A method and system for identifying readings of wheel-type instruments

Abstract

The present invention belongs to the technical field of instrument reading recognition, and provides a method and system for recognizing the reading of a wheel-type instrument. The method for recognizing the reading of a wheel-type instrument includes obtaining an image of a wheel-type instrument dial; using a high-order recognition model and a low-order recognition model to infer the image of the wheel-type instrument dial respectively, and obtaining the category, confidence and corresponding center coordinate distribution of each digit in the high-order digital area and the lowest-order digital area in a preset coordinate system, and then judging whether the reasoning of the high-order recognition model and the low-order recognition model is valid; if the reasoning is valid, the value corresponding to the corresponding digital area is determined; wherein, the value corresponding to the digit category of the high-order digital area is an integer corresponding to the phase digit, and the value corresponding to the digit category of the lowest-order digital area is a decimal containing the next digit of the lowest-order digit. It reduces the interference links that affect image recognition, has high recognition accuracy, and calculates the next digit of the lowest digit.

Description

Technical Field

[0001] The invention belongs to the technical field of meter reading recognition, and in particular relates to a method and system for recognizing the reading of a character wheel type meter. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In the field of wheel-type meter reading recognition technology, the wheel-type meter image is subjected to image segmentation and cropping, grayscale processing, image corrosion or expansion processing, and binarization processing. Each link of these image processing may affect the accuracy of recognition. For example, the image segmentation and cropping processing is affected by the rotation angle, and if it is not handled properly, it will affect the recognition accuracy; the binarization process may be affected by light, making the binarization threshold not fixed, and if the threshold is set unreasonably, it will also affect the recognition accuracy.

[0004] In addition, the traditional reading recognition principle of wheel-type meters adopts a customized method, such as the template matching method to compare the closest features bit by bit to achieve the purpose of recognition, which is closely related to information such as the shape of the numbers; and it can only recognize the data on the dial, and cannot deduce the next digit on the dial. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a method and system for identifying readings of a character wheel type instrument, which does not require image segmentation and cropping, grayscale processing, image corrosion or expansion processing and binarization processing of the character wheel type instrument image, reduces the interference links that affect image recognition, and can calculate the next lowest digit.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A first aspect of the present invention provides a method for identifying readings of a character wheel type meter.

[0008] In one or more embodiments, a method for identifying readings of a wheel-type meter is provided, comprising:

[0009] Acquire a wheel-type instrument dial image; the digital area in the wheel-type instrument dial image is composed of a high-order digital area and a low-order digital area; the high-order digital area contains at least one digital digit;

[0010] The high-order recognition model and the low-order recognition model are used to infer the dial image of the wheel-type instrument respectively, and the category, confidence and corresponding center coordinate distribution of each digit in the high-order digit area and the lowest-order digit area in a preset coordinate system are obtained; wherein the value corresponding to the digit category of the high-order digit area is an integer corresponding to the phase digit, and the value corresponding to the digit category of the lowest-order digit area is a decimal containing the next digit of the lowest-order digit;

[0011] Based on the number of categories, center coordinate distribution and confidence of each digit, the validity of the reasoning of the high-order recognition model and the low-order recognition model is judged respectively; if the reasoning of the high-order recognition model and the low-order recognition model are both valid, the numerical value corresponding to the corresponding digital area is determined according to the categories of the digits in the high-order digital area and the lowest-order digital area.

[0012] A second aspect of the present invention provides a character wheel type meter reading recognition system.

[0013] In one or more embodiments, a wheel-type meter reading recognition system includes:

[0014] A dial image acquisition module is used to acquire a dial image of a wheel-type instrument; the digital area in the wheel-type instrument dial image is composed of a high-order digital area and a low-order digital area; the high-order digital area contains at least one digital digit;

[0015] A digit category inference module, which is used to use a high-order recognition model and a low-order recognition model to respectively infer the image of the wheel-type instrument dial, and obtain the category, confidence and corresponding center coordinate distribution of each digit in the high-order digit area and the lowest-order digit area in a preset coordinate system; wherein the value corresponding to the digit category of the high-order digit area is an integer corresponding to the phase digit, and the value corresponding to the digit category of the lowest-order digit area is a decimal containing the next digit of the lowest-order digit;

[0016] A digit value determination module is used to determine whether the reasoning of the high-order recognition model and the low-order recognition model is valid based on the number of categories, center coordinate distribution and confidence of each digit; if the reasoning of the high-order recognition model and the low-order recognition model are both valid, the value corresponding to the corresponding digital area is determined according to the category of the digits in the high-order digital area and the lowest-order digital area.

