum scale marks on the dial, the angular range of the scale outline is determined, clarifying the distribution range of the scale marks on the dial.

[0103] In one possible implementation, the numerical value and unit represented by each degree are calculated according to the corresponding scale range and the scale recognition result, including:

[0104] The value represented by each degree is calculated based on the corresponding scale range and the angular range of the scale profile;

[0105] The unit of the value represented per degree is determined according to the unit of the range.

[0106] In this embodiment, the actual value represented by each degree (per unit angle) is calculated by dividing the scale range of the dial (e.g., the values ​​corresponding to the minimum and maximum scale marks) by the angular range of the scale outline. The units corresponding to the value represented by each degree are determined based on the meter's range units (e.g., amperes, volts, etc.), ensuring that the physical meaning of the calculation result is clear and facilitating subsequent reading calculations and use.

[0107] In a possible implementation, after calculating the pointer reading of the meter according to the corresponding value and unit represented by each degree and the pointer identification result, the method further includes:

[0108] Determine the intermediate scale and its position according to the positions of the minimum scale and the maximum scale;

[0109] The rotation angle of the pointer is calculated based on the intersection of the minimum scale line and the pointer and the circle. A coordinate system is established using the intersection of the minimum scale and the maximum scale and the middle scale, with the origin as the intersection point and the vertical axis as the line between the middle scale and the origin. A circle of a preset radius is constructed with the origin as the center.

[0110] The pointer recognition result is verified based on the pointer recognition result, the pointer rotation angle and the difference threshold.

[0111] In this embodiment, the middle scale on the dial and its specific position are determined based on the position of the minimum scale and the maximum scale. The actual rotation angle of the pointer is calculated using the minimum scale line and the intersection of the pointer and the circle. By establishing a coordinate system with the intersection of the minimum scale and the maximum scale as the origin, and constructing a circle with a preset radius with the origin as the center, the rotation angle of the pointer can be measured more accurately, which helps to improve the accuracy of the reading, especially when the pointer does not point to a clear scale. By combining the pointer recognition results (such as the deflection angle of the pointer), the calculated pointer rotation angle and the set difference threshold, the pointer recognition results are verified, which can correct possible recognition errors and ensure the accuracy of the pointer reading.

[0112] In one possible implementation, obtaining multiple dial images of meters to be identified includes:

[0113] Obtain multiple substation monitoring images to be identified;

[0114] Grayscale processing is performed on the monitoring image of the substation to be identified;

[0115] Filtering the grayscale substation monitoring image to be identified to obtain a preprocessed image;

[0116] The pre-processed image is segmented using a threshold to obtain the initial image of the dial;

[0117] The initial dial image is enhanced to obtain the dial image.

[0118] In this embodiment, a clear dial image is extracted by acquiring a monitoring image, preprocessing (grayscale and filtering), image segmentation (threshold segmentation), and enhancing the image.

[0119] In one possible implementation, the enhancement includes one or more of rotation, translation, scaling, and flipping.

[0120] In this embodiment, enhancement technology is used to improve image quality and ensure that the key features of the meter dial are clear and accurate, providing a basis for subsequent steps such as text extraction, scale recognition, and pointer recognition, thereby improving the accuracy and efficiency of the entire recognition process.

[0121] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0122] The following are device embodiments of the present invention. For details not fully described therein, reference may be made to the corresponding method embodiments described above.

[0123] Figure 2 A schematic diagram of the structure of a meter identification device provided by an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown, which are described in detail as follows:

[0124] like Figure 2 As shown, the meter identification device 2 includes:

[0125] An image acquisition module 21 is used to acquire multiple dial images of the meter to be identified;

[0126] A text extraction module 22 is configured to extract text from each meter dial image to obtain a text extraction result; wherein the text extraction result includes the meter type, range, and scale range, and the scale range includes the minimum scale and the maximum scale and their positions;

[0127] A calculation module 23 is configured to perform scale contour recognition on each dial image to obtain a scale recognition result, and calculate the value and unit represented by each degree based on the corresponding scale range and the scale recognition result; wherein the scale recognition result includes the angular range of the scale contour;

[0128] The pointer reading module 24 is used to perform pointer recognition on each dial image to obtain a pointer recognition result; calculate the pointer reading based on the corresponding numerical value and unit represented by each degree and the pointer recognition result; wherein the pointer recognition result includes the deflection angle of the pointer, and package the text extraction result, scale recognition result, numerical value and unit represented by each degree and the pointer recognition result with the corresponding dial image respectively, and classify and store the packaged data according to the text extraction result, scale recognition result, numerical value and unit represented by each degree and the pointer recognition result.

