Meter identification method

Abstract

The embodiment of the invention provides a meter identification method. The method comprises the following steps: acquiring a target meter picture and target point location information of the target meter picture; performing meter identification on the target meter picture according to the target calibration information and an identification rule indicated by the target calibration picture to obtain a first meter identification result; wherein the corresponding relationship among the at least one group of calibration pictures, the calibration information and the point location information is established by the following steps of: acquiring pictures shot by image acquisition equipment at each preset point location as the calibration pictures of each preset point location; obtaining point location information of each calibration picture; aiming at each calibration picture, calibrating the calibration picture by using an automatic calibration algorithm to obtain calibration information of the calibration picture; and correspondingly recording the calibration picture, the calibration information of the calibration picture and the point location information of the calibration picture. By applying the technical scheme provided by the embodiment of the invention, the meter identification can be efficiently realized.

Description

Technical Field

[0001] The present application relates to the technical field of information collection and analysis, and in particular to a meter identification method. Background Art

[0002] In complex production scenarios, such as power inspection scenarios and building scenarios, the types of meters in the scenarios are complex and diverse, and the number of meters is also large. A related art discloses a meter identification method and system (CN 117876648 A). In this method, when a first camera is added to the meter identification system, the user terminal configures the calibration set ID, wake-up plan, and communication address of the video recorder for the first camera, and controls the first camera to capture a first image. The video recorder obtains the calibration set ID of the first camera and the calibration parameters corresponding to the first camera. As a result, the first camera can capture a second image according to the wake-up plan and send the calibration set ID and the captured second image to the video recorder. The video recorder determines the calibration parameters used to identify the second image based on the received calibration set ID, thereby realizing the recognition of the meter reading in the second image.

[0003] However, it is understandable that the above method requires different calibration parameters to be used for identification of different meters. In other words, the above method requires setting up corresponding identification rules for each meter in advance. This is extremely labor-intensive and inefficient in scenarios with a large number of different meters. Therefore, how to set up an efficient meter identification method has become an urgent problem to be solved.

[0004] Based on this, the present application provides a meter identification method, which improves the meter identification efficiency in scenarios where there are a large number of different meters by improving the calibration method of calibration parameters. Summary of the Invention

[0005] The purpose of the embodiments of the present application is to provide a meter identification method to improve the efficiency of meter identification. The specific technical solution is as follows:

[0006] In a first aspect, an embodiment of the present application provides a meter identification method, which is applied to an inspection host device, and the method includes:

[0007] Acquire a target meter image and target point information of the target meter image; wherein the target point information is used to indicate the point where the image acquisition device was located when capturing the target meter image;

[0008] Based on the correspondence between the calibration image, the calibration information, and the point information, the calibration information and the calibration image corresponding to the target point information are determined as the target calibration information and the target calibration image; wherein the calibration information and the calibration image are used to instruct the inspection host device to perform recognition rules when performing meter recognition on the meter image taken at the corresponding point information;

[0009] performing meter recognition on the target meter image according to the recognition rules indicated by the target calibration information and the target calibration image, and obtaining a meter recognition result of the target meter image as a first meter recognition result;

[0010] The correspondence between at least one set of calibration images, calibration information, and point information is established in the following manner:

[0011] Acquire images taken by an image acquisition device at each preset point as calibration images of each preset point; and acquire point information of each calibration image; wherein the point information of the calibration image is used to indicate the point at which the image acquisition device was located when taking the calibration image;

[0012] For each calibration image, the calibration image is calibrated using an automatic calibration algorithm to obtain calibration information of the calibration image; and the calibration image, the calibration information of the calibration image, and the point information of the calibration image are correspondingly recorded.

[0013] In a second aspect, an embodiment of the present application further provides an electronic device, including:

[0014] Memory for storing computer programs;

[0015] The processor is configured to implement any of the above-mentioned meter identification methods when executing a program stored in the memory.

[0016] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any of the above-mentioned meter identification methods is implemented.

[0017] In a fourth aspect, an embodiment of the present application further provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute any of the meter identification methods described in the above embodiments.

[0018] Beneficial effects of the embodiments of the present application:

[0019] In the technical solution provided by the embodiments of this application, a correspondence is established between calibration images, calibration information, and point information. Specifically, the inspection host device obtains images taken by the image acquisition device at each preset point as calibration images for each preset point, and obtains the point information of each calibration image. At the same time, for each calibration image, the inspection host device calibrates the calibration image using an automatic calibration algorithm, thereby obtaining the calibration information of each calibration image. Based on this, the inspection device realizes the corresponding recording of the calibration image, the calibration information of the calibration image, and the point information of the calibration image.

[0020] Therefore, based on the corresponding recorded calibration pictures, calibration information, and point information, in actual application, after the inspection host device obtains the target meter picture and the target point information of the target meter picture, it can determine the calibration information and calibration picture corresponding to the target point information. The calibration information and calibration picture are used to indicate the inspection host device to use recognition rules when performing meter recognition on the meter picture taken at the corresponding point information. Therefore, by performing meter recognition on the target meter picture according to the recognition rules indicated by the target calibration information and the target calibration picture, the meter recognition result of the target meter picture can be obtained, thereby realizing automatic recognition of the meter.

[0021] It can be understood that in the present application, the calibration information of the calibration image is determined by an automatic calibration algorithm, thereby realizing the corresponding storage of the calibration information, calibration image and point information. Based on this, the calibration image and calibration information corresponding to the target meter image are determined by the point information, that is, the target calibration image and target calibration information. The target meter can be identified based on the recognition rules indicated by the target calibration image and the target calibration information. There is no need to manually set corresponding meter recognition rules for each type of meter. Therefore, no matter how many types and quantities of meters there are in the scene, meter recognition can be achieved more efficiently.

[0022] Of course, it is not necessary to achieve all the advantages described above at the same time when implementing any product or method of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other embodiments can also be obtained based on these drawings.

[0024] Figure 1 A schematic diagram of a flow chart of a meter identification method provided in an embodiment of the present application;

[0025] Figure 2A schematic diagram of a first process for determining a corresponding relationship provided in an embodiment of the present application;

[0026] Figure 3 A schematic diagram of a second process for determining a corresponding relationship provided in an embodiment of the present application;

[0027] Figure 4 A schematic diagram of a calibration information input interface provided in an embodiment of the present application;

[0028] Figure 5 A schematic diagram of a third process for determining a corresponding relationship provided in an embodiment of the present application;

[0029] Figure 6 A fourth flow chart of determining a corresponding relationship provided in an embodiment of the present application;

[0030] Figure 7 A schematic diagram of a process for registering a calibration image provided in an embodiment of the present application;

[0031] Figure 8 A schematic diagram of a meter identification system provided in an embodiment of the present application;

[0032] Figure 9 A schematic diagram of the process interaction of the automated calibration provided for the implementation of this application;

[0033] Figure 10 A schematic diagram of a process interaction for template registration provided in an embodiment of the present application;

[0034] Figure 11 A schematic diagram of the process interaction of the meter identification method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field based on this application are within the scope of protection of this application.

[0036] The meter identification method provided in the embodiment of the present application is described below with a specific example.

[0037] See also Figure 1 , Figure 1 A flow chart of a meter identification method provided in an embodiment of the present application includes the following steps:

[0038] Step S11: Acquire a target meter image and target point information of the target meter image;

[0039] The target point information is the point where the image acquisition device takes the target meter image;

[0040] Step S12: determining the calibration information and calibration image corresponding to the target point information through the correspondence between the calibration image, the calibration information, and the point information, as the target calibration information and the target calibration image;

[0041] The calibration information and calibration images are used to indicate the recognition rules for the inspection host device when performing meter recognition on the meter images taken at the corresponding point information;

[0042] Step S13: performing meter recognition on the target meter image according to the recognition rules indicated by the target calibration information and the target calibration image, and obtaining a meter recognition result of the target meter image as a first meter recognition result.

