Pointer instrument reading method, device and system and storage medium

Abstract

The embodiment of the invention provides a pointer instrument reading method, device and system and a storage medium. The method comprises the following steps: performing scale number identification on a pointer instrument image to obtain a scale number identification result; performing pointer segmentation on the pointer instrument image by taking the image position information of the detection frame of the pointer of the pointer instrument as prompt information; and the target image position information determined according to the pointer segmentation result and the scale number identification result is taken as the prompt information to perform scale segmentation on the pointer instrument image, so that the segmentation of the pointer and the scale can be completed under the condition of zero learning, and the pointer instrument corresponding to the pointer instrument image does not need to be learned; and semantic segmentation of the pointer instrument image can be realized. As the pointer instrument of a specific type does not need to be learned, pointer segmentation can be performed on the image of the pointer instrument of an unknown type, the universality of semantic segmentation of the image of the pointer instrument is improved, and the universality of reading of the pointer instrument can be further improved.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and particularly to a pointer instrument reading method, device, system and storage medium. Background Art

[0002] Pointer instruments are commonly used measuring instruments in the industrial field, generally consisting of scale lines, scale numbers and pointers. In some application scenarios, there are a large number of pointer instruments. Manually reading the pointer instruments is undoubtedly time-consuming and laborious.

[0003] In order to reduce the labor cost of reading pointer instruments, in some traditional solutions, the pointer instrument image is collected, and object detection and semantic segmentation models are used to locate the pointer and scale in the pointer instrument image. Then, the reading of the pointer instrument is calculated by using the known range information and the position ratio of the pointer end point relative to the scale after processing. However, this pointer instrument reading method cannot read any type of input pointer instrument image, and its generality and universality are relatively low. Summary of the Invention

[0004] Multiple aspects of this application provide a pointer instrument reading method, device, system and storage medium to improve the generality of the pointer instrument reading method.

[0005] An embodiment of this application provides a pointer instrument reading method, including:

[0006] Performing scale number recognition on a pointer instrument image of a dial image containing a pointer instrument to determine the scale number recognition result of the pointer instrument;

[0007] Performing pointer detection on the pointer instrument image to determine the image position information of the detection frame of the pointer of the pointer instrument;

[0008] Using the image position information of the detection frame of the pointer as a prompt message to perform image segmentation on the pointer instrument image to obtain a target pointer segmentation result;

[0009] Determining target image position information for scale segmentation prompt according to the target pointer segmentation result and the scale number recognition result;

[0010] Using the target image position information as a prompt message to perform image segmentation on the pointer instrument image to obtain a target scale segmentation result;

[0011] Determining the reading of the pointer instrument according to the scale number recognition result, the target pointer segmentation result and the target scale segmentation result.

[0012] An embodiment of the present application further provides a data center inspection system, including: an autonomous mobile device and a server device; the autonomous mobile device is provided with an image acquisition device;

[0013] The autonomous mobile device is configured to move in the data center and, during the movement, control the image acquisition device to perform image acquisition on the pointer instrument in the data center to obtain a pointer instrument image including a dial image of the pointer instrument; and provide the pointer instrument image to the server device;

[0014] The server device is configured to execute the steps in the above-mentioned pointer instrument reading method.

[0015] An embodiment of the present application further provides an electronic device, including: a memory and a processor; wherein, the memory is configured to store a computer program;

[0016] The processor is coupled to the memory and is configured to execute the computer program to execute the steps in the above-mentioned pointer instrument reading method.

[0017] An embodiment of the present application further provides a computer-readable storage medium storing computer instructions, which, when executed by one or more processors, cause the one or more processors to execute the steps in the above-mentioned pointer instrument reading method.

[0018] In the embodiment of the present application, scale digit recognition is performed on the pointer instrument image to obtain a scale digit recognition result; pointer segmentation is performed on the pointer instrument image using the image position information of the detection frame of the pointer of the pointer instrument as a prompt message; and scale segmentation is performed on the pointer instrument image using the target image position information determined according to the pointer segmentation result and the scale digit recognition result as a prompt message, so that the segmentation of the pointer and the scale can be completed without zero learning, without learning the pointer instrument corresponding to the pointer instrument image, and semantic segmentation of the pointer instrument image can also be achieved. Since there is no need to learn for a specific type of pointer instrument, the image of an unknown type of pointer instrument can be pointer-segmented, improving the generality and universality of the semantic segmentation of the pointer instrument image, and further improving the generality and universality of the pointer instrument reading. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The schematic embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0020] Figure 1a is a schematic structural diagram of the data center inspection system provided by the embodiment of the present application;

[0021] Figure 1bSchematic flowchart of the pointer instrument reading method provided by the embodiment of the present application;

[0022] Figure 1c Schematic diagram of the specific process of the pointer instrument reading method provided by the embodiment of the present application;

[0023] Figure 2 Schematic diagram of the pointer segmentation process provided by the embodiment of the present application;

[0024] Figure 3 Schematic diagram of the pointer segmentation result provided by the embodiment of the present application;

[0025] Figure 4 Schematic diagram of the dial layout of the pointer instrument provided by the embodiment of the present application;

[0026] Figure 5 Schematic diagram of the scale segmentation result provided by the embodiment of the present application;

[0027] Figure 6 Schematic diagram of the structure of the electronic device provided by the embodiment of the present application. Detailed implementation manners

[0028] To make the objectives, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be clearly and completely described below in conjunction with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present application.

[0029] In the traditional solution for determining a pointer instrument based on a pointer instrument image, a target detection and semantic segmentation model is used to locate the pointer and scale in the pointer instrument image. Then, the reading of the pointer instrument is calculated using the known range information and the position ratio of the pointer endpoint relative to the scale obtained through processing. In the above traditional pointer instrument reading solution, a semantic segmentation model for a specific pointer instrument usually needs to be trained, and a large number of semantic segmentation samples of pointer instrument images need to be marked for the specific pointer instrument; and the semantic segmentation samples of the marked pointer instrument images are used to train the semantic segmentation model to obtain a semantic segmentation model dedicated to the specific pointer instrument. That is, the trained semantic segmentation model can only perform semantic segmentation on the specific pointer instrument that has been learned, and cannot perform semantic segmentation on other types of pointer instruments that have not been learned, and thus cannot use the semantic segmentation result to read the unlearned types of pointer instruments. Therefore, the versatility and universality of the traditional pointer instrument reading solution are relatively low.

[0030] In some embodiments of the present application, in order to improve the generality and universality of the pointer instrument reading method, a solution is proposed. The basic idea is as follows: perform scale digit recognition on the pointer instrument image to obtain the scale digit recognition result; use the image position information of the detection frame of the pointer of the pointer instrument as a prompt message to segment the pointer of the pointer instrument image; and use the target image position information determined according to the pointer segmentation result and the scale digit recognition result as a prompt message to segment the scale of the pointer instrument image. The segmentation of the pointer and the scale can be completed without zero learning, without learning the pointer instrument corresponding to the pointer instrument image, and the semantic segmentation of the pointer instrument image can also be realized. Since there is no need to learn for a specific type of pointer instrument, the pointer of an unknown type of pointer instrument image can be segmented, which improves the generality and universality of the semantic segmentation of the pointer instrument image, and further improves the generality and universality of the pointer instrument reading.

[0031] The following will describe in detail the technical solutions provided by the embodiments of the present application with reference to the accompanying drawings.

[0032] It should be noted that the same reference numerals represent the same object in the following drawings and embodiments. Therefore, once an object is defined in one drawing or embodiment, it does not need to be further discussed in the subsequent drawings and embodiments.

[0033] Figure 1a It is a schematic structural diagram of the data center inspection system provided by the embodiment of the present application. As Figure 1b shown, the data center inspection system may include: an autonomous mobile device 10 and a server device 20.

[0034] There are various different pointer instruments in the data center, which are used to monitor various performance indicators of the data center. For example, a pressure gauge is set in the data center to monitor the pressure of the cooling system of the data center; an ammeter and a voltmeter are used to monitor the current and voltage of the cabinets in the data center, etc. The data center maintenance personnel can understand the health status of the data center according to the monitoring results, that is, the readings, of these pointer instruments.

[0035] In order to reduce the manual training cost, the autonomous mobile device 10 is often used to inspect the data center. In this embodiment, the autonomous mobile device 10 can move autonomously and complete some operation tasks on the basis of autonomous movement. In the embodiments of the present application, the specific implementation form of the autonomous mobile device 10 is not limited. The autonomous mobile device 10 can be implemented as a robot or a drone, etc. Among them, the appearance of the robot can be humanoid, animal-shaped, vehicle-shaped or puppet-shaped, etc. In this embodiment, as Figure 1aAs shown, an image acquisition device 11 is installed on the autonomous mobile device 10. In this embodiment, the image acquisition device can be implemented by any device with image acquisition function, such as a camera, a camera, a video recorder, etc. The images acquired by the image acquisition device can be independent frames of images, or video frames in a video.