[0129] In a possible implementation, the calculation module 23 is further configured to:

[0130] Get the scale range and scale recognition results;

[0131] The scale range and scale recognition result are matched and verified based on the dial image pixels to obtain the corresponding scale range and scale recognition result.

[0132] In a possible implementation, the pointer reading module 24 is further configured to:

[0133] Obtain the value and unit of each degree in the scale recognition result and the pointer recognition result;

[0134] Based on the dial image pixels, the numerical value represented by each degree, the unit and the pointer recognition result are matched and checked to obtain the corresponding numerical value represented by each degree, the unit and the pointer recognition result.

[0135] In a possible implementation, the pointer reading module 24 is further configured to:

[0136] Perform edge detection on each dial image;

[0137] Extracting the pointer outline from the dial image after edge detection; wherein the pointer outline includes the position of the pointer outline;

[0138] The deflection angle of the pointer is determined based on the extracted pointer profile and the minimum scale.

[0139] In a possible implementation, the pointer reading module 24 is further configured to:

[0140] Calculate the pointer reading of the meter based on the corresponding pointer deflection angle, the value represented per degree, and the unit.

[0141] In a possible implementation, the text extraction module 22 is further configured to:

[0142] Perform edge detection on each dial image and perform binarization;

[0143] Perform connected domain analysis on the binarized dial image and perform segmentation;

[0144] Perform text recognition on the segmented dial image and extract text features; the text features include the position, size and shape of the text;

[0145] Calculate the similarity between text features and characters in the template library, and determine the text extraction result based on the similarity and similarity threshold.

[0146] In a possible implementation, the calculation module 23 is further configured to:

[0147] Perform edge detection on each dial image;

[0148] Extract scale contours from the dial image after edge detection;

[0149] The angle between the minimum scale and the maximum scale of the dial is calculated according to the extracted scale profile to obtain the angle range of the scale profile.

[0150] In a possible implementation, the calculation module 23 is further configured to:

[0151] The value represented by each degree is calculated based on the corresponding scale range and the angular range of the scale profile;

[0152] The unit of the value represented per degree is determined according to the unit of the range.

[0153] In an embodiment of the present invention, by processing text extraction tasks for multiple images in parallel, the efficiency of text extraction can be improved. Similarly, processing scale recognition and calculation tasks for multiple images in parallel can increase the speed of scale recognition and calculation. In addition, processing pointer recognition and reading calculation tasks for multiple images in parallel can also improve the efficiency of pointer recognition and reading calculation. Compared with traditional methods, this method can process multiple images simultaneously, fully mobilize and utilize computing resources, reduce processing time, and avoid the problem of idle resources caused by processing images one by one in traditional methods.

[0154] Figure 3 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, the electronic device 3 of this embodiment includes a processor 30 and a memory 31. The memory 31 stores a computer program 32. When the processor 30 executes the computer program 32, the steps of the above-described method embodiments are implemented. Alternatively, when the processor 30 executes the computer program 32, the functions of the modules / units in the above-described device embodiments are implemented.

[0155] Exemplarily, the computer program 32 may be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 30 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program 32 in the electronic device 3.

[0156] The electronic device 3 may include, but is not limited to, a processor 30 and a memory 31. Those skilled in the art will appreciate that Figure 3 It is only an example of electronic device 3 and does not constitute a limitation of electronic device 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, electronic device 3 may also include input and output devices, network access devices, buses, etc.

[0157] For the sake of convenience and brevity, the division of the above functional modules / units is only used as an example. In actual applications, the above functions can be assigned to different functional modules / units as needed. The above modules / units can be implemented in the form of hardware, software, or a combination of hardware and software.

[0158] In the above embodiments, the descriptions of each embodiment have their own focus. For parts not described or recorded in detail in one embodiment, please refer to the relevant descriptions of other embodiments. Unless otherwise specified or there is a logical conflict, the terms and / or descriptions between different embodiments are consistent and can be referenced to each other. The technical features of different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0159] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A meter identification method, characterized in that: include: Acquire multiple dial images of meters to be identified; Perform text extraction on each dial image to obtain a text extraction result; wherein the text extraction result includes the meter category, range and scale range, and the scale range includes the minimum scale and maximum scale and their positions; perform scale contour recognition on each dial image to obtain a scale recognition result, and calculate the value and unit represented by each degree based on the corresponding scale range and scale recognition result; wherein the scale recognition result includes the angular range of the scale contour; perform pointer recognition on each dial image to obtain a pointer recognition result; calculate the pointer reading based on the corresponding value and unit represented by each degree and the pointer recognition result; wherein the pointer recognition result includes the deflection angle of the pointer, and package the text extraction result, scale recognition result, value and unit represented by each degree, and pointer recognition result with the corresponding dial image, and classify and store the packaged data according to the text extraction result, scale recognition result, value and unit represented by each degree, and pointer recognition result.