[0043] Among them, the correspondence between at least one set of calibration pictures, calibration information and point information is established in the following manner: obtaining pictures taken by the image acquisition device at each preset point as the calibration pictures of each preset point; and obtaining the point information of each calibration picture; wherein the point information of the calibration picture is used to indicate the point at which the image acquisition device was located when taking the calibration picture; for each calibration picture, calibrating the calibration picture using an automatic calibration algorithm to obtain the calibration information of the calibration picture; and correspondingly recording the calibration picture, the calibration information of the calibration picture, and the point information of the calibration picture.

[0044] In the technical solution provided by the embodiments of this application, a correspondence is established between calibration images, calibration information, and point information. Specifically, the inspection host device obtains images taken by the image acquisition device at each preset point as calibration images for each preset point, and obtains the point information of each calibration image. At the same time, for each calibration image, the inspection host device calibrates the calibration image using an automatic calibration algorithm, thereby obtaining the calibration information of each calibration image. Based on this, the inspection device realizes the corresponding recording of the calibration image, the calibration information of the calibration image, and the point information of the calibration image.

[0045] Therefore, based on the corresponding recorded calibration pictures, calibration information, and point information, in actual application, after the inspection host device obtains the target meter picture and the target point information of the target meter picture, it can determine the calibration information and calibration picture corresponding to the target point information. The calibration information and calibration picture are used to indicate the inspection host device to use recognition rules when performing meter recognition on the meter picture taken at the corresponding point information. Therefore, by performing meter recognition on the target meter picture according to the recognition rules indicated by the target calibration information and the target calibration picture, the meter recognition result of the target meter picture can be obtained, thereby realizing automatic recognition of the meter.

[0046] It can be understood that in the present application, the calibration information of the calibration image is determined by an automatic calibration algorithm, thereby realizing the corresponding storage of the calibration information, calibration image and point information. Based on this, the calibration image and calibration information corresponding to the target meter image are determined by the point information, that is, the target calibration image and target calibration information. The target meter can be identified based on the recognition rules indicated by the target calibration image and the target calibration information. There is no need to manually set corresponding meter recognition rules for each type of meter. Therefore, no matter how many types and quantities of meters there are in the scene, meter recognition can be achieved more efficiently.

[0047] In step S11, the image acquisition device is a pre-installed device used to capture meter images. A meter image is an image of a meter captured by the image acquisition device. The target meter image is the image of the meter for which meter identification is required, i.e., the meter reading within the image needs to be determined. The point information indicates the point at which the image acquisition device captured the meter image. The target point information indicates the point at which the image acquisition device captured the target meter image. The point can be understood as the position of the image acquisition device.

[0048] Regarding the point information, specifically, the point information may include the point where the image acquisition device is located when taking the meter image, and may also include its own device identification, wherein the device identification is used to identify the uniqueness of the image acquisition device.

[0049] It can be understood that when the image acquisition device takes a picture of the meter, it can obtain its own point position. Furthermore, when the image acquisition device sends the meter picture to the inspection host device, it can send the point position obtained when taking the meter picture to the inspection host device. Alternatively, when the image acquisition device takes a picture of the meter, it can obtain its own point position. Furthermore, when the image acquisition device sends the meter picture to the inspection host device, it can send the point position obtained when taking the meter picture and its own device identification to the inspection host device. Specifically, the inspection device in this embodiment obtains the target meter picture and the point position information of the target meter picture, which can be understood as the inspection host device obtaining the target meter picture sent by the image acquisition device and the point position information of the target meter picture sent by the image acquisition device.

[0050] In addition, the point where the image acquisition device captures the meter image is a pre-set point (i.e., a preset point). It is understandable that the user knows where the meter is located in the scene, and when deploying the image acquisition device, the image acquisition device is pre-set in which position it can capture the meter image. In this way, each time the image acquisition device runs to the preset point, it can capture the corresponding meter image. Based on this, the point where the image acquisition device is located can be directly represented in the form of coordinates. For example, when the image acquisition device captures the meter image, it obtains the coordinates of its own position in real time, and uses the coordinates of its own position as the point where the meter image was captured. Alternatively, the point where the image acquisition device is located can also be represented in the form of an identifier. For example, a correspondence between each identifier and position is pre-established. When the image acquisition device captures the meter image, it obtains its own position in real time, and then determines the identifier corresponding to its own position, and uses the determined identifier to represent the point where the meter image was captured.

[0051] In the case where the point information includes the point at which the image acquisition device itself is located and the device identifier of the image acquisition device, each point under the same image acquisition device is unique, while each point under different image acquisition devices may not be unique. Specifically, if the point is represented in the form of an identifier, then for the same image acquisition device, the identifier of each point under the image acquisition device is unique, while for different image acquisition devices, the identifier of each point under different image acquisition devices may not be unique. For example, for the same image acquisition device, the identifier of each point under the image acquisition device can be set as point identifier 1, point identifier 2, and point identifier 3, and each point identifier is unique; while for different image acquisition devices, for example, for image acquisition device 1, the identifier of each point under image acquisition device 1 can be set as point identifier 1, point identifier 2, and point identifier 3; for image acquisition device 2, the identifier of each point under image acquisition device 2 can also be set as point identifier 1, point identifier 2, and point identifier 3. That is, when there are multiple image acquisition devices, only the identifiers of the points under the same image acquisition device may be kept unique.

[0052] In the case where the point information is the point where the image acquisition device itself is located, each point under the same image acquisition device is unique, and each point under different image acquisition devices is also unique.

[0053] Taking the use of identifiers to represent points as an example, in this application, the uniqueness of each point under the same image acquisition device means that the identifiers used to represent each point under the same image acquisition device are different; the uniqueness of each point under different image acquisition devices means that the identifiers used to represent each point between different image acquisition devices are different.

[0054] In the above step S12, after the inspection host device obtains the target calibration image and the target point information, it determines the point information consistent with the target point information through the correspondence between the calibration image, calibration information and point information, and then determines the calibration image and calibration information that have a corresponding relationship with the point information. Therefore, the determined calibration image is the target calibration image, that is, the calibration image corresponding to the target point information; the determined calibration information is the target calibration information, that is, the calibration information corresponding to the target point information.

[0055] It can be understood that the calibration information and calibration images are used to indicate the recognition rules used by the inspection host device when performing meter recognition on the meter images captured at the corresponding point information. In other words, the calibration information and calibration images corresponding to the point information are used to indicate the recognition rules for the meter images captured by the image acquisition device when the image acquisition device is at the point information. Therefore, after obtaining the calibration information and calibration images corresponding to the target point information, the target meter image can be recognized according to the recognition rules indicated by the calibration information and calibration images corresponding to the target point information.

[0056] To illustrate with a specific example, for example, the correspondence package is: (point information 1, calibration picture 1, calibration information 1), (point information 2, calibration picture 2, calibration information 2), where calibration picture 1 and calibration information 1 are used to indicate the recognition rules for the meter picture taken at point information 1; calibration picture 2 and calibration information 2 are used to indicate the recognition rules for the meter picture taken at point information 2. Based on this, if the target meter picture obtained by the inspection host device is meter picture 3, and the target point information is point information 1, then it can be determined that the calibration picture corresponding to point information 1 is calibration picture 1, that is, calibration picture 1 is the target calibration picture, and the calibration information corresponding to point information 1 is calibration information 1. Then, according to the recognition rules indicated by calibration picture 1 and calibration information 1, the meter recognition of meter picture 3 can be achieved.

