Instrument reading recognition method and device, electronic equipment and medium

Abstract

The invention discloses an instrument reading recognition method and device, electronic equipment and a medium. The instrument reading identification method comprises the following steps: obtaining a reading identification model obtained through training, wherein the reading identification model comprises a scale information identification module and a pointer information identification module; obtaining a to-be-recognized image; pointer information corresponding to a to-be-recognized image is obtained through a pointer information recognition module, scale information corresponding to the to-be-recognized image is obtained through a scale information recognition module, the scale information comprises scale position information and scale value information, and the scale information recognition module is a key point recognition model; and obtaining an instrument reading corresponding to the to-be-identified image through the scale information and the pointer information. According to the instrument reading identification method provided by the embodiment of the invention, a template drawing does not need to be configured, so that the consumption of manpower and material resources is reduced, and accurate reading identification of the instrument can be realized under the interference of a complex environment.

Description

Technical Field

[0001] The present application relates to the field of meter readings, and in particular to a meter reading recognition method, a recognition device, an electronic device and a medium. Background Art

[0002] The maintenance points of substations are widely distributed and the equipment is complex and diverse. Intelligent inspections based on cameras, robots, and drones are gradually replacing manual inspections. Since substations have a large number of pointer instruments for monitoring equipment, instrument readings have become one of the essential functions in the intelligent inspection system. The current large instrument reading solution is a semi-automatic solution that combines manual template configuration and deep learning models. It is relatively labor-intensive and resource-intensive. In addition, due to factors such as changes in ambient lighting and offset jitter of image acquisition equipment, it is easy to cause difficulty in aligning the template image with the image to be analyzed, resulting in reading errors. Summary of the invention

[0003] The present application provides an instrument reading recognition method, recognition device, electronic device and medium, which do not require configuration of a template diagram and can achieve accurate reading recognition of the instrument under interference from complex environments, thereby saving manpower and material resources.

[0004] In a first aspect, an embodiment of the present application provides an instrument reading recognition method, which includes: obtaining a trained reading recognition model, the reading recognition model including a scale information recognition module and a pointer information recognition module; obtaining an image to be recognized; obtaining pointer information corresponding to the image to be recognized through the pointer information recognition module, obtaining scale information corresponding to the image to be recognized through the scale information recognition module, the scale information including scale position information and scale value information, the scale information recognition module being a key point recognition model; obtaining the instrument reading corresponding to the image to be recognized through the scale information and the pointer information.

[0005] According to the aforementioned implementation manner of the first aspect of the present application, the method of obtaining the trained reading recognition model includes: obtaining a scale information recognition module by training a key point model, and the method of obtaining the scale information recognition module by training a key point model includes: obtaining a number of instrument image samples, and the instrument image samples include annotations of scale key points and scale value categories to obtain a first training atlas; based on the first training atlas, training the key point model to obtain the scale information recognition module, and the scale information recognition module is used to recognize instrument images to obtain corresponding scale position information and scale value information.

[0006] According to any of the aforementioned implementations of the first aspect of the present application, obtaining the scale information corresponding to the image to be identified through the scale information identification module includes: identifying the scale key points of the instrument image in the image to be identified through the scale information identification module to obtain scale position information; identifying the scale value category of the instrument image in the image to be identified through the scale information identification module to obtain scale value information.

[0007] According to any of the aforementioned implementations of the first aspect of the present application, the pointer information recognition module is a segmentation model, and obtaining the pointer information corresponding to the image to be identified through the pointer information recognition module includes: performing regional segmentation on the instrument image in the image to be identified through the pointer information recognition module to obtain pointer information corresponding to the pointer in the instrument image.

[0008] According to any of the aforementioned embodiments of the first aspect of the present application, the reading recognition model is configured to output the scale position information, the scale value and the information pointer information in multiple ways.