[0036] In this embodiment, the autonomous mobile device 10 can move in the data center. In the embodiments of the present application, the autonomous mobile device 10 refers to a device with an independent power system. The autonomous mobile device 10 can move by using its own power system. The power system can include drive wheels, drive motors, transmission devices, etc. The autonomous mobile device 10 can automatically move along the inspection route by using its own power system. For example, the autonomous mobile device 10 can automatically plan the inspection route; and automatically move along the inspection route. Of course, the autonomous mobile device 10 can also be controlled by a user or other devices to move in the data center. For example, a computing device (such as the server device 20) etc. controls the autonomous mobile device 10 to move in the data center. Optionally, the computing device can send the inspection route to the autonomous mobile device 10, and control the autonomous mobile device 10 to move in the data center according to the inspection route. For another example, the user can control the autonomous mobile device 10 to move in the data center through terminals such as mobile phones and remote controls, etc. For the autonomous mobile device 10, it can respond to the terminal control signal and move by using its own power system.

[0037] Specifically, the autonomous mobile device 10 moves in the data center, and controls the image acquisition device 11 to acquire images of the pointer instruments in the data center during the movement, so as to obtain a pointer instrument image including the image of the dial of the pointer instrument.

[0038] In the embodiments of the present application, in order to reduce repeated image acquisition, a plurality of acquisition positions can also be preset on the inspection route. The pointer instrument images acquired by the image acquisition device 11 at the plurality of acquisition positions can cover all the pointer instruments in the data center. In the embodiments of the present application, the plurality of acquisition positions can be determined according to the acquisition perspective of the image acquisition device 11 and the distance between the inspection route and the pointer instrument. Of course, it is also possible to manually test the acquisition range of the pointer instrument acquired by the image acquisition device on the inspection route in advance; and determine the plurality of acquisition positions according to the manual test results, so that the pointer instrument images acquired by the image acquisition device 11 at the plurality of acquisition positions can cover all the pointer instruments in the data center. Among them, each acquisition position is used to acquire images of some pointer instruments in the data center.

[0039] Based on a plurality of preset acquisition positions, the autonomous mobile device 10 can perform autonomous positioning during the process of moving in the data center according to the set inspection route to determine the position information of the current position where the autonomous mobile device 10 moves to. In the embodiments of the present application, the specific implementation manner of the autonomous mobile device for autonomous positioning is not limited. In some embodiments, the autonomous mobile device 10 can adopt the Simultaneous Localization And Mapping (SLAM) technology for autonomous positioning.

[0040] Specifically, the autonomous mobile device 10 acquires the environmental information around its current position; and locates its pose in the stored environmental map according to the environmental information around its current position. Optionally, the autonomous mobile device 10 can construct a temporary map according to the environmental information obtained during the movement process; and compare the constructed temporary map with the stored environmental map to determine the pose of the robot in the stored environmental map. Among them, an optional implementation manner of comparing the constructed temporary map with the stored environmental map to determine the pose of the robot in the stored environmental map is: based on the matching algorithm, traverse each pose of the constructed temporary map on the stored environmental map. For example, if the grid size is 5 cm, a step size of 5 cm can be selected. For the temporary map, cover the possible poses in the stored environmental map, and then take a step size of 5 degrees for the angle, including the orientation parameters in all poses. When the grid representing the obstacle on the temporary map hits the grid representing the obstacle on the stored environmental map, points are added, and the pose with the highest score is determined as the pose of the global optimal solution; then, calculate the matching rate of the pose of the global optimal solution, and when the matching rate of the pose of the global optimal solution is greater than the preset matching rate threshold, determine the pose of the global optimal solution as the pose information of the autonomous mobile device 10. Among them, the pose information of the autonomous mobile device 10 includes: the position information and the orientation information of the autonomous mobile device 10.

[0041] After the autonomous mobile device 10 determines the position information of the current position it moves to, it can match the position information of the current position where the autonomous mobile device moves to among the set plurality of acquisition positions to determine whether the autonomous mobile device 10 moves to the set acquisition position. If the position information of the current position where the autonomous mobile device 10 moves to matches among the set plurality of acquisition positions, it is determined that the autonomous mobile device 10 moves to the set acquisition position P. Further, the autonomous mobile device 10 can acquire the pointer instrument image at the set acquisition position P.

[0042] In this embodiment, the autonomous mobile device 10 and the server device 20 can be communicatively connected. The connection between the autonomous mobile device 10 and the server device 20 can be wireless or wired. Optionally, the autonomous mobile device 10 and the server device 20 can be communicatively connected via the Internet. Of course, the autonomous mobile device 10 and the server device 20 can be communicatively connected to the terminal device 10b through a mobile network. Optionally, the autonomous mobile device 10 and the server device 20 can also be communicatively connected by means of Bluetooth, Wireless Fidelity (WiFi), or infrared rays, etc.

[0043] Based on the communication link between the autonomous mobile device 10 and the server device 20, the autonomous mobile device 10 can provide the captured pointer instrument image to the server device 20, and the server device 20 reads the pointer instrument according to the pointer instrument image. Among them, the server device 20 can be a single server device, or a cloudified server array, or a virtual machine (Virtual Machine, VM) running in the cloudified server array. In addition, the server device 20 can also refer to other computing devices with corresponding service capabilities, such as terminal devices (running service programs) such as computers, etc.

[0044] The method for the server device 20 to read the pointer instrument according to the pointer instrument image will be exemplarily described below. In the embodiments of the present application, the pointer instrument image includes the dial image of the pointer instrument.

[0045] Figure 1b It is a schematic flowchart of the pointer instrument reading method provided by the embodiments of the present application. As Figure 1b shown, the pointer instrument reading method mainly includes:

[0046] 101. Perform scale digit recognition on the pointer instrument image including the dial image of the pointer instrument to determine the scale digit recognition result of the pointer instrument.

[0047] 102. Perform pointer detection on the pointer instrument image to determine the image position information of the detection frame of the pointer of the pointer instrument.

[0048] 103. Use the image position information of the detection frame of the pointer as a prompt message to perform image segmentation on the pointer instrument image to obtain the target pointer segmentation result.

[0049] 104. Determine the target image position information for scale segmentation prompt according to the target pointer segmentation result and the scale digit recognition result.

[0050] 105. Use the target image position information as a prompt message to perform image segmentation on the pointer instrument image to obtain the target scale segmentation result.

[0051] 106. Determine the reading of the pointer instrument according to the scale number recognition result, the target pointer segmentation result, and the target scale segmentation result.

[0052] For a pointer instrument, the components of the pointer instrument generally include: scale numbers, scales, and pointers. When reading manually, the reading of the pointer instrument can be determined according to the target scale and scale numbers pointed by the pointer. To determine the reading of the pointer instrument, it is necessary to identify the scale numbers from the pointer instrument image. Based on this, combined with Figure 1a and Figure 1b in step 101, for the pointer instrument image, the scale numbers of the pointer instrument image can be recognized to determine the scale number recognition result of the pointer instrument. Among them, the scale number recognition result can include: the scale numbers of the pointer instrument, such as Figure 1a the numbers 0-10 on the pointer instrument image in . Of course, the scale number recognition result can also include: the image position information of the scale numbers, that is, the position information of the scale numbers on the pointer instrument image, which can be represented by the pixel coordinates corresponding to the scale numbers on the pointer instrument image.

[0053] In the embodiments of the present application, the specific implementation manner of recognizing the scale numbers of the pointer instrument image is not limited. In some embodiments, a digital recognition model can be used to recognize the scale numbers of the pointer instrument image to determine the scale number recognition result of the pointer instrument. In the embodiments of the present application, the specific implementation form of the digital recognition model is not limited. Optionally, the digital recognition model can be a neural network model, such as a convolutional neural network (CNN) model, a recurrent neural network (RNN), a deep neural network (DNN), or a feedforward neural network (FNN) model, etc.

[0054] Since the neural network model needs to be pre-trained for scale number recognition, a large number of sample images marked with numbers need to be provided. Not only is the workload of sample marking for model training large, but the time cost of training the neural network model is also high.

[0055] In order to reduce the cost of model training, in some other embodiments, the Optical Character Recognition (OCR) technology can be used to perform character recognition on the pointer instrument image to determine the alphanumeric characters and the image position information of the characters included in the pointer instrument image. The OCR technology can detect the characters on the pointer instrument image, determine the shape of the characters by detecting the dark and bright patterns, and then translate the shape of the characters into computer text by using character recognition methods. That is, for the characters on the pointer instrument image, the characters in the pointer instrument image are converted into a black-and-white dot-matrix image file in an optical manner, and the text in the image file is converted into a text format by using recognition software.

[0056] Specifically, as Figure 1c shown in step 1, the OCR technology can be used to perform character detection on the pointer instrument image to determine the image position information of the characters included in the pointer instrument image; further, the OCR technology can be used to perform character recognition on the characters included in the pointer instrument image to determine the characters included in the pointer instrument image, that is, to determine the specific character content included in the pointer instrument image.