[0057] Among them, the correspondence between at least one set of calibration pictures, calibration information and point information is established in the following way, specifically: the inspection host device obtains the pictures taken by the image acquisition device at each preset point as the calibration pictures of each preset point; and obtains the point information of each calibration picture; then, for each calibration picture, the inspection host device calibrates the calibration picture using an automatic calibration algorithm to obtain the calibration information of the calibration picture, and correspondingly records the calibration picture, the calibration information of the calibration picture, and the point information of the calibration picture. For details, please refer to the following Figure 2-Figure 4 The description in , will not be repeated here.

[0058] In step S13, the first meter recognition result is a meter recognition result of the target meter image obtained by performing meter recognition on the target meter image according to the recognition rules indicated by the target calibration information and the target calibration image. Specifically, in some embodiments, the first meter recognition result may be a meter reading in the target meter image.

[0059] Alternatively, in other embodiments, the first meter recognition result may also be a picture for representing the reading result. Specifically, the picture may be a processed target meter picture, the picture including a marking frame for marking the meter portion in the picture, and the reading result for the meter. When the meter in the target meter picture is a pointer meter, the picture may also include annotations for the reading represented by the main scale of the meter. The marking frame for marking the meter portion in the picture may be referred to below. Figure 4 The dashed box shown in the figure; for notes on the readings represented by the main scales in the meter, see the following Figure 4 0, 0.5, 1, 1.5, 2, 2.5, 3 marked in the box.

[0060] For the correspondence between the calibration image, calibration information and point information mentioned in the above step S12, at least one set of correspondence is obtained by the following method: Figure 2 The provided method is implemented. Figure 2 , Figure 2 The first flow chart of determining the corresponding relationship provided in the embodiment of the present application includes the following steps:

[0061] Step S21: obtaining images taken by an image acquisition device at each preset point as calibration images of each preset point; and obtaining point information of each calibration image;

[0062] The point information of the calibration image is used to indicate the point where the image acquisition device was located when taking the calibration image;

[0063] Step S22: for each calibration image, calibrate the calibration image using an automatic calibration algorithm to obtain calibration information of the calibration image; and correspondingly record the calibration image, the calibration information of the calibration image, and the point information of the calibration image.

[0064] The technical solution provided in the embodiment of the present application provides a method for determining the calibration information of a calibration image, so as to realize the corresponding recording of the calibration image, calibration information and point information, and thus realize the automatic identification of the meter in the target meter image.

[0065] In the above step S21, the calibration image is a meter image taken by the image acquisition device at a preset point. The preset point is a pre-set point where the image acquisition device takes the meter image. For a description of the preset point, please refer to the content in the above step S11.

[0066] In the embodiment of the present application, images captured by the image acquisition device at each preset location are obtained as calibration images for each preset location, and the location information of the calibration images is obtained. That is, for each preset location, a meter image captured at the preset location is obtained as the calibration image for the preset location. Furthermore, the inspection host device also obtains the location information of each calibration image when it captures the calibration image. The location information of the calibration image indicates the location at which the image acquisition device captured the calibration image.

[0067] In some embodiments, pictures taken by the image acquisition device at each preset point are obtained as calibration pictures of each preset point, and point information of the calibration pictures is obtained. Specifically, pictures taken at each preset point and sent by the image acquisition device are obtained as calibration pictures of each preset point, and point information of each calibration picture sent by the image acquisition device is obtained.

[0068] In the above step S22, for each calibration image, the inspection host device calibrates each calibration image using an automatic calibration algorithm, thereby obtaining calibration information of each calibration image, and then can correspondingly record the calibration image, the calibration information of the calibration image, and the point information of the calibration image.

[0069] It is understandable that there is at least one set of correspondences between calibration images, calibration information and point information. Figure 2 Of course, all the correspondences between the calibration images, calibration information and point information can be realized by the above Figure 2 The method shown is implemented.

[0070] Below, the correspondence between all calibration images, calibration information and point information is given above. Figure 2 Taking the implementation method shown as an example, the contents indicated by the above steps S21 and S22 are explained through specific examples.

[0071] For example, the preset points are point 1, point 2, and point 3; the inspection equipment needs to obtain the meter picture taken by the image acquisition device at point 1, as the calibration picture of point 1, recorded as calibration picture 1, and obtain the point information when taking calibration picture 1, which is point information 1; obtain the meter picture taken by the image acquisition device at point 2, as the calibration picture of point 2, recorded as calibration picture 2, and obtain the point information when taking calibration picture 2, which is point information 2; obtain the meter picture taken by the image acquisition device at point 3, as the calibration picture of point 3, recorded as calibration picture 3, and obtain the point information when taking calibration picture 3, which is point information 3. Then, the automatic calibration algorithm is used to calibrate calibration image 1, obtaining calibration information for calibration image 1, recorded as calibration information 1. The automatic calibration algorithm is used to calibrate calibration image 2, obtaining calibration information for calibration image 2, recorded as calibration information 2. The automatic calibration algorithm is used to calibrate calibration image 3, obtaining calibration information for calibration image 3, recorded as calibration information 3. Based on this, calibration image 1, calibration information 1, and point information 1 are recorded accordingly; calibration image 2, point information 2, and calibration information 2 are recorded accordingly; and calibration image 3, point information 3, and calibration information 3 are recorded accordingly.

[0072] It is understandable that in some embodiments, when only part of the correspondence between the calibration image, calibration information and point information is determined by the above Figure 2 When the method shown is implemented, another part of the corresponding relationship can be achieved by manually calibrating the calibration information of the calibration image, thereby achieving the corresponding relationship between the calibration image, calibration information, and point information.

[0073] In some embodiments, for the correspondence between the calibration image, calibration information, and point information mentioned in step S12, at least one set of correspondence is obtained through the above Figure 2 The automatic calibration method shown is implemented, and at least one set of corresponding relationships is achieved through the following Figure 3 The automatic calibration method shown is implemented. Figure 3 , Figure 3 The second flow chart of determining the corresponding relationship provided in the embodiment of the present application may include the following steps:

[0074] Step S31: obtaining images taken by an image acquisition device at each preset point as calibration images of each preset point; and obtaining point information of each calibration image;

[0075] The point information of the calibration image is used to indicate the point where the image acquisition device was located when taking the calibration image;

[0076] Step S32: at least one set of correspondence is achieved by: for each calibration image, calibrating the calibration image using an automatic calibration algorithm to obtain calibration information of the calibration image; and correspondingly recording the calibration image, the calibration information of the calibration image, and the point information of the calibration image.

[0077] Step S33: at least one set of correspondence is achieved by: for each calibration image, receiving calibration information input for the calibration image, obtaining calibration information of the calibration image; and correspondingly recording the calibration image, the calibration information of the calibration image, and the point information of the calibration image.

[0078] The technical solution provided in the embodiment of the present application provides another method for determining the calibration information of the calibration image, so as to realize the corresponding recording of the calibration image, calibration information and point information, and thus realize the automatic identification of the meter in the target meter image.

[0079] The above step S31 is the same as step S21 and will not be repeated here.

[0080] In the above step S32, for a part of the correspondence, it is implemented in the manner shown in the above step S22. It can be understood that since step S32 only determines a part of the correspondence, the calibration image mentioned in step S32 is the calibration image that needs to use step S32 to determine the correspondence.

[0081] In the above step S33, for another part of the corresponding relationships, such as the corresponding relationships other than the corresponding relationships determined in step S22, the method shown in step S33 is used.

[0082] Specifically, for each calibration image, the calibration information input for the calibration image is received to obtain the calibration information of the calibration image, and the calibration image, the calibration information of the calibration image, and the point information of the calibration image are recorded accordingly. It can be understood that since step S33 only determines a part of the corresponding relationship, the calibration image mentioned in step S33 is the image that needs to use step S33 to determine the corresponding relationship. Among them, for each calibration image, receiving the calibration information input for the calibration image to obtain the calibration information of the calibration image can be understood as, for each calibration image, receiving the calibration information input by the user for each calibration image, and obtaining the calibration information of each calibration image.