2. The meter identification method according to claim 1, characterized in that: Before calculating the value and unit of each degree according to the corresponding scale range and scale recognition result, including: Get the scale range and scale recognition results; Performing a matching check on the scale range and the scale recognition result based on the dial image pixels to obtain a corresponding scale range and scale recognition result; Before calculating the meter pointer reading according to the corresponding value per degree and unit and pointer identification results, including: Obtain the value and unit of each degree in the scale recognition result and the pointer recognition result; Based on the dial image pixels, the numerical value represented by each degree, the unit and the pointer recognition result are matched and checked to obtain the corresponding numerical value represented by each degree, the unit and the pointer recognition result.

3. The meter identification method according to claim 1, characterized in that: The step of performing pointer recognition on each dial image to obtain a pointer recognition result includes: Perform edge detection on each dial image; Extracting the pointer outline from the dial image after edge detection; wherein the pointer outline includes the position of the pointer outline; The deflection angle of the pointer is determined according to the extracted pointer profile and the minimum scale.

4. The meter identification method according to claim 3, characterized in that: The step of calculating the pointer reading of the meter according to the corresponding numerical value and unit represented by each degree and the pointer identification result includes: The pointer reading of the meter is calculated according to the corresponding pointer deflection angle and the value and unit represented by each degree.

5. The meter identification method according to claim 1, characterized in that: The step of extracting text from each dial image to obtain a text extraction result includes: Perform edge detection on each dial image and perform binarization; Perform connected domain analysis on the binarized dial image and perform segmentation; Perform text recognition on the segmented dial image and extract text features; wherein the text features include the position, size and shape of the text; The similarity between the text feature and the characters in the template library is calculated, and the text extraction result is determined according to the similarity and a similarity threshold.

6. The meter identification method according to claim 1, characterized in that: The step of performing scale contour recognition on each dial image to obtain a scale recognition result includes: Perform edge detection on each dial image; Extract scale contours from the dial image after edge detection; The angle between the minimum scale and the maximum scale of the dial is calculated according to the extracted scale profile to obtain the angle range of the scale profile.

7. The meter identification method according to claim 6, characterized in that: The calculation of the value and unit represented by each degree according to the corresponding scale range and scale recognition result includes: The value represented by each degree is calculated based on the corresponding scale range and the angular range of the scale profile; The unit of the value expressed per degree is determined according to the unit of the range.

8. The meter identification method according to claim 1, characterized in that: After calculating the pointer reading of the meter according to the corresponding value and unit represented by each degree and the pointer identification result, it also includes: Determine the intermediate scale and its position according to the positions of the minimum scale and the maximum scale; The rotation angle of the pointer is calculated based on the intersection of the minimum scale line and the pointer and the circle; wherein a coordinate system is established using the intersection of the minimum scale and the maximum scale and the intermediate scale, with the origin being the intersection point, the ordinate axis being the straight line between the intermediate scale and the origin, and a circle of a preset radius being constructed with the origin as the center; The pointer recognition result is verified based on the pointer recognition result, the pointer rotation angle and the difference threshold.

9. A meter identification device, characterized in that: include: An image acquisition module, used to acquire multiple dial images of meters to be identified; a text extraction module, configured to extract text from each dial image to obtain a text extraction result; wherein the text extraction result includes the meter type, range, and scale range, and the scale range includes the minimum scale and the maximum scale and their positions; a calculation module for performing scale contour recognition on each dial image to obtain a scale recognition result, and calculating the value and unit represented by each degree based on the corresponding scale range and the scale recognition result; wherein the scale recognition result includes the angular range of the scale contour; The pointer reading module is used to perform pointer recognition on each dial image to obtain a pointer recognition result; calculate the pointer reading based on the corresponding numerical value and unit represented by each degree and the pointer recognition result; wherein the pointer recognition result includes the deflection angle of the pointer, and the text extraction result, scale recognition result, numerical value and unit represented by each degree and pointer recognition result are respectively packaged with the corresponding dial image, and the packaged data are classified and stored according to the text extraction result, scale recognition result, numerical value and unit represented by each degree and pointer recognition result.

10. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

Images