[0057] Since there may be other characters on the dial of the pointer instrument in addition to the scale numbers, such as the model number, manufacturer, and / or the unit of measurement of the reading of the pointer instrument, etc., but not limited thereto. Therefore, it is also necessary to screen out the scale numbers of the pointer instrument from the characters included in the recognized pointer instrument image. The inventors of the present application have found that the position distribution characteristics of the scale numbers of the pointer instrument on the dial, that is, the position distribution characteristics, have certain rules, generally distributed in an arc shape on the dial; and there is a certain logical relationship between the scale numbers of the pointer instrument, such as generally the scale numbers form an arithmetic sequence. Based on this, as Figure 1c shown in step 1, the scale numbers can be screened out from the characters included in the pointer instrument image according to the characters and the image position information of the characters included in the pointer instrument image.

[0058] Specifically, the position distribution characteristics of the characters can be determined according to the image position information of the characters; and the logical relationship between the characters can be determined according to the characters included in the pointer instrument image; then, the scale numbers can be screened out from the characters included in the pointer instrument image according to the position distribution characteristics of the characters and the logical relationship between the characters. Specifically, the characters with the position distribution characteristics of being distributed on the same arc can be screened out from the characters included in the pointer instrument image; and from the characters distributed on the same arc, the characters with the logical relationship of an arithmetic sequence can be selected as the scale numbers of the pointer instrument.

[0059] After determining the scale numbers of the pointer instrument, the image position information of the scale numbers of the pointer instrument can also be obtained from the image position information of the characters included in the pointer instrument image. Among them, the aforementioned scale number recognition result may include: the scale numbers of the pointer instrument and the image position information of the scale numbers. The image position information of the scale numbers specifically refers to the position information of the scale numbers on the pointer instrument image, generally represented by the pixel coordinates of the scale numbers in the pointer instrument image.

[0060] The implementation manner of recognizing the scale numbers of the pointer instrument image shown in the above embodiments is only an exemplary illustration and does not constitute a limitation.

[0061] If it is necessary to determine the reading of the pointer instrument, in addition to recognizing the scale numbers, it is also necessary to obtain the information of the pointer and the scale of the pointer instrument. Therefore, in the embodiments of the present application, in combination with Figure 1a , Figure 1b step 102 in Figure 1b and step 2 in

[0062]

[0063] In practical applications, when performing pointer detection on an image, a rectangular detection frame is usually used to label the pointer image included in the image. The image position information of the rectangular detection frame is specifically the image position information of the rectangular detection frame on the pointer instrument image, which can reflect the spatial distribution of the pointer in the pointer instrument image. Among them, the image position information of the detection frame includes: the center position, size, etc. of the detection frame. Optionally, the center position of the detection frame can be represented by the center coordinates of the rectangular detection frame, and the size of the detection frame can be represented by the width and height of the rectangular detection frame. Correspondingly, the image position information of the detection frame can be expressed as (x, y, w, h). Among them, (x, y) represents the center coordinates of the rectangular detection frame, that is, the pixel coordinates of the center of the rectangular detection frame in the pointer instrument image, and w and h respectively represent the width and height of the rectangular detection frame. Alternatively, the image position information of the rectangular detection frame includes: the vertex coordinates of the rectangular detection frame, etc. The center coordinates and vertex coordinates of the rectangular detection frame are both pixel coordinates on the pointer instrument image.

[0064] Among them, the number of pointers included in one frame of pointer instrument image can be 1 or more. More than one means 2 or more. Each pointer can correspond to a detection frame.

[0065] In the embodiments of the present application, before using the pointer detection model for pointer detection, the pointer detection model needs to be trained. Among them, the model architecture of the pointer detection model can be a Single Shot Detector (SSD) model, a YOLO (You only Look Once) series model, a CenterNet model, a Spatial Pyramid Pooling (SPPNet) model, a Feature Pyramid Networks (FPN) model, etc., but not limited thereto. The battery anomaly recognition model can be a neural network model, etc. The neural network model can be a CNN, RNN, DNN, or FNN model, etc.

[0066] In the embodiments of the present application, the pointer detection model can be trained using pointer instrument sample images. There are multiple pointer instrument sample images. Multiple means 2 or more. The pointer instruments corresponding to the pointer instrument sample images can be various types of pointer instruments. Various means 2 or more. Among them, the pointer instrument sample image can be a pointer instrument image with the image position information of the pointer pre-annotated. Optionally, the pointer annotation can be performed on the pointer instrument sample image by manual annotation to obtain the image position information of the reference detection box. In the embodiments of the present application, the reference detection box refers to the detection box that pre-annotates the pointer in the pointer instrument sample image.

[0067] Furthermore, the pointer detection model can be trained using the pointer instrument sample images to obtain the pointer detection model. The initially trained model of the pointer detection model is called the initial detection model. Among them, the model architecture of the initial detection model is the same as that of the pointer detection model finally obtained through model training, that is, the parameters of the model are the same. The model training in this embodiment mainly refers to: training the parameters of the pointer detection model using the pointer instrument sample images to minimize the loss function. That is, with the minimization of the loss function as the training objective, the pointer instrument sample images are used for model training to obtain the final pointer detection model. Among them, the loss function can be determined according to the image position information of the detection box obtained through model training and the image position information of the reference detection box obtained by pre-annotating the pointer in the pointer instrument sample image before model training.

[0068] Optionally, the loss function L x can be expressed as:

[0069] L x = L center + λ scale L scale + λ offset L offset (1)

[0070] In the loss function (1), L center represents the center loss, that is, the loss between the center coordinates of the detection box obtained by model training and the center coordinates of the reference detection box; L scale represents the scale loss, that is, the loss between the width and height of the detection box obtained by model training and the width and height of the reference detection box; L offset represents the offset loss, that is, the offset loss of the center coordinates of the detection box obtained by model training compared to the center coordinates of the reference detection box. λ scale 、λ offset and λ θ respectively represent the weights of the scale loss and the offset loss, which can be flexibly set according to the actual situation.

[0071] Among them, the training process of the above pointer detection model can be executed on the server device or on any other computing device. After the pointer detection model is trained, the pointer detection model can be pre-set in the server device 20. In this way, the pointer detection model can be used to detect the pointer in the pointer instrument image to obtain the image position information of the detection box of the pointer included in the pointer instrument image. Figure 2 The detection box shown in the (b) diagram of Figure 2 is the detection box of the pointer included in the pointer instrument image shown in the (a) diagram of

[0072] Furthermore, in combination with Figure 1a 、 Figure 1b step 103 in Figure 1c and step 2 in

[0073] Specifically, as Figure 1cAs shown in the corresponding image segmentation process, the image position information of the detection frame of the pointer can be used as the prompt information (Prompt), and the pointer instrument image can be used as the image to be segmented and input into the image segmentation model; in the image segmentation model, the pointer instrument image and the image position information of the detection frame of the pointer are respectively encoded to obtain the image feature vector of the pointer instrument image and the prompt feature vector of the image position information of the detection frame of the pointer; further, the image feature vector of the pointer instrument image and the prompt feature vector of the image position information of the detection frame of the pointer are subjected to mask mapping to obtain as shown in Figure 2 the mask map of the pointer shown in Figure (c) in

[0074] In this embodiment, taking the image position information of the detection frame of the pointer as the prompt information (Prompt), and using the image segmentation model to perform pointer segmentation on the pointer instrument image, the pointer segmentation can be completed without zero learning. That is, the image segmentation model can be a general image segmentation model, without the need to learn the image of the pointer instrument, and there is no need to perform pointer annotation on the images of various types of pointer instruments, which can reduce the pointer annotation cost of the image segmentation model. On the other hand, since the image segmentation model does not need to learn the pointer instrument to be read, it can perform pointer segmentation on the images of pointer instruments with unknown types, improving the versatility and universality of pointer segmentation.

[0075] In the embodiments of the present application, the specific implementation form of the image segmentation model is not limited. The image segmentation model can be a neural network model. In some embodiments, when the number of parameters of the image segmentation model is large, for example, when the number of parameters of the image segmentation model is millions, hundreds of millions, billions or even more, the image segmentation model can also be called a large model. In the embodiments of the present application, the large model is defined as a neural network model whose number of model parameters meets the preset parameter range. Among them, the number of model parameters corresponding to the preset parameter range is very large, which can be millions, hundreds of millions, billions or even more, and the specific value can be determined by the standards in the field of artificial intelligence.