[0083] In some embodiments, see Figure 4 , Figure 4 A schematic diagram of the calibration information input interface provided in an embodiment of the present application. In this interface, the meter images 1-5 on the left side of the interface can be understood as the calibration images whose corresponding relationships need to be determined according to step S33. The words "to be confirmed" are marked on the meter images 1-5, which indicates that the user has not yet entered calibration information for these images. Figure 4 The middle of the interface is used to display the calibration picture where the calibration information is being input. For this picture, the user needs to mark the key parts that need to be identified, such as the parts that are helpful for reading the meter reading results. Specifically, Figure 4 As shown in the figure, mark the dial and pointer of the pointer meter in the calibration picture in the form of a dotted frame. In addition, the user needs to mark the reading of the scale indicated by the important pointer on the picture, such as Figure 4 0, 0.5, 1, 1.5, 2, 2.5, 3 are marked in the box. In addition, the user needs to mark the reading results of the meter in the picture on the picture, such as Figure 4 0.367 marked above the dashed box.

[0084] Figure 4 The left side of the interface is set with "Confirmation Status", "Calibration Set ID", "Point Name", "Remarks", "Meter Type", "Subtype", "Location Box", "Meter Name", and "Scale" from top to bottom. Users can Figure 4 Enter the corresponding content in the input box below the words set on the left side of the interface shown, such as "to be confirmed" in the box below "Confirmation Status", "2" in the box below "Calibration Set ID", "2" in the box below "Point Name", and "Note Information" in the input box. You can enter or not enter, enter "Pointer Type" in the box below "Meter Type", and enter "Reading Pointer" in the box below "Sub Type". Alternatively, you can also Figure 4 The words set on the left side of the interface are used to calibrate the calibration image. For example, the user can click Figure 4 The control corresponding to the "position box" in the interface enters the area selection state, and then the user selects the area to be selected in the picture, and the area selected by the user is marked in the picture, such as Figure 4 The area selected by the user is marked in the form of a dotted box, and the content in the dotted box is the area selected by the user. For example, when the user clicks Figure 4 The control corresponding to "Scale" in the interface enters the scale annotation state and annotates the readings of the main scales in the meter, such as Figure 4 The values ​​marked in the boxes are 0, 0.5, 1, 1.5, 2, 2.5, and 3. The calibration set ID (Identifier) ​​is an identifier used to indicate the uniqueness of the calibration image.

[0085] Based on the above Figure 2 The embodiment of the present application also provides a method for determining the corresponding relationship, see Figure 5 , Figure 5The third flow diagram for determining the corresponding relationship provided in the embodiment of the present application may include the following steps:

[0086] Step S51: obtaining images taken by an image acquisition device at each preset point as calibration images of each preset point; and obtaining point information of each calibration image;

[0087] The point information of the calibration image is used to indicate the point where the image acquisition device was located when taking the calibration image;

[0088] Step S52: for each calibration image, calibrate the calibration image using an automatic calibration algorithm to obtain calibration information of the calibration image; and correspondingly record the calibration image, the calibration information of the calibration image, and the point information of the calibration image;

[0089] Step S53: for each calibration image, perform meter recognition on the calibration image according to the calibration information corresponding to the calibration image and the recognition rule indicated by the calibration image, and obtain a meter recognition result of the calibration image as the second meter recognition result;

[0090] Step S54: for the unqualified calibration picture whose accuracy of the second meter recognition result is lower than the preset threshold, receiving the calibration information input for the target calibration picture as new calibration information;

[0091] Step S55: updating the calibration information corresponding to the uncalibrated image to the new calibration information.

[0092] The technical solution provided in the embodiment of this application is based on the above Figure 2 The examples shown are expanded, and it can be understood that the above Figure 2In the example shown, the calibration image is calibrated using an automatic calibration algorithm to obtain calibration information for the calibration image. However, in scenarios with a large number of meter types, the automatic calibration algorithm may not necessarily accurately obtain calibration information for each calibration image. Therefore, in the technical solution provided in the embodiment of the present application, after recording the calibration map, calibration information, and point information, the calibration image is subjected to meter recognition for each calibration image according to the calibration information corresponding to the calibration image and the recognition rules indicated by the calibration image, obtaining a meter recognition result as the second meter recognition result. Furthermore, a substandard calibration image is identified for which the accuracy of the second meter recognition result is lower than a preset threshold, i.e., a calibration image is identified for which the corresponding calibration information may be abnormal. Thus, for the substandard calibration image, the calibration information input for the substandard calibration image is received as new calibration information, and the calibration information corresponding to the substandard calibration image is updated to the new calibration information. In this way, it is possible to make up for the problem that the current automatic calibration algorithm may not be able to accurately calibrate certain calibration images, thereby causing inaccurate calibration information. That is, based on the technical solution provided in the embodiment of the present application, it is possible to update the calibration information that may be erroneous, thereby improving the accuracy of meter recognition.

[0093] The above steps S51 and S52 are the same as steps S21 and S22, and will not be repeated here.

[0094] In the above step S53 , the second meter recognition result is a meter recognition result of the calibration picture obtained after performing meter recognition on the calibration picture according to the calibration information corresponding to the calibration picture and the recognition rule indicated by the calibration picture.

[0095] In the embodiment of the present application, for each calibration image, meter recognition is performed on each calibration image according to the calibration information corresponding to each calibration image and the recognition rules indicated by the calibration image, and the meter recognition result of each calibration image is obtained as the second meter recognition result of each calibration image.

[0096] To illustrate with a specific example, for example, based on step S52 above, the corresponding relationships determined are (calibration image 1, calibration information 1, point information 1), (calibration image 2, calibration information 2, point information 2), and (calibration image 3, calibration information 3, point information 3). Then, meter recognition is performed on calibration image 1 according to the recognition rules indicated by calibration information 1 and calibration image 1, obtaining a meter recognition result for calibration image 1, recorded as meter recognition result 1. Furthermore, meter recognition is performed on calibration image 2 according to the recognition rules indicated by calibration information 2 and calibration image 2, obtaining a meter recognition result for calibration image 2, recorded as meter recognition result 2. Meter recognition is performed on calibration image 3 according to the recognition rules indicated by calibration information 3 and calibration image 3, obtaining a meter recognition result for calibration image 3, recorded as meter recognition result 3. Meter recognition result 1, meter recognition result 2, and meter recognition result 3 are all second meter recognition results.

[0097] In the above step S54, the uncalibrated picture is a picture whose second meter recognition result accuracy is lower than a preset threshold. Specifically, if the accuracy of the second meter recognition result of a certain picture is lower than the threshold, the picture is an uncalibrated picture.

[0098] Regarding how to determine whether the accuracy of the second meter identification result is lower than a preset threshold, in some real-time examples, the user can set the accuracy of the second meter identification result to be output during the output of the second meter identification result, and pre-set the threshold. Thus, whether the accuracy of the second meter identification result is lower than the threshold can be determined based on the accuracy of the output second meter identification result and the preset threshold. Alternatively, in other embodiments, the user can determine whether the second meter identification result of each calibration result is correct one by one. If the inspection host device receives a user-input instruction indicating that the second meter identification result is correct, it indicates that the accuracy of the second identification result is not lower than the preset threshold; if the inspection host device receives a user-input instruction indicating that the second meter identification result is incorrect, it indicates that the accuracy of the second identification result is lower than the preset threshold.