[0017] A third aspect of the present invention provides a computer-readable storage medium.

[0018] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps in the above-mentioned method for identifying readings of a wheel-type meter.

[0019] A fourth aspect of the present invention provides an electronic device.

[0020] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the above-mentioned method for identifying readings of a wheel-type meter are implemented.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] (1) The present invention utilizes a high-bit recognition model and a low-bit recognition model to process the dial image of the character wheel instrument respectively, without the need for image segmentation and cropping, grayscale processing, image corrosion or expansion processing, and binarization processing, thereby reducing interference links that affect image recognition and improving the accuracy of image processing results.

[0023] (2) The present invention utilizes a high-order recognition model to process the image of the wheel-type instrument dial, and can infer that the value corresponding to the digit category of the high-order digit area is the integer corresponding to the phase digit; and utilizes a low-order recognition model to process the image of the wheel-type instrument dial, and can infer that the value corresponding to the digit category of the lowest-order digit area is the decimal containing the next digit of the lowest-order digit, and the entire recognition process is simple, highly accurate and fast. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0025] Figure 1 1 is a flow chart of a method for identifying readings of a wheel-type meter according to an embodiment of the present invention;

[0026] Figure 2 It is a high-position image sample on the dial of the character wheel type instrument in an embodiment of the present invention;

[0027] Figure 3 It is a low-bit image sample on the dial of the character wheel type instrument in the embodiment of the present invention;

[0028] FIG4 is a schematic diagram of the center position of a digit of training data according to an embodiment of the present invention;

[0029] Among them, (a) is a schematic diagram of the center position of the first digit of the training data; (b) is a schematic diagram of the center position of the second digit of the training data; (c) is a schematic diagram of the center position of the third digit of the training data;

[0030] Figure 5 is a high-level reasoning schematic diagram of an embodiment of the present invention;

[0031] Figure 6is a low-level reasoning schematic diagram of an embodiment of the present invention;

[0032] Figure 7 2 is a schematic diagram of the structure of a character wheel type meter reading recognition system according to an embodiment of the present invention;

[0033] Figure 8 is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0035] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0036] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0037] In the present invention, the character wheel type instrument can be an electric energy meter, a water meter, etc. Moreover, the character wheel in the character wheel type instrument can be a decimal system, an octal system or other artificially defined systems.

[0038] It should be noted here that different base numbers are numbers that use corresponding numbers to equally divide one circle of the outer surface of the roller. For example, the decimal system uses 0 to 9 to equally divide one circle of the outer surface of the roller into ten.

[0039] Figure 1 FIG. 1 is a flow chart of a method for identifying readings of a wheel-type meter according to an embodiment of the present invention. Figure 1 The method for identifying the readings of a wheel-type meter in the embodiment shown may include:

[0040] S101, acquiring a dial image of a wheel-type instrument; the digital area in the wheel-type instrument dial image is composed of a high-order digital area and a low-order digital area; the high-order digital area contains at least one digital digit;

[0041] S102, using a high-order recognition model and a low-order recognition model to infer the dial image of the wheel-type instrument respectively, to obtain the category, confidence and corresponding center coordinate distribution of each digit in the high-order digit area and the lowest-order digit area in a preset coordinate system; wherein the value corresponding to the digit category of the high-order digit area is an integer corresponding to the phase digit, and the value corresponding to the digit category of the lowest-order digit area is a decimal containing the next digit of the lowest-order digit;

[0042] S103, based on the number of categories, center coordinate distribution and confidence of each digital digit, determine whether the reasoning of the high-order recognition model and the low-order recognition model is valid; if the reasoning of the high-order recognition model and the low-order recognition model are both valid, determine the value corresponding to the corresponding digital area according to the category of the digital digits in the high-order digital area and the lowest-order digital area.

[0043] This embodiment uses a high-bit recognition model and a low-bit recognition model to process the image of the wheel-type instrument dial respectively. There is no need to separate and crop the image, grayscale it, erode or dilate it, or binarize it, which reduces interference factors that affect image recognition and improves the accuracy of image processing results.

[0044] This embodiment uses a high-bit recognition model to process the image of the wheel-type instrument dial, and can infer that the value corresponding to the digit category of the high-bit digit area is the integer corresponding to the phase digit; uses a low-bit recognition model to process the image of the wheel-type instrument dial, and can infer that the value corresponding to the digit category of the lowest-bit digit area is the decimal containing the next digit of the lowest-bit digit, and the entire recognition process is simple, with high accuracy and high speed.