[0076] Among them, the large model can be trained in two stages such as pre-training and fine-tuning. In the pre-training stage, the model is trained on a large-scale general-domain image to learn the basic structure of the language and various common senses. Then, in the fine-tuning stage, the model is further trained on a smaller and more specific-domain dataset. Fine-tuning can enable the model to better understand and generate the language of this specific domain, so as to better complete specific tasks. In the embodiments of the present application, the image segmentation model can adopt a large model completed by pre-training, without the need to perform model learning and training for specific types of pointer instruments, that is, the image segmentation model can adopt a general-domain image segmentation model, and there is no need to perform pointer annotation on the images of various types of pointer instruments, which can reduce the pointer annotation cost of the image segmentation model.

[0077] In some embodiments, the image segmentation model can be a Segment-anything Model (SAM). The SAM model can complete the segmentation of the pointer in a zero-shot manner.

[0078] Among them, as Figure 1c shown, the image segmentation model (such as the SAM model) can include: an image encoder, a prompt encoder, and a mask decoder. Among them, the image encoder performs image encoding on the pointer instrument image to obtain an image feature vector; the prompt encoder is used to perform prompt encoding on the image position information of the detection box of the pointer to obtain a prompt feature vector; the mask decoder is used to map the image feature vector of the pointer instrument image and the text prompt feature vector of the image position information of the detection box of the pointer to a mask, perform mask map prediction of the pointer, and obtain the mask map of the pointer.

[0079] Furthermore, the pointer instrument image can be segmented according to the mask map of the pointer to obtain the pointer segmentation result as shown in Figure 2 Figure (d). The pointer segmentation result can include: the image position information of the pointer (such as the pixel coordinates of the pointer in the pointer instrument image) and / or the image of the pointer. The image of the pointer is as shown in Figure 2 Figure (d).

[0080] Among them, the pointer segmentation result can be one or more. Multiple means two or more. For some image segmentation models (such as the SAM model), the pointer segmentation result includes multiple pointer segmentation results with different granularities. The pointer segmentation results with different granularities can correspond to different pointer regions. For example, as Figure 3 shown, using the pointer segmentation method shown in the foregoing embodiments, the pointer instrument image shown in Figure 3 Figure (a) is segmented to obtain Figure 3 the three pointer segmentation results with different granularities shown in (1), (2), and (3) in the figure. The regions or granularities of the pointers included in different pointer segmentation results are different. Based on this, the target pointer segmentation result can also be determined from the pointer segmentation results.

[0081] In some embodiments, if the pointer segmentation result is one, it can be determined that this pointer segmentation result is the target pointer segmentation result. In other embodiments, if the pointer segmentation result is multiple pointer segmentation results with different granularities, the target pointer segmentation result can be determined from the multiple pointer segmentation results (that is, Figure 1c "screen the target pointer segmentation result" in the figure).

[0082] Specifically, the pointer segmentation result may include: the image position information of the pointer, that is, the pixel coordinates of the pointer. Accordingly, the lengths of the pointers segmented by multiple pointer segmentation results can be determined based on the image position information of the pointers respectively included in the multiple pointer segmentation results; further, from the multiple pointer segmentation results, the pointer segmentation result whose segmented pointer length meets the set length condition can be selected as the target pointer segmentation result.

[0083] Among them, the set length condition can be the maximum length. Accordingly, from the multiple pointer segmentation results, the pointer segmentation result with the maximum segmented pointer length can be selected as the target pointer segmentation result. Or, the set length condition is that the length of the pointer is within a set length range. The set length range is pre-set according to the pointer length of the actual pointer instrument. Accordingly, from the multiple pointer segmentation results, the pointer segmentation result whose segmented pointer length is within the set length range can be selected as the target pointer segmentation result.

[0084] If it is necessary to determine the reading of the pointer instrument, in addition to identifying the scale numbers and the pointer in the pointer instrument image, it is also necessary to obtain the information of the scale of the pointer instrument. Therefore, it is also necessary to segment the scale from the pointer instrument image. If a scale segmentation model is used to segment the scale of the pointer instrument image, it is necessary to pre-label the scale of a specific type of pointer instrument and use the pointer instrument image of this specific type to perform large-scale model training on the scale segmentation model. This scale segmentation method has two problems. One is that the scale annotation cost for training the scale segmentation model is relatively high. The other is that it can only segment the scale of a specific type of pointer instrument (that is, the pointer instrument learned by the scale segmentation model), and cannot segment the scale of the pointer instrument that has not been learned.

[0085] To solve the above technical problems, the inventors of this application have found through research that there are certain regularities in the relative position relationships among the scale, scale numbers, and pointer of the pointer instrument. Based on this, combined with Figure 1a 、 Figure 1b step 104 in Figure 1c and step 3 in Figure 1c the target image position information for scale segmentation hint can be determined according to the scale number recognition result and the target pointer segmentation result obtained in the foregoing embodiments, that is,

[0086] the hint information for scale segmentation determined in

[0087] The inventors of the present application have found through research that there are certain regularities in the relative positional relationship among the scale, scale numbers, and the pointer of a pointer instrument, that is, as Figure 2 and Figure 3 shown, the scale of the pointer instrument is located between the arc formed by the rotation of the pointer (defined as the first arc) and the arc formed by the scale numbers (defined as the second arc). Based on this, as Figure 4 shown, the image position information of the first arc formed by the rotation of the pointer can be determined according to the image position information of the pointer; and the image position information of the second arc formed by the scale numbers can be determined according to the image position information of the scale numbers; then, the target image area located between the first arc and the second arc can be determined from the pointer instrument image according to the image position information of the first arc and the image position information of the second arc. Further, a plurality of pixel coordinates can be selected from the pixel coordinates of the target image area as the target image position information.

[0088] For example, in some embodiments, a set number of pixel coordinates can be randomly selected from the pixel coordinates of the target image area as the target image position information; the set number is greater than or equal to 2 and is an integer. Or, the pixel coordinates located on the third arc with a target radius r can be determined from the pixel coordinates of the target image area; and a plurality of pixel coordinates can be randomly selected from the pixel coordinates located on the third arc as the target image position information. In some embodiments, as Figure 4 shown, the scale numbers are located outside the scale, and the pointer end is located inside the scale, then the target radius r is greater than the radius r1 of the first arc and less than the radius r2 of the second arc. In other embodiments, as Figure 2 and Figure 3 shown, the scale numbers are located inside the scale, and the pointer end is located outside the scale or coincides with the arc of the scale, then the target radius r is the radius r2 of the second arc and less than the radius r1 of the first arc. The target radius r can be any value between the radius r1 of the first arc and the radius r2 of the second arc.

[0089] Or, a plurality of pixel coordinates can also be selected from the pixel coordinates located on the third arc at a set central angle interval as the target image position information.

[0090] The embodiments shown above for determining the target image position information for scale segmentation prompts are only exemplary descriptions and do not constitute limitations.

[0091] Further, in combination with Figure 1a 、 Figure 1b step 105 in Figure 1cIn step 3, the pointer instrument image can be segmented using the target image position information as a prompt to obtain the target scale segmentation result. The target image position information provides a prompt for scale segmentation. Based on this prompt, scale segmentation of the pointer instrument image can be completed in a zero-shot manner, that is, without learning the pointer instrument corresponding to the pointer instrument image, scale segmentation of the image of the pointer instrument can also be achieved. Since there is no need to learn the pointer instrument to be read, the image of a pointer instrument with an unknown type can be segmented, improving the generality and universality of scale segmentation.

[0092] Specifically, as Figure 1c shown in the image segmentation process, the target image position information can be used as a prompt, and the pointer instrument image can be input into the image segmentation model as the image to be segmented. In the image segmentation model, the pointer instrument image and the target image position information are encoded respectively to obtain the image feature vector of the pointer instrument image and the prompt feature vector of the target image position information. Further, a mask mapping is performed on the image feature vector of the pointer instrument image and the prompt feature vector of the target image position information to obtain the mask map of the scale.

[0093] In this embodiment, using the target image position information as a prompt and using the image segmentation model to perform scale segmentation on the pointer instrument image, scale segmentation can be completed in a zero-shot manner. The image segmentation model can be a general image segmentation model, without the need to learn the image of the pointer instrument, and thus there is no need to perform scale annotation on the images of various types of pointer instruments, which can reduce the scale annotation cost of the image segmentation model. On the other hand, since the image segmentation model does not need to learn the pointer instrument to be read, the image of a pointer instrument with an unknown type can be segmented, improving the generality and universality of scale segmentation. For the specific implementation form of the image segmentation model, reference can be made to the relevant content of the foregoing embodiments, which will not be elaborated here.

[0094] After determining the mask map of the scale, the pointer instrument image can be segmented according to the mask map of the scale to obtain the scale segmentation result. The scale segmentation result may include: the image position information of the scale (such as the pixel coordinates of the scale in the pointer instrument image) and / or the image of the scale. The image of the pointer is as Figure 5 shown in Figure (b).

[0095] Among them, the scale segmentation result can be one or more. Multiple means two or more. For the SAM model, the scale segmentation result includes multiple scale segmentation results with different granularities. The scale segmentation results with different granularities can correspond to different scale regions. The regions or granularities of the scales included in different scale segmentation results are different. Based on this, the target scale segmentation result can also be determined from the scale segmentation results.