[0099] For the calibration picture that does not meet the standard, the inspection host device receives the calibration information input for the calibration picture that does not meet the standard as new calibration information. For example, continuing with the example shown in the above step S53, it is determined that the second meter recognition result of calibration picture 1 is lower than the preset threshold, then calibration picture 1 is a calibration picture that does not meet the standard. As a result, the inspection host device receives the calibration information input for calibration picture 1, and obtains new calibration information for calibration picture 1, which is recorded as calibration information 4. The way in which the inspection host device receives the calibration information input for the calibration picture that does not meet the standard is the same as the way of determining the calibration information shown in the above step S33, which can be referred to the above steps S33 and Figure 4 The description in , will not be repeated here.

[0100] In the above step S55, the calibration information corresponding to the non-standard calibration picture is updated to the new calibration information. That is, after the calibration picture, calibration information, and point information are recorded correspondingly based on the above steps S51-S52, if the non-standard calibration picture and the calibration information of the non-standard calibration picture are determined based on the subsequent steps S53-S54, the calibration information of the non-standard calibration picture in the previously recorded correspondence is updated to the new calibration information. Continuing with the example shown in the above steps S53 and S54, the calibration information 1 corresponding to calibration picture 1 is updated to calibration information 4. After the update, the corresponding recorded correspondences are (calibration picture 1, calibration information 1, point information 4), (calibration picture 2, calibration information 2, point information 2), and (calibration picture 3, calibration information 3, point information 3).

[0101] In some embodiments, after updating the calibration information corresponding to the substandard calibration image based on step S55, the updated calibration information and the recognition rules indicated by the calibration image can be used again to perform meter recognition on the calibration image, and determine whether the accuracy of the meter recognition result of the calibration image is lower than a preset threshold. If it is lower than the preset threshold, the calibration image is determined to be substandard, and the user can again correct the calibration information input in step S54 to obtain new calibration information. The calibration information corresponding to the substandard calibration image is then updated with the new calibration information. In this way, the accuracy of the updated calibration information is re-determined. If it is inaccurate, further adjustments are made until the calibration information is accurate, thereby improving the accuracy of meter recognition.

[0102] In some embodiments, it can be understood that for the above Figure 5In the example shown, the meter recognition result of the calibration image is obtained by performing meter recognition based on the calibration information of the calibration image and the recognition rules indicated by the calibration image. Therefore, if the accuracy of the meter recognition result of the calibration image is lower than the preset threshold, the calibration information may be abnormal. The calibration information is determined according to the automatic calibration algorithm. Therefore, if the calibration information of a calibration image is abnormal, it indicates that the current automatic calibration algorithm may not be good at recognizing the meter in the calibration image. Based on this, the embodiment of the present application also provides a technical solution. Specifically, the current automatic calibration algorithm is trained with the substandard calibration image determined in the above step S54 and the new calibration information to obtain an updated automatic calibration algorithm as the first automatic calibration algorithm. That is, the substandard calibration image is used as a sample and the new calibration information is used as the true value to train the current automatic calibration algorithm to obtain the first automatic calibration algorithm. Therefore, when step S22 is subsequently executed, if there are meters in the calibration image that are the same as or similar to the meters in the non-standard calibration image, the first automatic algorithm can calibrate this type of meter more accurately, that is, more accurately calibrate the calibration image containing this type of meter. In this way, the calibration ability of the current automatic calibration algorithm for calibration images that it is not good at can be improved, the accuracy of the calibration information can be improved, and the accuracy of meter identification can be improved.

[0103] Based on the above Figure 3 The embodiment of the present application also provides a method for determining the corresponding relationship, see Figure 6 , Figure 6 The fourth flow diagram for determining the corresponding relationship provided in the embodiment of the present application may include the following steps:

[0104] Step S61: obtaining images taken by an image acquisition device at each preset point as calibration images of each preset point; and obtaining point information of each calibration image;

[0105] The point information of the calibration image is used to indicate the point where the image acquisition device was located when taking the calibration image;

[0106] Step S62: at least one set of correspondences is achieved by: for each calibration image, calibrating the calibration image using an automatic calibration algorithm to obtain calibration information of the calibration image; and correspondingly recording the calibration image, the calibration information of the calibration image, and the point information of the calibration image;

[0107] Step S63: at least one set of correspondence relationships is achieved by: for each calibration image, receiving calibration information input for the calibration image, obtaining calibration information of the calibration image; and correspondingly recording the calibration image, the calibration information of the calibration image, and the point information of the calibration image;

[0108] Step S64: for each calibration image, identify the calibration image according to the calibration information corresponding to the calibration image and the identification rule indicated by the calibration image, and obtain a meter identification result of the calibration image as the second meter identification result;

[0109] Step S65: For the substandard calibration image whose accuracy of the second meter recognition result is lower than the preset threshold, if the calibration information corresponding to the substandard calibration image is obtained through input, the input correction of the calibration information of the substandard calibration image is received to obtain new calibration information; if the calibration information corresponding to the substandard calibration image is obtained through an automatic calibration algorithm, the input calibration information for the substandard calibration image is received to obtain new calibration information;

[0110] Step S66: updating the calibration information corresponding to the uncalibrated image to the new calibration information.

[0111] The technical solution provided in the embodiment of the present application shows a method for correcting the calibration information according to the calibration method of the calibration information when there is an abnormality in the calibration information.

[0112] The above steps S61 to S63 are the same as steps S31 to S32, step S64 is the same as step S53, and step S66 is the same as step S55, and they will not be repeated here.

[0113] In the above step S65, the method for determining the unqualified calibration image is the same as the method for determining the unqualified calibration image in the above step S54, and will not be repeated here.

[0114] For each non-standard calibration image, if the calibration information corresponding to the non-standard calibration image is obtained by input (i.e., obtained by the method shown in step S33 above), the input correction of the calibration information of the non-standard calibration image is received to obtain new calibration information. It can be understood that since the calibration information corresponding to the non-standard calibration image is obtained by user input, the user can check whether the previously input correction information is correct and correct any inaccuracies, thereby obtaining new calibration information for the non-standard calibration image.

[0115] At the same time, for the non-standard calibration image, if the calibration information corresponding to the non-standard calibration image is obtained by the automatic calibration algorithm (i.e., obtained by the method shown in step S32 above), the calibration information input for the non-standard calibration image is received to obtain new calibration information. It can be understood that in this case, the calibration information corresponding to the non-standard calibration image is obtained by the automatic calibration algorithm. Therefore, when the calibration information may be abnormal, it is no longer appropriate to calibrate again using the calibration algorithm to obtain new calibration information. The user needs to manually input the calibration information for the non-standard calibration image to obtain new calibration information for the target calibration image, thereby improving the accuracy of the calibration information and further improving the accuracy of meter recognition.

[0116] In some embodiments, the calibration information corresponding to the non-standard calibration picture is obtained by an automatic calibration algorithm, and the user needs to manually input the calibration information for the non-standard calibration picture to obtain new calibration information for the non-standard calibration picture. After executing step S66, that is, updating the calibration information corresponding to the non-standard calibration picture to the new calibration information, the non-standard calibration picture can be used as a sample and the new calibration information as the true value to train the current automatic calibration algorithm to obtain a new automatic calibration algorithm. Therefore, when step S22 is subsequently executed, the new automatic calibration algorithm can be used to calibrate each calibration picture to obtain the calibration results of each calibration picture. In this way, the calibration ability of the current automatic calibration algorithm for calibration pictures that it is not good at can be improved, the accuracy of the calibration information can be improved, and the accuracy of meter recognition can be improved.