[0045] In step S101, taking a decimal wheel type instrument as an example:

[0046] The lowest digit area in the dial image of the wheel-type instrument contains the unit digit, and the high digit area contains the tens digit and digits above it, such as the four digits of ten thousand, thousand, hundred, and ten.

[0047] In the specific implementation process, the acquired dial image of the character wheel instrument and the image samples for training the high-position recognition model or the low-position recognition model are all acquired by the image acquisition device at the same camera position.

[0048] In step S102, the high-order recognition model and the low-order recognition model are both existing neural networks or deep learning models, such as Yolo series models.

[0049] The high-order recognition model and the low-order recognition model can adopt a neural network or a deep learning model architecture and be trained simultaneously.

[0050] In the specific implementation process, the number of recognition types of the high-order recognition model is the same as the base number of the digits of the wheel-type instrument; the number of recognition types of the low-order recognition model is the second power of the base number of the digits of the wheel-type instrument.

[0051] Taking the decimal system of the digital position of the wheel-type instrument as an example, the training sample formation process of the high-bit recognition model and the low-bit recognition model is given below:

[0052] (1) Data collection before training;

[0053] First, record a video of the instrument panel code rotating; among them:

[0054] Video requirements: Fixed camera position, with the disc code located in the center of the video and horizontal;

[0055] Video coverage: from 00000 to 99999;

[0056] Frame rate: 60FPS, resolution 320*320;

[0057] The dial rotates at a constant speed of 36 degrees per second. The dial code changes by one digit every second.

[0058] (2) Pre-training data processing;

[0059] Extract one frame for every set frame (e.g. 6 frames): total number of frames s = 99999*1*60 = 5999940;

[0060] Frame extraction formula: (k+1)%6=1, k represents the frame number; if k satisfies (k+1)%6=1, then k is the target frame, and the original jpg file is obtained, with a size of 15-25KB. In order to obtain a smaller image, the current frame is compressed first to obtain a new jpg file with a file size of 1.5-2.5KB, and then saved as a jpg file with the file name k.jpg.

[0061] Because the video frame rate is 60FPS, 10 pictures are drawn in 1 second, and the rotation is 36 degrees in 1 second. After 1 second, the disk code changes a number, so these 10 pictures only look at the lowest digit, changing evenly from a to b, such as a.0, a.1, ..., a.9, b.0, and generate a corresponding new file, the file name is k.txt, which is used to record the corresponding readings, including the readings and the coordinates of each number, to obtain data for model training.

[0062] When the digits of the wheel-type instrument are in decimal, the number of identification types of the high-order identification model is ten, corresponding to the digits 0 to 9 respectively. Figure 2As shown, the high-order recognition model only trains high-order (ten thousand, thousand, hundred, ten), a total of 10 categories, marked as categories A, B, C, D, E, F, G, H, I, J. Each category has 4000 sample data.

[0063] When the digital digits of the wheel-type instrument are in decimal, the number of identification types of the low-order identification model is one hundred, corresponding to the values ​​of one digit after the decimal point in the range of [0.0, 9.9]. Figure 3 As shown, the low-order recognition model only trains the lowest order (units), with a total of 100 categories, marked as categories Aa, Ab...Aj, Ba, Bb,..., Ji, Jj. Each category has 10,000 sample data.

[0064] In the specific implementation process, the center coordinates of each digit in the high-order digit area and the lowest-order digit area are measured at the standard position. Through conversion, the coordinates are set to [0,1]. The coordinate value range of the training data is also [0,1], and the inferred coordinate value range is also [0,1].

[0065] As shown in Figure 4, taking the decimal system of the digits of the wheel-type instrument as an example, the green frame and the yellow frame are standard inference areas. In this embodiment, the origin of the coordinate system is: the lower left point of the highest digit area in the wheel-type instrument image.

[0066] The x-coordinate of the center of the first category (red frame) inferred in Figure 4(b) can fall at the center of the first green frame, close to 0.125. The same is true for other categories.

[0067] In Figure 4(a) (the camera needs to move to the left to achieve the effect in the middle), the inferred center of the first category (red frame) has an x-coordinate that falls on the left edge of the first green frame, close to 0.03125. The same is true for other categories.