[0096] In some embodiments, the scale segmentation result is one, and it can be determined that this scale segmentation result is the target scale segmentation result. In other embodiments, the scale segmentation result is multiple scale segmentation results with different granularities, then the target scale segmentation result can be determined from the multiple scale segmentation results (corresponding Figure 1c to screening the target scale segmentation result).

[0097] Specifically, the scale segmentation result may include: the image position information of the scale, that is, the pixel coordinates of the scale and / or the image of the scale. The inventors of the present application have found that the scales of pointer instruments are generally distributed in an arc. Based on this, the arc fitting can be performed on the scale distribution shapes corresponding to the multiple scale segmentation results according to the image position information of the scales included in the multiple scale segmentation results respectively, so as to obtain the arc fitting degrees of the scale distribution shapes corresponding to the multiple scale segmentation results respectively.

[0098] Among them, the arc fitting degree can be represented by the square of the correlation coefficient R (R-Square) obtained by performing arc fitting on the scale distribution shape corresponding to the scale segmentation result X. Among them, R-Square is a value that measures the degree of correlation between variables, also known as the goodness of fit or the coefficient of determination, and is used to represent the percentage of the change in the dependent variable that can be explained by the fitted arc. The closer R-Squard is to 1, the better the fitting effect.

[0099] After obtaining the circular arc fitting degrees of the scale distribution shapes corresponding to multiple scale segmentation results, from the multiple scale segmentation results, a scale segmentation result whose circular arc fitting degree of the scale distribution shape meets the set circular arc fitting degree condition can be selected as the target scale segmentation result. Among them, the set circular arc fitting degree condition can be that the larger the circular arc fitting degree is. Correspondingly, from the multiple scale segmentation results, the scale segmentation result with the largest circular arc fitting degree of the scale distribution shape can be selected as the target scale segmentation result. Or, the set circular arc fitting degree condition can be that the circular arc fitting degree is greater than or equal to a set circular arc fitting degree threshold. Correspondingly, from the multiple scale segmentation results, a scale segmentation result whose circular arc fitting degree of the scale distribution shape is greater than or equal to the set circular arc fitting degree threshold can be selected as the target scale segmentation result. If the number of scale segmentation results whose circular arc fitting degree of the scale distribution shape is greater than or equal to the set circular arc fitting degree threshold is multiple, then from the scale segmentation results whose circular arc fitting degree of the scale distribution shape is greater than or equal to the set circular arc fitting degree threshold, the scale segmentation result with the largest circular arc fitting degree of the scale distribution shape can be selected as the target scale segmentation result.

[0100] In some other embodiments, when determining the target scale segmentation result from multiple scale segmentation results, the blank areas corresponding to the multiple scale segmentation results can be identified from the images of the scales respectively included in the multiple scale segmentation results, and the areas of the blank areas corresponding to the multiple scale segmentation results can be calculated; according to the areas of the blank areas corresponding to the multiple scale segmentation results, a scale segmentation result whose blank area meets the set area condition can be selected from the multiple scale segmentation results as the target scale segmentation result.

[0101] Among them, the set area condition can be that the length is the largest. Correspondingly, from the multiple scale segmentation results, the scale segmentation result with the largest blank area can be selected as the target scale segmentation result. Or, the set area condition is that the area of the blank area is within a set area range. The set area range is flexibly set according to actual needs. Correspondingly, from the multiple scale segmentation results, a scale segmentation result whose blank area is within the set area range can be selected as the target scale segmentation result. If the number of scale segmentation results whose blank area is within the set area range is multiple, then from the scale segmentation results whose blank area is within the set area range, the scale segmentation result with the largest blank area can be selected as the target scale segmentation result.

[0102] In some other embodiments, the arc fitting degree of the scale distribution shape corresponding to the above scale segmentation result may be further combined with selecting a target scale segmentation result from multiple scale segmentation results and selecting a target scale segmentation result from multiple scale segmentation results according to the area of the blank area corresponding to the scale segmentation result. Specifically, a scale segmentation result that satisfies the set arc fitting degree condition for the scale distribution shape and the set area condition for the area of the blank area may be selected from multiple scale segmentation results as the target scale segmentation result.

[0103] The embodiments of determining the target scale segmentation result of the pointer instrument image shown in the above embodiments are only exemplary descriptions and do not constitute limitations.

[0104] Since the pointer reading is determined by the scale numbers of the pointer instrument, the scale of the pointer instrument, and the position of the pointer of the pointer instrument, after determining the scale number recognition result, the target scale segmentation result, and the target pointer segmentation result, the reading of the pointer instrument may be determined according to the scale number recognition result, the target pointer segmentation result, and the target scale segmentation result.

[0105] Specifically, since the scale number recognition result includes the scale numbers of the pointer instrument, the range of the pointer instrument may be determined according to the scale number recognition result (specifically, the scale numbers). Further, in combination with Figure 1a and Figure 1b Step 106, the graduation value of the pointer instrument may also be determined according to the target scale segmentation result and the range of the pointer instrument. The graduation value is the minimum scale value of the pointer instrument, which is the minimum value that can be read on the pointer instrument and is the value of the smallest grid between two adjacent scales on the pointer instrument. This process Figure 1c is not shown.

[0106] Specifically, the number of scales included in the target scale segmentation result may be determined according to the image of the scales in the target scale segmentation result; then, the graduation value of the pointer instrument may be determined according to the range of the pointer instrument and the number of scales included in the target scale segmentation result. The quotient of the range of the pointer instrument divided by the number of scales included in the target scale segmentation result is the graduation value of the pointer instrument.

[0107] After determining the graduation value of the pointer instrument, the target scale pointed by the pointer may be determined according to the target scale segmentation result and the target pointer recognition result. Specifically, according to the image position information of the scales in the target scale segmentation result and the image position information of the pointer in the target pointer recognition result, the scale whose image position information overlaps with the image position information of the pointer may be determined from the scales of the pointer instrument as the target scale pointed by the pointer, and thus it is determined which scale the pointer specifically points to.

[0108] After that, according to the graduation value of the pointer instrument and the target scale, the reading corresponding to the target scale can be determined, which is the reading of the pointer instrument. Specifically, the graduation value of the pointer instrument can be multiplied by the sequence number of the target scale (i.e., which scale the target scale is), and the obtained reading corresponding to the target scale is the reading of the pointer instrument.

[0109] In the embodiment of the present application, scale number recognition is performed on the pointer instrument image to obtain a scale number recognition result; the pointer of the pointer instrument is segmented from the pointer instrument image using the image position information of the detection frame of the pointer as a prompt message; and scale segmentation is performed on the pointer instrument image using the target image position information determined according to the pointer segmentation result and the scale number recognition result. The segmentation of the pointer and the scale can be completed without zero learning, without learning the pointer instrument corresponding to the pointer instrument image, and semantic segmentation of the pointer instrument image can also be realized. Since there is no need to learn the pointer instrument to be read, the pointer of the image of the pointer instrument with unknown type can be segmented, improving the generality and universality of the semantic segmentation of the pointer instrument image, and further improving the generality and universality of the reading of the pointer instrument.

[0110] On the other hand, in the embodiment of the present application, only the sample images used by the pointer detection model for pointer detection need to be pointer-annotated, and the image segmentation model used for pointer and scale segmentation does not need to be annotated, that is, there is no need to annotate semantic segmentation samples, which can reduce the workload and cost of sample annotation.

[0111] In addition, since scale number recognition is performed on the pointer instrument image in the embodiment of the present application to obtain a scale number recognition result; the scale number recognition result can reflect the range of the pointer instrument. Therefore, the range of the pointer instrument can be determined according to the scale number recognition result. When reading the pointer instrument subsequently, the reading of the pointer instrument can be performed according to the range of the pointer instrument, the scale segmentation result, and the pointer segmentation result determined by the scale number recognition result. Therefore, the pointer instrument reading method provided by the present application can also read the pointer instrument with an unknown range, further improving the generality and universality of the reading of the pointer instrument.

[0112] Since the aforementioned target image position information for scale segmentation prompt is determined according to the scale number recognition result and the pointer segmentation result, it has a certain degree of randomness and uncertainty, which in turn leads to a certain degree of uncertainty in the accuracy of the scale result segmented by the target image position information. Therefore, in order to improve the accuracy of the pointer instrument reading, the aforementioned target image position information for scale segmentation prompt can also be optimized. Specifically, as Figure 1c shown in step 4, the reliability score of the target scale segmentation result can be determined according to the aforementioned target scale segmentation result.

[0113] In the embodiments of the present application, the specific implementation manner of determining the reliability score of the target scale segmentation result based on the target scale segmentation result is not limited. The following is an exemplary illustration in combination with several implementation manners.