[0117] In some embodiments, if the calibration information corresponding to the substandard calibration image is obtained by an automatic calibration algorithm, and the user needs to manually input the calibration information for the substandard calibration image to obtain new calibration information for the substandard calibration image, after executing step S66, that is, updating the calibration information corresponding to the substandard calibration image with the new calibration information, the calibration image can be again used to identify the meter using the updated calibration information and the recognition rules indicated by the calibration image, and it is determined whether the meter recognition result of the calibration image is lower than a preset threshold. If it is lower than the preset threshold, the calibration image is determined to be a substandard calibration image, and the user again corrects the calibration information input in step S65 to obtain new calibration information. The calibration information corresponding to the substandard calibration image is then updated with the new calibration information. In this way, it is again determined whether the updated calibration information is accurate. If it is inaccurate, it is further adjusted until the calibration information is accurate, thereby improving the accuracy of meter recognition.

[0118] In some embodiments, for the above Figure 2 In the example shown, before calibrating the calibration image using the automatic calibration algorithm, the automatic calibration algorithm can be updated in advance in the following ways, such as Figure 7 As shown, the following steps may be included:

[0119] Step S71: Obtain an unregistered calibration image;

[0120] Step S72: receiving calibration information input for the unregistered calibration image, and obtaining calibration information of the unregistered calibration image;

[0121] Step S73: correspondingly record the unregistered calibration picture, the calibration information of the unregistered calibration picture, and the point information of the unregistered calibration picture; and, train the automatic calibration algorithm according to the unregistered calibration picture and the calibration information of the unregistered calibration picture to obtain an updated automatic calibration algorithm as the second automatic calibration algorithm.

[0122] In the embodiment of the present application, an unregistered calibration image can be understood as a calibration image that cannot be accurately calibrated using the current automatic calibration algorithm. It is understandable that there are many different types of meters currently, and with the development of technology, new meters often appear. In this regard, in actual application, it is difficult to collect all meter types for training the automatic calibration algorithm. Therefore, it is difficult for the conventional automatic calibration algorithm to achieve accurate calibration of all meters. However, based on the technical solution provided in the embodiment of the present application, in actual application, if an unregistered calibration image is found, the calibration information of the calibration image can be obtained through manual calibration, that is, the calibration information of the unregistered calibration image, and then the unregistered calibration image and the calibration information of the unregistered calibration image are used to train the current automatic calibration algorithm. In this way, the second automatic calibration algorithm obtained after training can accurately calibrate calibration images of the same type as the unregistered calibration image, wherein the image of the same type as the unregistered calibration image means that the meter type in the image is the same or similar to the meter type in the unregistered calibration image. It can be seen that through the technical solution provided in the embodiment of the present application, by continuously training and updating the automatic calibration algorithm during actual application, the automatic calibration algorithm can be continuously enabled to accurately calibrate more meter types. Therefore, the automatic identification method provided by the present application is more applicable and can be used in any scenario that requires meter identification, especially for scenarios with more complex meter types.

[0123] In step S71 above, the inspection host device obtains an unregistered calibration image. In the embodiment of the present application, since there are calibration images whose calibration information cannot be accurately determined by the current automatic calibration algorithm, or the current automatic calibration algorithm does not support calibration images, these calibration images need to be manually calibrated and the calibration information, calibration images, and point information are recorded accordingly. This is equivalent to registering such calibration images. Therefore, such calibration images are referred to as unregistered calibration images.

[0124] In the above step S72, the inspection host device receives the calibration information input for the unregistered calibration image and obtains the calibration information of the unregistered calibration image, which is consistent with the above step S33 and Figure 4 The principle of inputting calibration information is the same as that of FIG, and only the calibration image in step S33 needs to be replaced with an unregistered calibration image.

[0125] In step S73, the unregistered calibration image, its calibration information, and its point information are recorded. Furthermore, the automatic calibration algorithm is trained using the unregistered calibration image as a sample and its calibration information as the ground truth, resulting in an updated automatic calibration algorithm, which serves as the second automatic calibration algorithm.

[0126] After the second automatic calibration algorithm is obtained, when step S22 is subsequently executed, each calibration image is calibrated using the second automatic calibration algorithm to obtain calibration information of the calibration image.

[0127] Understandably, the above Figure 7 The example shown can be executed only when there is an unregistered calibration picture. When there is no unregistered calibration picture, it may not be executed. Figure 7 The example shown.

[0128] To facilitate understanding of the technical solution provided in the embodiment of the present application, the meter identification method provided in the embodiment of the present application is described below with a specific example:

[0129] See also Figure 8 , Figure 8 A schematic diagram of a meter identification system provided in an embodiment of the present application. Figure 8 As shown, the meter recognition system includes a storage medium, a camera device, a platform, and a meter inspection host device. The storage medium can be a hard drive or an SSD (Solid State Drive). The camera device is an image acquisition device. The meter inspection host device is an inspection host device, which includes a meter smart application (Application), a storage subsystem, a computing management subsystem, an alarm subsystem, and an engine subsystem. The computing management subsystem includes algorithm package management, template synchronization, and task management. The engine subsystem includes multiple engines, such as Engine 1…Engine N.

[0130] The camera device can capture calibration images, target meter images, and point information through camera capture, and send the acquired calibration images, target meter images, and point information to the platform. The platform can send calibration images in batches to the meter smart APP (see step S21 above), submit automatic calibration tasks (see step S22 above), submit template registration tasks, submit algorithm pre-allocation, submit meter verification tasks (see step S23 above), and send the calibration images to the meter smart APP (see step S24 above). Figure 5 or Figure 6 Example shown), submitting meter analysis tasks (see above Figure 1 content shown).

[0131] The meter smart APP can manage the calibration library and template library stored in the storage subsystem. The calibration library is used to record the corresponding relationship between the calibration image, the calibration information of the calibration image, and the point information of the calibration image. The template library is used to record the corresponding relationship between the unregistered calibration image, the calibration information of the unregistered calibration image, and the point information of the unregistered calibration image. The corresponding relationship can be confirmed by referring to the above. Figure 2 or Figure 3 The storage subsystem can interact with the storage medium and store the calibration library and template library in the storage medium.

[0132] The meter smart app can query algorithm information and task execution status through the calculation management subsystem and issue images and tasks to the calculation management subsystem. These include issuing calibration images and target meter images, issuing meter recognition tasks or automatic calibration tasks, deleting tasks, and querying tasks. Furthermore, the meter smart app can interact with the calculation management subsystem to register templates, such as using unregistered calibration images and their calibration information to train the current automatic calibration algorithm, and deleting templates. Algorithm package management in the calculation management subsystem pre-allocates algorithms for engine management; template synchronization in the calculation management subsystem synchronizes template data with the engine management; and task management in the calculation management subsystem issues analysis images and tasks to the engine management. It also receives analysis results from the engine subsystem and sends them to the alarm subsystem. The alarm subsystem converts the analysis results into alarm messages according to the meter smart app's alarm subscription rules and sends them to the meter smart app for alarm notification.

[0133] The engine subsystem loads algorithms issued by the engine management in the calculation management subsystem, receives template data synchronized by the engine management, and receives issued image tasks and analysis images. Image tasks can be understood as recognition tasks for target meter images, and analysis images can be understood as target meter images.

[0134] Specifically, in some embodiments, manual calibration is achieved by the following steps: the calibration image is first manually calibrated, that is, the user inputs the calibration information of the calibration image to the inspection host device, for details, see the above steps S33 and Figure 4 After the calibration image is manually calibrated to obtain the calibration information, the reading verification is performed, that is, the calibration image is identified according to the calibration information corresponding to the calibration image and the identification rules indicated by the calibration image to determine whether the calibration image is a substandard calibration image. If it is a substandard calibration image, new calibration information needs to be determined. For details, please refer to the above Figure 6 Description in .