[0009] According to any of the aforementioned implementations of the first aspect of the present application, before the step of obtaining the pointer information corresponding to the image to be identified through the pointer information identification module and obtaining the scale information corresponding to the image to be identified through the scale information identification module, the instrument reading identification method also includes: obtaining a trained instrument detection model; identifying the image to be identified through the instrument detection model to obtain a first recognition result to determine whether the image to be identified contains an instrument image and whether the instrument image has defects.

[0010] According to any of the aforementioned implementations of the first aspect of the present application, obtaining the trained instrument detection model includes: obtaining a number of instrument image samples, the instrument image samples including annotations of whether they have defects, to obtain a second training atlas; based on the second training atlas, training the YOLO-detect model to obtain the instrument detection model.

[0011] According to any of the aforementioned implementations of the first aspect of the present application, the instrument reading recognition method also includes: if the first recognition result is that there is no instrument image in the image to be recognized, outputting first abnormal information; if the first recognition result is that there is an instrument image in the image to be recognized and the instrument image is defective, outputting second abnormal information.

[0012] According to any of the aforementioned implementations of the first aspect of the present application, obtaining the instrument reading corresponding to the image to be identified through the scale information and the pointer information includes: obtaining the center of the arc where the multiple scale key points are located according to the multiple scale key points in the scale information; obtaining the corresponding pointer pointing according to the pointer information; obtaining the instrument reading corresponding to the instrument image in the image to be identified according to the scale value information corresponding to each of the scale key points in the scale information, the center of the circle, and the pointer pointing.

[0013] In a second aspect, an embodiment of the present application provides an instrument reading recognition device, which includes: a reading recognition model training module, used to obtain a trained reading recognition model, the reading recognition model including a scale information recognition module and a pointer information recognition module; an image acquisition module, used to obtain an image to be recognized; a reading information recognition module, used to obtain pointer information corresponding to the image to be recognized through the pointer information recognition module, and obtain scale information corresponding to the image to be recognized through the scale information recognition module, the scale information including scale position information and scale value information, and the scale information recognition module is a key point recognition model; a reading calculation module, used to obtain the instrument reading corresponding to the image to be recognized through the scale information and the pointer information.

[0014] In a third aspect, an embodiment of the present application provides an electronic device, the electronic device comprising a memory and at least one processor, the memory being electrically connected to the at least one processor, the memory storing instructions, the at least one processor calling the instructions in the memory so that the electronic device executes the meter reading identification method according to any of the aforementioned implementations of the first aspect of the present application.

[0015] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement a meter reading identification method according to any of the aforementioned implementations of the first aspect of the present application.

[0016] According to the meter reading recognition method of the embodiment of the present application, a trained reading recognition model is obtained, and the reading recognition model includes a scale information recognition module and a pointer information recognition module. After obtaining the image to be recognized, the pointer information corresponding to the image to be recognized is obtained through the pointer information recognition module, and the scale information corresponding to the image to be recognized is obtained through the scale information recognition module. The scale information includes scale position information and scale value information. The scale information recognition module is a key point recognition model, so that the meter reading corresponding to the image to be recognized can be obtained through the scale information and the pointer information. The above-mentioned meter reading recognition method can simultaneously identify and obtain the scale position information and scale value information of the instrument through the scale information recognition module of the reading recognition model, without the need to configure a template diagram, thereby saving manpower and material resources. The scale position is taken as the key point and the scale value is taken as the category. The scale position information and scale value information of the instrument are directly identified and obtained by using the scale information recognition module, which saves the step of obtaining the scale value information by using optical character recognition (OCR). Compared with the solution of obtaining the scale value information by using OCR, the method is not affected by the rotation of the scale value, and has higher recognition accuracy. The robustness of the instrument reading recognition method in complex environments is increased, and accurate reading recognition of the instrument can be achieved under the interference of complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0018] Figure 1 This is a flow chart of an embodiment of the meter reading recognition method of the present application;

[0019] Figure 2 This is a flow chart of obtaining the instrument reading corresponding to the image to be identified through scale information and pointer information in an embodiment of the instrument reading identification method of the present application;

[0020] Figure 3 This is a schematic diagram of the structure of an embodiment of the instrument reading recognition device of the present application;

[0021] Figure 4 This is a schematic diagram of the hardware structure of an embodiment of the electronic device of the present application.