[0068] In Figure 4(c) (the camera needs to move to the right to achieve the effect in the middle), the inferred center of the first category (red frame) has an x-coordinate that falls on the left edge of the first green frame, close to 0.21850. The same is true for other categories.

[0069] So the center coordinates of the first category are x∈[0.03125,0.21850); the same is true for other categories.

[0070] Therefore, the standard range of the horizontal coordinate of the center point corresponding to the high-order digital position obtained by the high-order recognition model is as follows: Figure 5 As shown, from small to large they are:

[0071] The coordinates of the ten thousandth digit x∈[0.03125,0.21850);

[0072] Thousands coordinates x∈[0.21850,0.40625);

[0073] Hundreds coordinate x∈[0.40625,0.59375);

[0074] The tens coordinate x∈[0.59375,0.78125).

[0075] The standard range of the horizontal coordinate of the center point corresponding to the low-order digital position obtained by the low-order recognition model is as follows: Figure 6 Shown as: the coordinates of the units digit x∈[0.78125,0.96875).

[0076] In step S102, the reading cannot be directly recognized using the high-order recognition model and the low-order recognition model. What is inferred is the category, confidence level and corresponding center coordinate distribution of each digit in the high-order digit area and the lowest-order digit area in the preset coordinate system, which requires secondary processing.

[0077] In step S103, for the inference result of the high-order recognition model, when the number of digits recognized in the inference result of the high-order recognition model is less than the number of digits actually contained in the high-order digit area, the inference of the high-order recognition model is judged to be invalid;

[0078] Regarding the reasoning results of the high-bit recognition model, when the reasoning of the high-bit recognition model is valid, when there are multiple category reasoning results for a digit, the category with the largest confidence is taken as the category of the corresponding digit.

[0079] For the inference result of the high-order recognition model, when the number of digits recognized in the inference result of the high-order recognition model is greater than or equal to the number of digits m actually contained in the high-order digit area:

[0080] According to the distribution of the center coordinates corresponding to the digits identified in the inference result in the preset coordinate system, the digits are sorted from small to large according to the horizontal coordinates, and the first m identified digits are taken;

[0081] In the same preset coordinate system, determine whether the center coordinates of the m recognized digits are distributed within the center coordinate range of the digits actually contained in the high-order digital area. If so, the reasoning of the high-order recognition model is valid; otherwise, the reasoning of the high-order recognition model is invalid.

[0082] Take the decimal system of the digital digits of the word wheel instrument as an example:

[0083] When the inference result of the high-level recognition model is 5 categories, first sort the inference results from small to large according to the horizontal coordinates of the central coordinates corresponding to the results, and take the first 4 categories. If each category is distributed in the four standard intervals of ten thousand, thousand, hundred, and ten, the inference is valid, corresponding to ten thousand, thousand, ten, and hundred respectively; otherwise, the inference fails.

[0084] When the inference result of the high-level recognition model is 4 categories, first sort them from small to large according to the horizontal coordinates of the central coordinates corresponding to the inference results. If the horizontal coordinates of the central coordinates of each category are distributed in the four standard intervals of ten thousand, thousand, hundred, and ten, the inference is valid, corresponding to ten thousand, thousand, ten, and hundred respectively; otherwise, the inference fails.

[0085] When the inference result of the high-order recognition model is less than 4 categories, the inference fails.

[0086] When the inference result of the high-order recognition model is greater than 5 categories, if the horizontal coordinates of the center coordinates of each category are distributed in the four standard intervals of ten thousand, thousand, hundred, and ten, and if there are multiple records in an interval, the record with the highest confidence is taken, and the inference is valid, otherwise the inference fails;

[0087] The above completes the high-order reasoning and calculates the ten thousand, thousand, hundred and ten digits.

[0088] For the reasoning results of the low-order recognition model, when the central coordinates of the category results are distributed within the central coordinate range of the digits in the lowest-order digital area, the reasoning of the low-order recognition model is judged to be invalid; when there are multiple category reasoning results for the digits in the lowest-order digital area, the category with the largest confidence is taken as the category of the corresponding digit.

[0089] In one or more embodiments, when the reasoning of either the high-order recognition model or the low-order recognition model is invalid, the corresponding model with invalid reasoning is updated, and the reasoning of the corresponding digital area is re-performed until the reasoning is valid. This can improve the image recognition accuracy and ensure the recognition effect.