[0135] In some embodiments, automatic calibration is achieved through the following steps: the calibration image is first automatically calibrated, that is, the calibration information of the calibration image is determined by the automatic calibration algorithm, and the details can be found in the description of the above step S22. After the calibration image is automatically calibrated to obtain the calibration information, the meter reading is performed, that is, the calibration image is identified according to the calibration information corresponding to the calibration image and the recognition rules indicated by the calibration image, so as to obtain the second meter identification result. For details, please refer to the description of the above step S53. Afterwards, it is manually determined whether the second meter identification result of the calibration image is correct. If it is correct, it ends. If it is not correct, manual calibration is required. After manual calibration, the reading verification can be performed again, that is, the reading verification is performed again with the new calibration information. For details, please refer to the description of the above steps S54-step S55. Alternatively, you can also refer to Figure 6 Description in .

[0136] See also Figure 9 , Figure 9 A schematic diagram of an automated calibration process provided for the implementation of this application may include the following steps:

[0137] Step S1.1: The platform allocates algorithm resources.

[0138] It is understandable that before creating an automatic calibration task, the device algorithm engine resources may be pre-allocated, that is, the engine is pre-loaded with the algorithms required for operation.

[0139] Step S1.2: The platform sends the algorithm pre-allocation protocol to the meter APP.

[0140] Step S1.3: The meter APP sends the pre-allocation information modification to the calculation management. After the calculation management successfully modifies the information, it returns the modification success information to the meter APP. The meter APP responds to the platform in response to the modification success information.

[0141] Step S2.1: The platform adds / modifies calibration images;

[0142] Step S2.2: The platform sends a calibration information add protocol to the meter APP.

[0143] Step S2.3: The meter APP sends the calibration image to the data storage for storage.

[0144] Step S2.4: Data is stored for data persistence and an image URL is generated.

[0145] Step S2.5: The meter APP sends the calibration data to the data storage for structured storage.

[0146] Step S2.6: The data storage performs data persistence and returns a response to the meter APP, which then returns a response to the platform.

[0147] Step S3.1: The platform selects automatic calibration conditions.

[0148] Step S3.2: The platform creates an automatic calibration task to the meter APP.

[0149] Step S3.3: The meter APP creates an automatic calibration thread.

[0150] Step S3.4: The meter APP retrieves the calibration image to be calibrated in the calibration library.

[0151] Step S3.5: The meter APP sends a synchronization image analysis task to the calculation management.

[0152] Step S3.6: The meter APP synchronously waits for the calibration analysis results.

[0153] Step S3.7: The computing management notifies the engine to start the image task.

[0154] Step S3.8: The calculation management sends the image to the engine for intelligent analysis.

[0155] Step S3.9: The calculation management sends the calibration analysis result output to the alarm management.

[0156] Step S3.10: The alarm management broadcasts the analysis results.

[0157] Step S3.11: The meter APP subscribes to the alarm management analysis results by ID.

[0158] Step S3.12: The meter APP performs callback analysis and processing of the analysis results.

[0159] Step S3.13: The meter APP updates the calibration information of the calibration library to the data storage.

[0160] Step S3.14: Data storage for data persistence.

[0161] It can be understood that the above steps 3.5-3.14 are steps for calibrating the image to be calibrated, that is, steps for calibrating the image that needs to be automatically calibrated.

[0162] In some embodiments, template registration is achieved through the following steps: the calibration image obtains calibration information through automatic registration, and then the meter recognition of the calibration image is achieved through the calibration information and the calibration image, that is, the meter reading is achieved, and then the correctness of the meter reading is manually determined, that is, the unregistered calibration image is determined. If it is correct, it ends. If it is incorrect, it is manually calibrated to obtain new calibration information, and the unregistered calibration image, the calibration information of the unregistered calibration image, and the point information of the unregistered calibration image are recorded in the template library. In addition, the automatic calibration algorithm is trained according to the unregistered calibration image and the calibration information of the unregistered calibration image to obtain a new automatic calibration algorithm, and the new automatic calibration algorithm is subsequently used to calibrate the calibration image. For details, please refer to the above Figure 7 Description in .

[0163] See also Figure 10 , Figure 10 A schematic diagram of a template registration process interaction provided in an embodiment of the present application may include the following steps:

[0164] Step S1.1: The platform allocates algorithm resources;

[0165] It is understandable that before creating an automatic calibration task, the device algorithm engine resources can be pre-allocated, that is, the engine is pre-loaded with the algorithm required for operation;

[0166] Step S1.2: The platform sends the algorithm pre-allocation agreement to the meter APP;

[0167] Step S1.3: The meter APP sends the pre-allocation information modification to the calculation management. After the calculation management successfully modifies the information, it returns the modification success information to the meter APP. The meter APP responds to the platform in response to the modification success information.

[0168] Step S2.1: The platform adds / modifies calibration images;

[0169] Step S2.2: The platform sends a calibration information add protocol to the meter APP;

[0170] Step S2.3: The meter APP sends the calibration image to the data storage for storage;

[0171] Step S2.4: Data storage for data persistence and generating image URL;

[0172] Step S2.5: The meter APP sends the calibration data to the data storage for structured storage;

[0173] Step S2.6: The data storage performs data persistence and returns a response to the meter APP, which then returns a response to the platform.

[0174] Step S4.1: The platform selects the calibration image for template registration;

[0175] Step S4.2: The platform sends the template registration agreement to the meter APP;

[0176] Step S4.3: The meter APP queries the template library to see if there is an identical template. If so, registration is not allowed.

[0177] Step S4.4: The meter APP sends the template image to the data storage for transfer;

[0178] Step S4.5: Data storage is used to persist data and generate image URLs;

[0179] Step S4.6: The meter APP sends the synchronization image analysis task to the calculation management;

[0180] Step S4.7: The meter APP synchronously waits for the template analysis result;

[0181] Step S4.8: The table calculation manager notifies the engine to start the image task.

[0182] Step S4.9: The calculation management sends the image to the engine for intelligent analysis;

[0183] Step S4.10: The calculation management sends the template analysis results to the alarm management;

[0184] Step S4.11: The alarm management broadcasts the analysis results;

[0185] Step S4.12: The meter APP subscribes to the analysis results from the alarm management according to the ID;

[0186] Step S4.13: The meter APP performs callback analysis and processing of the analysis results.

[0187] Step S4.14: The meter APP updates the template library data information to the data storage.

[0188] Step S4.15: Data storage for data persistence;

[0189] Step S4.16: The meter APP synchronously sends the template data to the calculation management;

[0190] Step S4.17: The calculation management synchronizes the template data to the engine.

[0191] See also Figure 11 , Figure 11The meter provided in the embodiment of the present application is a schematic diagram of a process interaction of the algorithm, which may include the following steps:

[0192] Step S5.1: The platform allocates algorithm resources;

[0193] It is understandable that before creating an automatic calibration task, the device algorithm engine resources can be pre-allocated, that is, the engine is pre-loaded with the algorithm required for operation;

[0194] Step S5.2: The platform sends the algorithm pre-allocation agreement to the meter APP;

[0195] Step S5.3: The meter APP sends the pre-allocation information modification to the calculation management. After the calculation management successfully modifies the information, it returns the modification success information to the meter APP. The meter APP responds to the platform in response to the modification success information.

[0196] Step S5.4: The platform configures task parameters;

[0197] Step S5.5: The platform sends the image analysis task to the meter APP;

[0198] Step S5.6: The meter APP parses the task protocol message;

[0199] Step S5.7: The meter APP obtains the calibration information image rule message according to the calibration set ID;

[0200] Step S5.8: The meter APP sends the image analysis task to the calculation management;

[0201] Step S5.9: The computing management notifies the engine to start the image task;

[0202] Step S5.10: The computing manager sends the image to the engine for intelligent analysis;

[0203] Step S5.11: Calculation management sends alarm picture storage to data storage;

[0204] Step S5.12: Data management performs data persistence and generates image URLs;

[0205] Step S5.13: The calculation management sends the analysis results to the alarm management;

[0206] Step S5.14: The alarm management plays the analysis results.