[0022] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0024] It should be noted that all directional indications in the embodiments of the present application (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0025] In addition, the descriptions of "first", "second", etc. in this application are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in this field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0026] An embodiment of the present application provides a meter reading identification method, which is used, for example, to identify the readings of pointer-type meters in a substation. The pointer-type meters in the substation can be of various specific styles.

[0027] Figure 1 This is a flow chart of an embodiment of the meter reading recognition method of the present application. The meter reading recognition method includes steps S120, S130, S150, and S160.

[0028] In step S120, a trained reading recognition model is obtained, where the reading recognition model includes a scale information recognition module and a pointer information recognition module.

[0029] In some embodiments, obtaining the trained reading recognition model includes: obtaining a scale information recognition module by training a key point model. The key point model may be a YOLO-pose model (You Only Look Once-pose model, YOLO pose estimation model).

[0030] In some embodiments, the scale information recognition module obtained by training the key point model includes: obtaining a number of instrument image samples, the instrument image samples include scale key points and scale value category annotations, and obtaining a first training atlas; based on the first training atlas, the key point model is trained to obtain a scale information recognition module, the scale information recognition module is used to identify the instrument image, and obtain the corresponding scale position information and scale value information. In some embodiments, based on the first training atlas, the YOLO-pose model can be trained to obtain the scale information recognition module. The YOLO-pose model can be a YOLOv8-pose model.

[0031] In some embodiments, obtaining the trained reading recognition model includes: obtaining a pointer information recognition module by training a segmentation model. The segmentation model may be a YOLO-segment model (You Only Look Once-segment model, YOLO segmentation model).

[0032] In some embodiments, the pointer information recognition module obtained by training the segmentation model includes: obtaining a number of instrument image samples, the instrument image samples include annotations of the pointer area, and obtaining a third training atlas; based on the third training atlas, the segmentation model is trained to obtain the pointer information recognition module, and the scale information recognition module is used to recognize the instrument image and obtain the corresponding pointer information.

[0033] In some embodiments, the reading recognition model is configured to output scale position information, scale value, and information pointer information in multiple heads. At this time, obtaining the trained reading recognition model may include: obtaining a number of instrument image samples, the instrument image samples include scale key points, scale value categories, and pointer area annotations to obtain a training atlas; based on the training atlas, a YOLO-pose+segment model (You Only Look Once-pose+segment model, a YOLO pose estimation and segmentation model) is trained to obtain a reading recognition model, which is a multi-head output model. The YOLO-pose+segment model is, for example, a YOLO v8-pose+segment model.

[0034] In this embodiment, the structure of the YOLOv8 model network is improved, the output head of the YOLOv8 model is modified, the scale value category, scale key points and segmentation mask are combined into the first prediction matrix, and the global proto matrix of the output mask is the second prediction matrix. During training, the first prediction matrix and the second prediction matrix obtained by inference of the input instrument image sample are multiplied by the matrix to calculate the loss functions of target detection, key points and instance segmentation respectively, and the gradient of the neural network weight decrease is obtained by back propagation, and the network weight of the iterative YOLOv8 model is optimized during training.

[0035] In step S130, the image to be identified is acquired. In some embodiments, the image to be identified is acquired by an image acquisition device. In some embodiments, the image acquisition device is arranged near the instrument to be identified.

[0036] In step S150, the pointer information corresponding to the image to be identified is obtained through the pointer information identification module, and the scale information corresponding to the image to be identified is obtained through the scale information identification module. The scale information includes scale position information and scale value information. The scale information identification module is a key point recognition model.

[0037] In some embodiments, obtaining scale information corresponding to the image to be identified through the scale information identification module includes: identifying scale key points of the instrument image in the image to be identified through the scale information identification module to obtain scale position information; identifying the scale value category of the instrument image in the image to be identified through the scale information identification module to obtain scale value information.