[0090] In other embodiments, the categories of the digits in the high-order digit area and the lowest-order digit area and their corresponding values ​​are pre-stored in corresponding tables, respectively, and the values ​​corresponding to the corresponding digit areas are determined by querying the tables.

[0091] Specifically, the categories of the digits in the high-order digit area are pre-stored via a first table; the first table is a two-column or two-row table, which respectively correspond to the high-order category codes and their corresponding values, as shown in Table 1.

[0092] Table 1 Categories of digits in the high-order digit area and their corresponding values

[0093]

[0094]

[0095] Specifically, the categories of the digital digits in the lowest digit area are pre-stored through a second table; the second table is an (n+1)*(n+1) table; n represents the base number of the digital digits of the digital wheel instrument; the categories of the digital digits in the lowest digit area are encoded in a two-bit format, and the two-bit codes correspond to the headers of the second table after being split, and the middle values ​​of the second table are the values ​​corresponding to each category, as shown in Table 2.

[0096] Table 2 Categories of digits in the low-order digit area and their corresponding values

[0097]

[0098]

[0099] The reading recognition accuracy of the character wheel type meter of this embodiment is improved from the traditional 90%-95% to more than 99.21%, and the response speed is improved from the traditional 100ms-3000ms to 10ms-100ms.

[0100] Figure 7 Schematic diagram of a wheel-type meter reading recognition system in an embodiment of the present invention. Figure 1 The reading recognition method of the wheel type instrument corresponds to Figure 7 As shown, the wheel-type meter reading recognition system in this embodiment may include:

[0101] The dial image acquisition module 701 is used to acquire a dial image of a wheel-type instrument; the digital area in the wheel-type instrument dial image is composed of a high-order digital area and a low-order digital area; the high-order digital area contains at least one digital digit;

[0102] The digit category inference module 702 is used to use the high-order recognition model and the low-order recognition model to respectively infer the image of the wheel-type instrument dial, and obtain the category, confidence and corresponding center coordinate distribution of each digit in the high-order digit area and the lowest-order digit area in a preset coordinate system; wherein the value corresponding to the digit category of the high-order digit area is the integer corresponding to the phase digit, and the value corresponding to the digit category of the lowest-order digit area is the decimal containing the next digit of the lowest-order digit;

[0103] The digit value determination module 703 is used to determine whether the reasoning of the high-order recognition model and the low-order recognition model is valid based on the number of categories, center coordinate distribution and confidence of each digit; if the reasoning of the high-order recognition model and the low-order recognition model are both valid, the value corresponding to the corresponding digital area is determined according to the categories of the digits in the high-order digital area and the lowest-order digital area.

[0104] It should be noted here that Figure 7 The dial image acquisition module 701, the digit category inference module 702 and the digit value determination module 703 in the digital wheel meter reading recognition system are Figure 1 The specific implementation process of steps S101 to S103 in the character wheel meter reading recognition method is the same and will not be repeated here.

[0105] Reference Figure 8 , a schematic diagram of an electronic device is given. It should be noted that, Figure 8 The electronic device 800 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0106] like Figure 8 As shown, electronic device 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage part 808 into a random access memory (RAM) 803. In RAM 803, various programs and data required for system operation are also stored. Central processing unit 801, ROM 802 and RAM 803 are connected to each other via a bus 804. Input / output (I / O) interface 805 is also connected to bus 804.

[0107] The following components are connected to the I / O interface 805: an input section 806 including a keyboard, a mouse, etc.; an output section 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a local area network (LAN) card, a modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed, so that a computer program read therefrom is installed into the storage section 808 as needed.

[0108] When the central processing unit 801 in the electronic device of this embodiment executes the program, the following is achieved: Figure 1 The steps in the method for identifying the readings of a wheel-type meter are shown.

[0109] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, the computer program including a computer program for executing Figure 1In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 809, and / or installed from a removable medium 811. When the computer program is executed by the central processing unit 801, various functions defined in the apparatus of the present application are executed.