[0207] Step S6.1: The meter APP subscribes to the analysis results from the alarm management according to the ID;

[0208] Step S6.2: The meter APP performs callback analysis of the analysis results;

[0209] Step S6.3: The meter APP sends the analysis results to the alarm management for structured storage;

[0210] Step S6.4: Data storage for data persistence;

[0211] Step S6.5: The meter APP obtains the destination address;

[0212] Step S6.6: The meter APP sends an alarm message to the alarm management;

[0213] Step S6.7: The alarm management sends the alarm message to the platform according to the destination address.

[0214] The meter recognition algorithm provided in the embodiments of the present application can achieve the following beneficial effects:

[0215] (1) The platform supports batch distribution of calibration images and the core automatic calibration function; specifically, by creating automatic calibration tasks, images can be automatically calibrated in batches, which can greatly improve calibration efficiency and accuracy.

[0216] (2) After manual or automatic calibration, the meter calibration and verification can be performed directly through the device platform interface to verify the calibration effect in advance, which can improve the accuracy and efficiency of meter analysis.

[0217] (3) By adding a new template registration function, meters that do not support automatic calibration in new scenarios can also be easily automatically calibrated through template registration, which makes them more adaptable to different scenarios.

[0218] An embodiment of the present application further provides an electronic device, comprising:

[0219] Memory for storing computer programs;

[0220] The processor is configured to implement any of the above-mentioned meter identification methods when executing a program stored in the memory.

[0221] Furthermore, the electronic device may further include a communication bus and / or a communication interface, and the processor, the communication interface, and the memory 161 communicate with each other via the communication bus.

[0222] The communication bus mentioned in the electronic device mentioned above may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. This communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is used in the figure, but this does not mean that there is only one bus or only one type of bus.

[0223] The communication interface is used for communication between the above electronic device and other devices.

[0224] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage. Alternatively, the memory may be at least one storage device located away from the processor.

[0225] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0226] In another embodiment provided by the present application, a computer-readable storage medium is further provided, wherein a computer program is stored in the computer-readable storage medium. When the computer program is executed by a processor, the steps of any of the above-mentioned meter identification methods are implemented.

[0227] In another embodiment provided by the present application, a computer program product including instructions is also provided, which, when executed on a computer, enables the computer to execute any meter identification method in the above embodiments.

[0228] In the above embodiments, all or part of the embodiments may be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part.

[0229] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0230] Each embodiment in this specification is described in a related manner. Similar portions between the various embodiments can be referenced to each other. Each embodiment focuses on the differences between the other embodiments. In particular, the electronic device, storage medium, and computer program product embodiments are generally similar to the method embodiments, so their descriptions are relatively simple. For related portions, reference can be made to the descriptions of the method embodiments.

[0231] The above description is only a preferred embodiment of the present application and is not intended to limit the scope of protection of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application are included in the scope of protection of the present application.

Claims

1. A meter identification method, characterized in that: Applied to inspection of host equipment, the method includes: Acquire a target meter image and target point information of the target meter image; wherein the target point information is used to indicate the point where the image acquisition device was located when capturing the target meter image; Based on the correspondence between the calibration image, the calibration information, and the point information, the calibration information and the calibration image corresponding to the target point information are determined as the target calibration information and the target calibration image; wherein the calibration information and the calibration image are used to instruct the inspection host device to perform recognition rules when performing meter recognition on the meter image taken at the corresponding point information; performing meter recognition on the target meter image according to the recognition rules indicated by the target calibration information and the target calibration image, and obtaining a meter recognition result of the target meter image as a first meter recognition result; The correspondence between at least one set of calibration images, calibration information, and point information is established in the following manner: Acquire images taken by an image acquisition device at each preset point as calibration images of each preset point; and acquire point information of each calibration image; wherein the point information of the calibration image is used to indicate the point at which the image acquisition device was located when taking the calibration image; For each calibration image, the calibration image is calibrated using an automatic calibration algorithm to obtain calibration information of the calibration image; and the calibration image, the calibration information of the calibration image, and the point information of the calibration image are correspondingly recorded.

2. The method according to claim 1, characterized in that The obtaining of the target meter image and the location information of the target meter image includes: Acquire a target meter image sent by the image acquisition device, and location information of the target meter image sent by the image acquisition device; The step of obtaining pictures taken by the picture acquisition device at each preset point as calibration pictures of each preset point, and obtaining point information of each calibration picture includes: The pictures taken at each preset point and sent by the image acquisition device are obtained as calibration pictures of each preset point; and the point information of each calibration picture sent by the image acquisition device is obtained.

3. The method according to claim 1, characterized in that The step of establishing the corresponding relationship further includes: For each calibration image, perform meter recognition on the calibration image according to the calibration information corresponding to the calibration image and the recognition rule indicated by the calibration image, and obtain a meter recognition result of the calibration image as a second meter recognition result; For a substandard calibration picture whose accuracy of the second meter recognition result is lower than a preset threshold, receiving calibration information input for the substandard calibration picture as new calibration information; The calibration information corresponding to the non-standard calibration image is updated to the new calibration information.

4. The method according to claim 3, characterized in that The method further comprises: Training the automatic calibration algorithm using the substandard calibration image and the new calibration information to obtain an updated automatic calibration algorithm as a first automatic calibration algorithm; The step of calibrating each calibration image by using an automatic calibration algorithm to obtain calibration information of the calibration image includes: For each calibration image, calibrate the calibration image using the first automatic calibration algorithm to obtain calibration information of the calibration image.

5. The method according to claim 1, characterized in that Before calibrating the calibration image using the automatic calibration algorithm, the automatic calibration algorithm is updated in advance in the following manner: Get unregistered calibration images; receiving calibration information input for the unregistered calibration image, and obtaining the calibration information of the unregistered calibration image; Correspondingly recording the unregistered calibration image, the calibration information of the unregistered calibration image, and the point information of the unregistered calibration image; and training the automatic calibration algorithm according to the unregistered calibration image and the calibration information of the unregistered calibration image to obtain an updated automatic calibration algorithm as a second automatic calibration algorithm; The step of calibrating each calibration image by using an automatic calibration algorithm to obtain calibration information of the calibration image includes: For each calibration image, the second automatic calibration algorithm is used to calibrate each calibration image to obtain calibration information of the calibration image.

6. The method according to claim 1, characterized in that The correspondence between at least one set of calibration images, calibration information, and point information is established in the following manner: Acquire images taken by an image acquisition device at each preset point as calibration images of each preset point; and acquire point information of each calibration image; wherein the point information of the calibration image is the point at which the image acquisition device was located when taking the calibration image; For each calibration picture, calibration information input for the calibration picture is received to obtain the calibration information of the calibration picture; and the calibration picture, the calibration information of the calibration picture, and the point information of the calibration picture are correspondingly recorded.

7. The method according to claim 6, characterized in that The step of establishing the corresponding relationship further includes: For each calibration image, identifying the calibration image according to the calibration information corresponding to the calibration image and the identification rule indicated by the calibration image, and obtaining a meter identification result of the calibration image as a second meter identification result; For a substandard calibration image whose accuracy of the second meter recognition result is lower than a preset threshold, if the calibration information corresponding to the substandard calibration image is obtained through input, receiving the input correction of the calibration information of the substandard calibration image to obtain new calibration information; if the calibration information corresponding to the substandard calibration image is obtained through an automatic calibration algorithm, receiving the input calibration information for the substandard calibration image to obtain new calibration information; The calibration information corresponding to the non-standard calibration image is updated to the new calibration information.

8. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the method according to any one of claims 1 to 7 when executing a program stored in a memory.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. A computer program product comprising instructions, characterized in that When the method is run on a computer, the computer is enabled to execute the method according to any one of claims 1 to 7.

Images