[0038] In the substation, the scale value categories of all pointer instruments are finite categories. Therefore, different scale values ​​can be marked as different categories. In the recognition stage, the scale value category of the instrument image is recognized to obtain the scale value information.

[0039] In some embodiments, the pointer information recognition module is a segmentation model. Obtaining pointer information corresponding to the image to be recognized through the pointer information recognition module includes: performing regional segmentation on the instrument image in the image to be recognized through the pointer information recognition module to obtain pointer information corresponding to the pointer in the instrument image, that is, segmenting to obtain a segmented area corresponding to the pointer in the instrument image.

[0040] In this embodiment, the reading recognition model is configured to output scale position information, scale value and information pointer information in multiple heads. The reading recognition model is obtained by training the YOLO-pose+segment model.

[0041] In step S160, the meter reading corresponding to the image to be identified is obtained through the scale information and the pointer information.

[0042] Figure 2This is a flowchart of obtaining the meter reading corresponding to the image to be identified through scale information and pointer information in an embodiment of the meter reading recognition method of the present application. In some embodiments, step S150 of obtaining the meter reading corresponding to the meter image through scale information and pointer information includes steps S151 to S153.

[0043] In step S151, the center of the arc where the multiple scale key points are located is obtained according to the multiple scale key points in the scale information.

[0044] In step S152, the corresponding pointer pointing is obtained according to the pointer information, that is, the pointer fitting line segment corresponding to the pointer information is obtained.

[0045] In step S153, the meter reading corresponding to the meter image in the image to be identified is obtained according to the scale value information, the center of the circle, and the pointer pointing corresponding to each scale key point in the scale information.

[0046] In some embodiments, the step of obtaining the meter reading corresponding to the meter image in the image to be identified according to the scale value information, the center of the circle, and the pointer pointing corresponding to each scale key point in the scale information includes obtaining the meter reading according to the following formula 1:

[0047]

[0048] In the above formula, s1 represents the scale value of the smaller scale point adjacent to the pointer fitting line segment, s2 represents the scale value of the larger scale point adjacent to the pointer fitting line segment, θ1 represents the angle between the line segment from the center of the circle to the smaller scale point and the pointer fitting line segment, θ2 represents the angle between the line segment from the center of the circle to the larger scale point and the pointer fitting line segment, and val represents the instrument reading.

[0049] like Figure 1 In some embodiments, before step S150 of obtaining pointer information corresponding to the image to be identified through the pointer information identification module and obtaining scale information corresponding to the image to be identified through the scale information identification module, the meter reading identification method also includes step S110 and step S140.

[0050] In step S110, a trained instrument detection model is obtained.

[0051] In some embodiments, the step S110 of obtaining the trained instrument detection model includes: obtaining a number of instrument image samples, the instrument image samples including annotations of whether they have defects, and obtaining a second training atlas; based on the second training atlas, training a YOLO-detect model (You Only Look Once-detect model, YOLO target detection model) to obtain an instrument detection model. The instrument detection model is used to identify and determine whether there is an instrument image in the image to be identified and whether the instrument image has defects. The YOLO-detect model is, for example, a YOLO v8-detect model.

[0052] In step S140, the image to be identified is identified by the instrument detection model to obtain a first identification result to determine whether the image to be identified contains an instrument image and whether the instrument image has defects.

[0053] If the first recognition result is that the image to be recognized contains an instrument image and the instrument image has no defects, step S150 and step S160 are performed.

[0054] In some embodiments, if the first recognition result is that the image to be recognized contains an instrument image and the instrument image has no defects, the image to be recognized is cropped to obtain a cropped image to be recognized, the frame of the cropped image to be recognized is smaller than the frame of the image to be recognized before cropping, and the cropped image to be recognized contains the instrument image, so that the cropped image to be recognized becomes a close-up image of the instrument to be recognized. After obtaining the cropped image to be recognized, steps S150 and S160 are performed.