[0114] Embodiment 1: The uniformity score of the scale can be determined according to the target scale segmentation result; and the reliability score of the target scale segmentation result can be determined according to the uniformity score of the scale. The reliability score of the target scale segmentation result is positively correlated with the uniformity score of the scale, that is, the higher the uniformity score of the scale (the more uniform the scale distribution), the higher the reliability score of the target scale segmentation result. In some embodiments, the uniformity score of the scale can be used as the reliability score of the target scale segmentation result.

[0115] Specifically, the pixel interval between adjacent scales can be determined according to the image position information of the scales included in the target scale segmentation result; and the uniformity score of the scale can be determined according to the difference in the pixel intervals between every two adjacent scales. Among them, the difference in the pixel intervals between every two adjacent scales can be represented by the mean square error of the pixel intervals between every two adjacent scales. For example, the difference in the pixel intervals between every two adjacent scales can be expressed as:

[0116]

[0117] In formula (2), N represents the total number of scales. i represents the i-th scale, i = 0, 1,..., (N - 1). (X i , Y i ) represents the image position information of the i-th scale; (X i+1 , Y i+1 ) represents the image position information of the (i + 1)-th scale. represents the pixel interval between the (i + 1)-th scale and the i-th scale; represents the pixel interval between the (i + 2)-th scale and the (i + 1)-th scale. Accordingly, ΔL represents the mean square error of the pixel intervals between every two adjacent scales.

[0118] Among them, the uniformity score of the scale is inversely correlated with the difference in the pixel intervals between every two adjacent scales, that is, the smaller the difference in the pixel intervals between every two adjacent scales, the higher the uniformity score of the scale.

[0119] Optionally, a mathematical relationship model between the uniformity score of the scale and the difference in the pixel intervals between every two adjacent scales can be preset. In this mathematical relationship model, the uniformity score of the scale is the dependent variable, the difference in the pixel intervals between every two adjacent scales is the independent variable, and the dependent variable is inversely correlated with the independent variable.

[0120] Alternatively, in some other embodiments, the image of the scale included in the target scale segmentation result may also be binarized to obtain a binarized image of the scale. In this binarized image, the scale is represented by black lines. Accordingly, the uniformity score of the scale can be determined based on the difference between the pixel intervals between the black lines in the binarized image of the scale. Among them, the uniformity score of the scale is inversely correlated with the difference between the pixel intervals between the black lines, that is, the smaller the difference between the pixel intervals between every two adjacent black lines, the higher the uniformity score of the scale. Among them, the difference between the pixel intervals between every two adjacent scales can be represented by the mean square deviation of the pixel intervals between every two adjacent black lines.

[0121] Optionally, a mathematical relationship model between the uniformity score of the scale and the difference between the pixel intervals between every two adjacent black lines can be preset. In this mathematical relationship model, the uniformity score of the scale is the dependent variable, the difference between the pixel intervals between every two adjacent black lines is the independent variable, and the dependent variable is inversely correlated with the independent variable.

[0122] In still some other embodiments, the number of scales between two adjacent scale numbers can be determined based on the image position information of the scale and the scale number recognition result. Specifically, the number of scales between two adjacent scale numbers can be determined based on the image position information of the scale and the image position information of the scale number. Further, the uniformity score of the scale line can be determined based on the difference between the number of scales between every two adjacent scale numbers. Among them, the uniformity score of the scale is inversely correlated with the difference between the number of scales between every two adjacent scale numbers, that is, the smaller the difference between the number of scales between every two adjacent scale numbers, the higher the uniformity score of the scale. Optionally, the difference between the number of scales between every two adjacent scale numbers can be represented by the mean square deviation of the number of scales between every two adjacent scale numbers. For example, the difference between the number of scales between every two adjacent scale numbers can be expressed as:

[0123]

[0124] In Equation (3), M represents the total number of scale numbers. i represents the i-th scale number, i = 0, 1,..., (M - 1). D i represents the number of scales between the (i + 1)-th scale number and the i-th scale number. D i+1 represents the number of scales between the (i + 2)-th scale number and the (i + 1)-th scale number. Accordingly, ΔD represents the mean square deviation of the number of scales between every two adjacent scale numbers.

[0125] Optionally, a mathematical relationship model can be preset between the uniformity score of the scale and the difference in the number of scales between every two adjacent scale numbers. In this mathematical relationship model, the uniformity score of the scale is the dependent variable, the difference in the number of scales between every two adjacent scale numbers is the independent variable, and the dependent variable is inversely correlated with the independent variable.

[0126] The implementation manner of determining the uniformity score of the scale shown in the above embodiments is only for illustrative purposes and does not constitute a limitation. In some embodiments, the reliability score of the target scale segmentation result can be determined according to the uniformity score of the scale. Among them, the reliability score of the target scale segmentation result is positively correlated with the uniformity score of the scale, that is, the higher the uniformity score of the scale (the more uniform the scale distribution), the higher the reliability score of the target scale segmentation result. In some embodiments, the uniformity score of the scale can be used as the reliability score of the target scale segmentation result. In other embodiments, a mathematical relationship model can be preset between the reliability score of the target scale segmentation result and the uniformity score of the scale. In this mathematical relationship model, the uniformity score of the scale is the independent variable, and the reliability score of the target scale segmentation result is the dependent variable, and the dependent variable is positively correlated with the independent variable. In this implementation manner, the uniformity score of the scale can be substituted into the mathematical relationship model to solve for the reliability score of the target scale segmentation result.

[0127] Embodiment 2: Determine the circular arc fitting degree of the distribution shape of the scale according to the target scale segmentation result. Specifically, according to the position distribution information of the scales included in the target scale segmentation result, circular arc fitting can be performed on the distribution shape of the scale to obtain the circular arc fitting degree of the distribution shape of the scale. For the implementation form of the circular arc fitting degree, reference can be made to the relevant content of the foregoing embodiments, which will not be elaborated here. Further, the reliability score of the target scale segmentation result can be determined according to the circular arc fitting degree of the distribution shape of the scale. Among them, the reliability score of the target scale segmentation result is positively correlated with the circular arc fitting degree of the distribution shape of the scale, that is, the larger the circular arc fitting degree of the distribution shape of the scale, the higher the reliability score of the target scale segmentation result.

[0128] In some embodiments, a mathematical relationship model can be preset between the reliability score of the target scale segmentation result and the circular arc fitting degree of the distribution shape of the scale. In this mathematical relationship model, the circular arc fitting degree of the distribution shape of the scale is the independent variable, and the reliability score of the target scale segmentation result is the dependent variable, and the dependent variable is positively correlated with the independent variable. In this implementation manner, the circular arc fitting degree of the distribution shape of the scale can be substituted into the mathematical relationship model to solve for the reliability score of the target scale segmentation result.

[0129] Embodiment 3: The arc fitting degree of the distribution shape of the above-mentioned scale and the score of the scale uniformity can be combined to determine the reliability score of the target scale segmentation result. Specifically, the reliability score of the target scale segmentation result can be determined according to the score of the scale uniformity and the arc fitting degree of the distribution shape of the scale.

[0130] In some embodiments, the score of the scale uniformity and the arc fitting degree of the distribution shape of the scale can be weighted and summed to obtain the reliability score of the target scale segmentation result. Alternatively, a mathematical relationship model between the score of the scale uniformity, the arc fitting degree of the distribution shape of the scale, and the reliability score of the target scale segmentation result can be preset in advance. In this mathematical relationship model, the score of the scale uniformity and the arc fitting degree of the distribution shape of the scale are independent variables, and the reliability score of the target scale segmentation result is the dependent variable, and the independent variable and the dependent variable are positively correlated. Based on this mathematical relationship model, the score of the scale uniformity and the arc fitting degree of the distribution shape of the scale obtained by using the above Embodiments 1 and 2 can be substituted into this mathematical relationship model for solution to obtain the reliability score of the target scale segmentation result.

[0131] The embodiments of determining the reliability score of the target scale segmentation result shown in the above embodiments are only exemplary descriptions and do not constitute limitations.

[0132] In the present application, the target image position information for scale segmentation prompt can be optimized based on the reliability score of the target scale segmentation result. Specifically, as Figure 1c shown in Step 4, the target image position information can be iteratively adjusted by an optimization method to obtain the image position information after each adjustment. For example, the target image position information can be iteratively adjusted by the gradient descent method to obtain the image position information after each adjustment.

[0133] After each adjustment of the target image position information, the pointer instrument image can be segmented with the image position information after each adjustment as the prompt information to obtain the target scale segmentation result corresponding to the image position information after each adjustment. For the specific implementation of segmenting the pointer instrument image with the image position information after each adjustment as the prompt information, reference can be made to the relevant content of segmenting the pointer instrument image with the target image position information as the prompt information described above, which will not be elaborated here.

[0134] Furthermore, the reliability score of the target scale segmentation result corresponding to the image position information after each adjustment can be determined according to the target scale segmentation result corresponding to the image position information after each adjustment. For the specific implementation of this step, reference can be made to the relevant content of determining the reliability score of the target scale segmentation result according to the target scale segmentation result described above, which will not be elaborated here.