[0110] in, Figure 1 The computer program instructions corresponding to the method shown may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0111] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0112] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for identifying readings of a wheel-type instrument, characterized in that: include: Get the image of the dial of the wheel-type instrument; The digital area in the dial image of the wheel-type instrument consists of a high-order digital area and a low-order digital area; the high-order digital area contains at least one digital digit; The high-order recognition model and the low-order recognition model are used to infer the dial image of the wheel-type instrument respectively, and the category, confidence and corresponding center coordinate distribution of each digit in the high-order digit area and the lowest-order digit area in a preset coordinate system are obtained; wherein the value corresponding to the digit category of the high-order digit area is an integer corresponding to the phase digit, and the value corresponding to the digit category of the lowest-order digit area is a decimal containing the next digit of the lowest-order digit; Based on the number of categories, center coordinate distribution and confidence of each digit, the validity of the reasoning of the high-order recognition model and the low-order recognition model is judged respectively; if the reasoning of the high-order recognition model and the low-order recognition model are both valid, the numerical value corresponding to the corresponding digital area is determined according to the categories of the digits in the high-order digital area and the lowest-order digital area.

2. The method for recognizing readings of a wheel-type meter as claimed in claim 1, characterized in that: When the reasoning of either the high-order recognition model or the low-order recognition model is invalid, the corresponding model whose reasoning is invalid is updated, and the reasoning of the corresponding digital area is re-performed until the reasoning is valid.

3. The method for recognizing readings of a wheel-type meter as claimed in claim 1, characterized in that: The number of identification types of the high-bit identification model is the same as the number of digits of the word wheel instrument; Or / and, the number of identification types of the low-bit identification model is the second power of the number of digits of the digital wheel instrument.

4. The method for identifying readings of a wheel-type meter as claimed in claim 1, characterized in that: The categories of the digits in the high-order digit area and the lowest-order digit area and their corresponding values ​​are pre-stored in corresponding tables respectively, and the values ​​corresponding to the corresponding digit areas are determined by querying the tables.

5. The method for recognizing readings of a wheel-type meter as claimed in claim 1, characterized in that: For the inference result of the high-order recognition model, when the number of digits recognized in the inference result of the high-order recognition model is less than the number of digits actually contained in the high-order digit area, the inference of the high-order recognition model is judged to be invalid; Or for the reasoning results of the high-bit recognition model, when the reasoning of the high-bit recognition model is valid, when there are multiple category reasoning results for a digit, the category with the largest confidence is taken as the category of the corresponding digit.

6. The method for recognizing readings of a wheel-type meter as claimed in claim 1, characterized in that: For the inference result of the high-order recognition model, when the number of digits recognized in the inference result of the high-order recognition model is greater than or equal to the number of digits m actually contained in the high-order digit area: According to the distribution of the center coordinates corresponding to the digits identified in the inference result in the preset coordinate system, the digits are sorted from small to large according to the horizontal coordinates, and the first m identified digits are taken; In the same preset coordinate system, determine whether the center coordinates of the m recognized digits are distributed within the center coordinate range of the digits actually contained in the high-order digital area. If so, the reasoning of the high-order recognition model is valid; otherwise, the reasoning of the high-order recognition model is invalid.

7. The method for recognizing readings of a wheel-type meter as claimed in claim 1, characterized in that: For the reasoning results of the low-order recognition model, when the central coordinates of the category results are distributed within the central coordinate range of the digits in the lowest-order digital area, the reasoning of the low-order recognition model is judged to be invalid; when there are multiple category reasoning results for the digits in the lowest-order digital area, the category with the largest confidence is taken as the category of the corresponding digit.

8. A wheel-type meter reading recognition system, characterized in that: include: A dial image acquisition module, which is used to acquire a dial image of a wheel-type instrument; The digital area in the dial image of the wheel-type instrument consists of a high-order digital area and a low-order digital area; the high-order digital area contains at least one digital digit; A digit category inference module, which is used to use a high-order recognition model and a low-order recognition model to respectively infer the image of the wheel-type instrument dial, and obtain the category, confidence and corresponding center coordinate distribution of each digit in the high-order digit area and the lowest-order digit area in a preset coordinate system; wherein the value corresponding to the digit category of the high-order digit area is an integer corresponding to the phase digit, and the value corresponding to the digit category of the lowest-order digit area is a decimal containing the next digit of the lowest-order digit; A digit value determination module is used to determine whether the reasoning of the high-order recognition model and the low-order recognition model is valid based on the number of categories, center coordinate distribution and confidence of each digit; if the reasoning of the high-order recognition model and the low-order recognition model are both valid, the value corresponding to the corresponding digital area is determined according to the category of the digits in the high-order digital area and the lowest-order digital area.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in the method for recognizing readings of a wheel-type meter as described in any one of claims 1 to 7 are implemented.

10. An electronic 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 program, the steps in the method for recognizing readings of a wheel-type meter as described in any one of claims 1 to 7 are implemented.

Images