[0055] In some embodiments, the meter reading identification method further includes step S170 and step S180.

[0056] In step S170, if the first recognition result is that there is no instrument image in the image to be recognized, the first abnormal information is output. Therefore, further, a corresponding abnormal warning can be generated according to the first abnormal information, indicating that there is no instrument image in the image to be recognized.

[0057] In step S180, if the first recognition result is that the image to be recognized has an instrument image and the instrument image is defective, the second abnormal information is output. Therefore, further, a corresponding abnormal warning can be generated according to the second abnormal information, indicating that the instrument image in the image to be recognized has defects.

[0058] According to the meter reading recognition method of the embodiment of the present application, a trained reading recognition model is obtained, and the reading recognition model includes a scale information recognition module and a pointer information recognition module. After obtaining the image to be recognized, the pointer information corresponding to the image to be recognized is obtained through the pointer information recognition module, and the scale information corresponding to the image to be recognized is obtained through the scale information recognition module. The scale information includes scale position information and scale value information. The scale information recognition module is a key point recognition model, so that the meter reading corresponding to the image to be recognized can be obtained through the scale information and the pointer information. The above-mentioned meter reading recognition method can simultaneously identify and obtain the scale position information and scale value information of the instrument through the scale information recognition module of the reading recognition model, without the need to configure a template diagram, thereby saving manpower and material resources. The scale position is taken as the key point and the scale value is taken as the category. The scale position information and scale value information of the instrument are directly identified and obtained by using the scale information recognition module, which saves the step of obtaining the scale value information by using optical character recognition (OCR). Compared with the solution of obtaining the scale value information by using OCR, the method is not affected by the rotation of the scale value, and has higher recognition accuracy. The robustness of the instrument reading recognition method in complex environments is increased, and accurate reading recognition of the instrument can be achieved under the interference of complex environments.

[0059] In some embodiments, the reading recognition model is configured to output scale position information, scale value and information pointer information in multiple ways. Therefore, when the meter image is recognized by the reading recognition model, the multi-head mechanism of the reading recognition model is utilized, so that one reading recognition model can simultaneously obtain scale position information, scale value and information pointer information, which greatly reduces the calculation time of the meter reading recognition method.

[0060] The embodiment of the present application also provides a device for identifying meter readings.

[0061] Figure 3 The schematic diagram of the structure of an embodiment of the instrument reading recognition device of the present application is as follows. The instrument reading recognition device comprises a reading recognition model training module 120 , an image acquisition module 130 , a reading information recognition module 150 , and a reading calculation module 160 .

[0062] The reading recognition model training module 120 is used to obtain a trained reading recognition model, and the reading recognition model includes a scale information recognition module and a pointer information recognition module.

[0063] The image acquisition module 130 is used to acquire the image to be identified;

[0064] The reading information recognition module 150 is used to obtain the pointer information corresponding to the image to be recognized through the pointer information recognition module, and obtain the scale information corresponding to the image to be recognized through the scale information recognition module. The scale information includes scale position information and scale value information. The scale information recognition module is a key point recognition model.

[0065] The reading calculation module 160 is used to obtain the instrument reading corresponding to the image to be identified through the scale information and the pointer information.

[0066] In some embodiments, the meter reading recognition device further includes a meter detection model training module 110 and a meter detection module 140 .

[0067] The instrument detection model training module 110 is used to obtain a trained instrument detection model.

[0068] The instrument detection module 140 is used to identify the image to be identified by using the instrument detection model to obtain a first recognition result, so as to determine whether the image to be identified contains an instrument image and whether the instrument image has defects.