[0135] Further, according to the reliability score of the target scale segmentation result and the reliability score of the target scale segmentation result corresponding to the target image position information after each adjustment, the target scale segmentation result with the reliability score meeting the set reliability score condition can be determined from the target scale segmentation result and the target scale segmentation result corresponding to the target image position information after each adjustment. For example, as shown in Figure 1c Step 4, the target scale segmentation result with the highest reliability score can be selected from the target scale segmentation result and the target scale segmentation result corresponding to the target image position information after each adjustment as the target scale segmentation result with the reliability score meeting the set reliability score condition, and so on.

[0136] Further, according to the target scale segmentation result with the reliability score meeting the set reliability score condition, the aforementioned scale digit recognition result, and the target pointer segmentation result, the reading of the pointer instrument can be determined. For the specific implementation manner of this step, reference can be made to the relevant content of determining the reading of the pointer instrument according to the target scale segmentation result, the aforementioned scale digit recognition result, and the target pointer segmentation result, which will not be elaborated here.

[0137] In this embodiment, through the optimization method, the scale segmentation result that makes the reliability score of the scale segmentation result meet the preset reliability score condition (such as the highest reliability score) is obtained iteratively. The reliability of the scale segmentation result reflects the accuracy of the scale segmentation. Therefore, this embodiment can improve the accuracy of the scale segmentation, and further improve the accuracy of the subsequent reading of the pointer instrument.

[0138] It should be noted that the above method for the server device 20 to read the pointer instrument based on the pointer instrument image can also be deployed on the autonomous mobile device 10, and the autonomous mobile device 10 can independently complete the processes of collecting the pointer instrument image and reading the pointer instrument based on the pointer instrument image. This process does not require the participation of the server device 20. For the specific implementation manner of the autonomous mobile device 10 to read the pointer instrument based on the pointer instrument image, reference can be made to the relevant content of the server device 20 to read the pointer instrument based on the pointer instrument image, which will not be elaborated here.

[0139] It is also worth noting that the pointer instrument reading method provided in the embodiments of the present application is not limited to reading the pointer instrument in the data center, and can be applied to reading the pointer instrument in any scenario. For example, it can read the pointer instrument in any industrial scenario. In addition, the execution subject of the pointer instrument reading method provided in the embodiments of the present application can be any device with computing functions, including but not limited to: terminal devices such as desktop computers, laptop computers, mobile phones, or Internet of Things devices; it can also be various server devices such as traditional servers, cloud servers, or server clusters.

[0140] It should be noted that the execution subject of each step of the method provided in the above embodiments can be the same device, or the method can also be executed by different devices. For example, the execution subject of steps 101 and 102 can be device A; for another example, the execution subject of step 101 can be device A, and the execution subject of step 102 can be device B; and so on.

[0141] In addition, in some of the processes described in the above embodiments and the accompanying drawings, there are multiple operations that appear in a specific order. However, it should be clearly understood that these operations can be executed not in the order in which they appear in this document or in parallel. The operation numbers such as 101, 102, etc. are only used to distinguish different operations, and the numbers themselves do not represent any execution order. In addition, these processes can include more or fewer operations, and these operations can be executed in sequence or in parallel.

[0142] Correspondingly, the embodiments of the present application also provide a computer-readable storage medium storing computer instructions, which when executed by one or more processors, cause the one or more processors to execute the steps in the above pointer instrument reading method.

[0143] Figure 6 It is a schematic structural diagram of the electronic device provided by the embodiments of the present application. As Figure 6 shown, the electronic device may include: a memory 60a and a processor 60b. Among them, the memory 60a is used to store computer programs.

[0144] The processor 60b is coupled to the memory 60a and is used to execute the computer program to execute the steps in the pointer instrument reading method provided in the foregoing embodiments. For the specific implementation manners of each step, reference can be made to the relevant descriptions in the foregoing embodiments, which will not be elaborated here.

[0145] In some alternative embodiments, as Figure 6 shown, the electronic device may further include: a communication component 60c, a power supply component 60d, a display component 60e, an audio component 60f and other optional components. In some embodiments, the electronic device may further include: an image acquisition device for acquiring an image of the pointer instrument to obtain an image of the pointer instrument, etc. Figure 6 Only some components are schematically shown, which does not mean that the electronic device must include Figure 6 all the components shown, nor does it mean that the electronic device can only include Figure 6 the components shown.

[0146] In addition, Figure 6The components within the dashed box are optional components, rather than essential components, and can be determined according to the product form of the specific visual electronic device. The electronic device in this embodiment can be implemented as a terminal device such as a desktop computer, a laptop computer, a mobile phone, or an Internet of Things device; it can also be various server devices such as a traditional server, a cloud server, or a server cluster. The electronic device can also be implemented as an autonomous mobile device. Correspondingly, it may further include: a driving component ( Figure 6 not shown in the figure) and the like. Among them, the driving component may include a driving wheel, a driving motor, a caster wheel, etc. The basic components included in different autonomous mobile devices and the composition of the basic components will vary. The examples listed in the embodiments of this application are only partial examples.

[0147] In the embodiments of this application, the memory is used to store computer programs and can be configured to store various other data to support operations on the device where it is located. Among them, the processor can execute the computer programs stored in the memory to implement corresponding control logics. The memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Electrical Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc.

[0148] In the embodiments of the present application, the processor may be any hardware processing device capable of executing the above method logic. Optionally, the processor may be a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), or a Microcontroller Unit (MCU); it may also be a programmable device such as a Field-Programmable Gate Array (FPGA), a Programmable Array Logic (PAL), a General Array Logic (GAL), or a Complex Programmable Logic Device (CPLD); or it may be an Advanced Reduced Instruction Set Compute (RISC) processor (Advanced RISC Machines, ARM) or a System on Chip (SoC), etc., but not limited thereto.

[0149] In the embodiments of the present application, the communication component is configured to facilitate communication between the device where it is located and other devices in a wired or wireless manner. The device where the communication component is located may access a wireless network based on a communication standard, such as Wireless Fidelity (WiFi), 2G or 3G, 4G, 5G, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component may also be implemented based on Near Field Communication (NFC) technology, Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wide Band (UWB) technology, Bluetooth (BT) technology, or other technologies.

[0150] In the embodiments of the present application, the display component may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the display component includes a touch panel, the display component can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also detect the duration and pressure associated with the touch or swipe operations.

[0151] In the embodiments of the present application, the power supply component is configured to provide power to various components of the device where it is located. The power supply component may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the device where the power supply component is located.

[0152] In the embodiments of the present application, the audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC). When the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive external audio signals. The received audio signals can be further stored in a memory or transmitted via a communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals. For example, for a device with a language interaction function, voice interaction with a user can be implemented through the audio component.

[0153] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data that have been authorized by the user or fully authorized by all parties. Moreover, the collection, use, and processing of relevant data need to comply with the relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or reject.

[0154] It also should be noted that the descriptions such as "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0155] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, compact disc read-only memory (CD-ROM), optical memory, etc.) containing computer-usable program code.

[0156] The present application is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in one or more flows in the flowchart and / or one or more blocks in the block diagram.

[0157] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in one or more flows in the flowchart and / or one or more blocks in the block diagram.

[0158] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Therefore, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows in the flowchart and / or one or more blocks in the block diagram.

[0159] In a typical configuration, a computing device includes one or more processors (such as CPUs), an input / output interface, a network interface, and a memory.

[0160] The memory may include non-permanent memory in the computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0161] The storage medium of a computer is a readable storage medium, also known as a readable medium. The readable storage medium includes permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of the storage medium of a computer include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other memory technologies, CD-ROM, digital versatile disc (DVD), or other optical storage, magnetic cassette tapes, disk storage, or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media do not include transitory media such as modulated data signals and carrier waves.

[0162] It should also be noted that the term "including", "comprising", or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "including a..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the above element.

[0163] The above content is only an embodiment of the present application and is not used to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for reading pointer instruments, It is characterized in that include: Performing scale digital recognition on a pointer instrument image including a dial image of a pointer instrument to determine a scale digital recognition result of the pointer instrument; Performing pointer detection on the pointer instrument image to determine image position information of a detection frame of a pointer of the pointer instrument; Using the image position information of the detection frame of the pointer as prompt information, performing image segmentation on the pointer instrument image to obtain a target pointer segmentation result; Determine the target image position information for scale segmentation prompt according to the target pointer segmentation result and the scale number recognition result; Using the target image position information as prompt information, performing image segmentation on the pointer instrument image to obtain a target scale segmentation result; The reading of the pointer instrument is determined according to the scale digit recognition result, the target pointer segmentation result and the target scale segmentation result.