[0069] According to the instrument reading recognition device of the embodiment of the present application, the reading recognition model training module 120 obtains the trained reading recognition model, and the reading recognition model includes a scale information recognition module and a pointer information recognition module. After obtaining the image to be recognized, the reading information recognition module 150 obtains the pointer information corresponding to the image to be recognized through the pointer information recognition module, and obtains the scale information corresponding to the image to be recognized through the scale information recognition module. The scale information includes scale position information and scale value information. The scale information recognition module is a key point recognition model, and the reading calculation module 160 can obtain the instrument reading corresponding to the image to be recognized through the scale information and the pointer information. In the above-mentioned instrument reading recognition process, the scale position information and scale value information of the instrument can be simultaneously identified and obtained through the scale information recognition module of the reading recognition model, without the need to configure a template diagram, thereby saving manpower and material resources. Taking the scale position as the key point and the scale value as the category, the scale information recognition module is used to directly identify and obtain the scale position information and scale value information of the instrument, eliminating the step of using OCR to obtain the scale value information. Compared with the solution of using OCR to obtain the scale value information, it is not affected by the rotation of the scale value, and the recognition accuracy is higher. The robustness of the instrument reading recognition device in complex environments is increased, and accurate reading recognition of the instrument can be achieved under complex environmental interference.

[0070] An embodiment of the present application also provides an electronic device. Figure 4The hardware structure diagram of an embodiment of an electronic device of the present application is shown in FIG. The electronic device includes a memory 410 and at least one processor 420, the memory 410 is electrically connected to the at least one processor 420, the memory 410 stores instructions, and the at least one processor 420 calls the instructions in the memory 410, so that the electronic device executes the meter reading recognition method according to any of the aforementioned embodiments of the present application.

[0071] The instrument reading recognition method includes: obtaining a trained reading recognition model, the reading recognition model includes a scale information recognition module and a pointer information recognition module; obtaining an image to be recognized; obtaining pointer information corresponding to the image to be recognized through the pointer information recognition module, obtaining scale information corresponding to the image to be recognized through the scale information recognition module, the scale information includes scale position information and scale value information, the scale information recognition module is a key point recognition model; obtaining the instrument reading corresponding to the image to be recognized through the scale information and the pointer information.

[0072] Specifically, the processor 420 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0073] The memory 410 may include a large capacity memory 410 for data or instructions. By way of example and not limitation, the memory 410 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 410 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 410 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 410 is a non-volatile solid-state memory. In a specific embodiment, the memory 410 includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or a flash memory, or a combination of two or more of these.

[0074] In one example, the control device may further include a communication interface 430 and a bus 440. The processor 420, the memory 410, and the communication interface 430 are connected via the bus 440 and communicate with each other.

[0075] The communication interface 430 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0076] Bus 440 includes hardware, software or both, and the parts of online data flow billing equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front-end bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory 410 bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 440 may include one or more buses. Although the present application embodiment describes and shows a specific bus, the present application considers any suitable bus or interconnection.

[0077] In addition, in combination with the meter reading recognition method in the above embodiment, the embodiment of the present application can provide a computer-readable storage medium for implementation. The computer-readable storage medium stores instructions, and when the instructions are executed by the processor, any one of the meter reading recognition methods in the above embodiment is implemented.

[0078] The instrument reading recognition method includes: obtaining a trained reading recognition model, the reading recognition model includes a scale information recognition module and a pointer information recognition module; obtaining an image to be recognized; obtaining pointer information corresponding to the image to be recognized through the pointer information recognition module, obtaining scale information corresponding to the image to be recognized through the scale information recognition module, the scale information includes scale position information and scale value information, the scale information recognition module is a key point recognition model; obtaining the instrument reading corresponding to the image to be recognized through the scale information and the pointer information.

[0079] The present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known methods is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.

[0080] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0081] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.

[0082] The above is only a specific implementation of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.

Claims

1. A method for identifying meter readings, characterized in that: include: Acquire a trained reading recognition model, wherein the reading recognition model includes a scale information recognition module and a pointer information recognition module; Obtain an image to be recognized; The pointer information corresponding to the image to be identified is obtained through the pointer information identification module, and the scale information corresponding to the image to be identified is obtained through the scale information identification module, wherein the scale information includes scale position information and scale value information, and the scale information identification module is a key point recognition model; The meter reading corresponding to the image to be identified is obtained through the scale information and the pointer information.

2. The meter reading recognition method according to claim 1, characterized in that: The method of acquiring the trained reading recognition model includes: obtaining a scale information recognition module by training a key point model, wherein the method of obtaining a scale information recognition module by training a key point model includes: Acquire a number of instrument image samples, wherein the instrument image samples include annotations of scale key points and scale value categories, to obtain a first training atlas; Based on the first training atlas, the key point model is trained to obtain the scale information recognition module, and the scale information recognition module is used to recognize the instrument image and obtain the corresponding scale position information and scale value information.

3. The meter reading recognition method according to claim 2, characterized in that: The obtaining the scale information corresponding to the image to be identified by the scale information identification module includes: The scale information recognition module recognizes the scale key points of the instrument image in the image to be recognized to obtain the scale position information; The scale value category of the instrument image in the image to be identified is identified by the scale information identification module to obtain scale value information.

4. The meter reading recognition method according to claim 1, characterized in that: The pointer information recognition module is a segmentation model, and obtaining the pointer information corresponding to the image to be recognized through the pointer information recognition module includes: The pointer information recognition module performs region segmentation on the instrument image in the image to be recognized to obtain pointer information corresponding to the pointer in the instrument image.

5. The meter reading recognition method according to claim 1, characterized in that: The reading recognition model is configured to output the scale position information, the scale value and the information pointer information in multiple ways.

6. The meter reading recognition method according to claim 1, characterized in that: Before the step of obtaining the pointer information corresponding to the image to be identified by the pointer information identification module and obtaining the scale information corresponding to the image to be identified by the scale information identification module, the meter reading identification method further includes: Get the trained instrument detection model; The image to be identified is identified by the instrument detection model to obtain a first identification result, so as to determine whether the image to be identified contains an instrument image and whether the instrument image has defects.

7. The meter reading recognition method according to claim 6, characterized in that: The step of obtaining the trained instrument detection model includes: Acquire a number of instrument image samples, wherein the instrument image samples include annotations of whether they have defects, to obtain a second training atlas; Based on the second training atlas, the YOLO-detect model is trained to obtain an instrument detection model.

8. The meter reading recognition method according to claim 6, characterized in that: Also includes: If the first recognition result is that there is no instrument image in the image to be recognized, outputting first abnormal information; If the first recognition result is that the image to be recognized contains an instrument image and the instrument image has defects, second abnormal information is output.

9. The meter reading recognition method according to claim 1, characterized in that: The obtaining of the meter reading corresponding to the image to be identified by using the scale information and the pointer information includes: According to the multiple scale key points in the scale information, the center of the arc where the multiple scale key points are located is obtained; Obtaining a corresponding pointer pointing according to the pointer information; According to the scale value information corresponding to each of the scale key points in the scale information, the center of the circle, and the pointer pointing, the instrument reading corresponding to the instrument image in the image to be identified is obtained.

10. A device for identifying meter readings, characterized in that: include: A reading recognition model training module, used to obtain a trained reading recognition model, wherein the reading recognition model includes a scale information recognition module and a pointer information recognition module; An image acquisition module, used for acquiring an image to be identified; A reading information recognition module, used to obtain the pointer information corresponding to the image to be recognized through the pointer information recognition module, and obtain the scale information corresponding to the image to be recognized through the scale information recognition module, wherein the scale information includes scale position information and scale value information, and the scale information recognition module is a key point recognition model; The reading calculation module is used to obtain the instrument reading corresponding to the image to be identified through the scale information and the pointer information.

11. An electronic device, characterized in that: The electronic device comprises a memory and at least one processor, wherein the memory is electrically connected to the at least one processor, and instructions are stored in the memory. The at least one processor calls the instructions in the memory so that the electronic device executes the meter reading identification method according to any one of claims 1 to 9.

12. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the meter reading recognition method according to any one of claims 1 to 9 is implemented.

Images