2. The method according to claim 1, It is characterized in that The step of performing scale digital recognition on the pointer instrument image including the dial image of the pointer instrument to determine the scale digital recognition result of the pointer instrument includes: Performing character recognition on the pointer instrument image by using optical character recognition technology to determine the characters contained in the pointer instrument image and the image position information of the characters; Filtering the scale numbers from the characters included in the pointer instrument image according to the characters included in the pointer instrument image and the image position information of the characters; The image position information of the scale numerals is obtained from the image position information of the characters; the scale numeral recognition result includes: the scale numerals and the image position information of the scale numerals.

3. The method according to claim 2, It is characterized in that The step of selecting the scale numbers from the characters included in the pointer instrument image according to the characters included in the pointer instrument image and the image position information of the characters includes: Determining position distribution characteristics of the character according to the image position information of the character; Determining the logical relationship between the characters according to the characters included in the pointer instrument image; The scale numbers are screened out from the characters included in the pointer instrument image according to the position distribution characteristics of the characters and the logical relationship between the characters.

4. The method according to claim 3, It is characterized in that The step of selecting the scale numbers from the characters included in the pointer instrument image according to the position distribution characteristics of the characters and the logical relationship between the characters includes: Filtering out characters whose position distribution characteristics are distributed on the same arc from the characters included in the pointer instrument image; From the characters distributed on the same arc, characters whose logical relationship is an arithmetic progression are selected as the scale numbers.

5. The method according to claim 1, It is characterized in that The step of performing image segmentation on the pointer instrument image using the image position information of the pointer as prompt information to obtain a target pointer segmentation result includes: Using the image position information of the detection frame of the pointer as prompt information and the pointer instrument image as the image to be segmented to input into the image segmentation model; In the image segmentation model, the pointer instrument image and the image position information of the pointer detection frame are respectively encoded to obtain an image feature vector of the pointer instrument image and a prompt feature vector of the image position information of the pointer detection frame; Performing mask mapping on the image feature vector of the pointer instrument image and the prompt feature vector of the image position information of the detection frame of the pointer to obtain a mask map of the pointer; Performing image segmentation on the pointer instrument image according to the mask image of the pointer to determine a pointer segmentation result; The target pointer segmentation result is determined from the pointer segmentation results.

6. The method according to claim 5, It is characterized in that The pointer segmentation result includes: a plurality of pointer segmentation results of different granularities; the pointer segmentation result includes: image position information of the pointer; The step of determining the target pointer segmentation result from the pointer segmentation result includes: determining the lengths of the pointers segmented by the multiple pointer segmentation results according to the image position information of the pointers respectively included in the multiple pointer segmentation results; From the plurality of pointer segmentation results, a pointer segmentation result whose length of the segmented pointer meets a set length condition is selected as the target pointer segmentation result.

7. The method according to claim 1, It is characterized in that The target pointer segmentation result includes: the image position information of the pointer; the scale number recognition result includes: the image position information of the scale number; the determining the target image position information for the scale segmentation prompt according to the pointer segmentation result and the scale number recognition result includes: Determining image position information of a first circular arc formed by the rotation of the pointer according to the image position information of the pointer; Determining image position information of a second circular arc formed by the scale numbers according to the image position information of the scale numbers; determining a target image region between the first arc and the second arc from the pointer instrument image according to the image position information of the first arc and the image position information of the second arc; A plurality of pixel coordinates are selected from the pixel coordinates of the target image area as the target image position information.

8. The method according to claim 7, It is characterized in that The step of selecting the plurality of pixel coordinates from the pixel coordinates of the target image area comprises: Randomly select a set number of pixel coordinates from the pixel coordinates of the target image area; the set number is greater than or equal to 2 and is an integer; or, Determine pixel coordinates located on a third circular arc having a radius equal to a target radius from pixel coordinates of the target image area; and randomly select the plurality of pixel coordinates from pixel coordinates located on the third circular arc; or, Determine the pixel coordinates of a third circular arc having a radius equal to the target radius from the pixel coordinates of the target image area; select the plurality of pixel coordinates according to the set center angle interval from the pixel coordinates of the third circular arc; The target radius is larger than the radius of the first arc and smaller than the radius of the second arc, or the target radius is larger than the radius of the second arc and smaller than the radius of the first arc.

9. The method according to claim 1, It is characterized in that The step of performing image segmentation on the pointer instrument image by using the target image position information as prompt information to obtain a scale segmentation result includes: Using the target image position information as prompt information and the pointer instrument image as the image to be segmented to input an image segmentation model; In the image segmentation model, the pointer instrument image and the target image position information are encoded respectively to obtain an image feature vector of the pointer instrument image and a prompt feature vector of the target image position information; Performing mask mapping on the image feature vector of the pointer instrument image and the prompt feature vector of the target image position information to obtain a mask map of the scale of the pointer instrument; According to the mask image of the scale, performing image segmentation on the pointer instrument image to determine the scale segmentation result; The target scale segmentation result is determined from the scale segmentation results.

10. The method according to claim 9, It is characterized in that The scale segmentation result includes: a plurality of scale segmentation results of different granularities; the scale segmentation result includes: image position information of the scale and / or an image of the scale; Determining the target scale segmentation result from the scale segmentation results includes: According to the image position information of the scales respectively included in the multiple scale segmentation results, arc fitting is performed on the scale distribution shapes corresponding to the multiple scale segmentation results to obtain the arc fitting degrees of the scale distribution shapes respectively corresponding to the multiple scale segmentation results; according to the arc fitting degrees of the scale distribution shapes respectively corresponding to the multiple scale segmentation results, a scale segmentation result whose arc fitting degree satisfies a set arc fitting degree condition is selected from the multiple scale segmentation results as the target scale segmentation result; and / or, From images of scales respectively contained in the multiple scale segmentation results, identify blank areas corresponding to the multiple scale segmentation results, and calculate the areas of the blank areas corresponding to the multiple scale segmentation results; based on the areas of the blank areas corresponding to the multiple scale segmentation results, select a scale segmentation result from the multiple scale segmentation results whose area of ​​the blank area satisfies a set area condition as the target scale segmentation result.

11. The method according to any one of claims 1 to 10, It is characterized in that Determining the reading of the pointer instrument according to the scale digit recognition result, the target pointer segmentation result and the target scale segmentation result includes: Determining a reliability score of the target scale segmentation result according to the target scale segmentation result; Iteratively adjusting the target image position information through an optimization method to obtain the image position information after each adjustment; Using the image position information after each adjustment as prompt information, performing image segmentation on the pointer instrument image to obtain a target scale segmentation result corresponding to the image position information after each adjustment; Determine the reliability score of the target scale segmentation result corresponding to the image position information after each adjustment according to the target scale segmentation result corresponding to the image position information after each adjustment; According to the reliability score of the target scale segmentation result and the reliability score of the target scale segmentation result corresponding to the target image position information after each adjustment, determine the target scale segmentation result whose reliability score meets the set reliability score condition from the target scale segmentation result and the target scale segmentation result corresponding to the target image position information after each adjustment; The reading of the pointer instrument is determined according to the target scale segmentation result that meets the set reliability score condition, the scale number recognition result and the target pointer segmentation result.

12. The method according to claim 11, It is characterized in that Determining the reliability score of the target scale segmentation result according to the target scale segmentation result includes: Determining a uniformity score of the scale according to the target scale segmentation result; determining a reliability score of the target scale segmentation result according to the uniformity score of the scale; or, Determining the arc fitting degree of the distribution shape of the scale according to the target scale segmentation result; determining the reliability score of the target scale segmentation result according to the arc fitting degree of the distribution shape of the scale; or, Determining the reliability score of the target scale segmentation result according to the uniformity score of the scale and the arc fitting degree of the distribution shape of the scale; Among them, the reliability score of the target scale segmentation result is positively correlated with the uniformity score of the scale; the reliability score of the target scale segmentation result is positively correlated with the arc fitting degree of the distribution shape of the scale.

13. The method according to claim 12, It is characterized in that The target scale segmentation result includes: image position information of the scale and / or an image of the scale; and determining the uniformity score of the scale according to the target scale segmentation result includes: Determine the pixel spacing between adjacent scales according to the image position information of the scale; determine the uniformity score of the scale according to the difference in the pixel spacing between every two adjacent scales; wherein the uniformity score of the scale is inversely correlated with the difference in the pixel spacing between every two adjacent scales; or, Binarizing the image of the scale to obtain a binary image of the scale; determining a uniformity score of the scale according to a difference between pixel intervals between black lines in the binary image of the scale; wherein the uniformity score of the scale is inversely correlated with the difference between pixel intervals between black lines; or, The number of scales between two adjacent scale numbers is determined based on the image position information of the scale and the scale number recognition result; the uniformity score of the scale line is determined based on the difference in the number of scales between every two adjacent scale numbers; wherein the uniformity score of the scale is inversely correlated with the difference in the number of scales between every two adjacent scale numbers.

14. The method according to claim 11, It is characterized in that The step of determining the reading of the pointer instrument according to the target scale segmentation result that satisfies the set reliability score condition, the scale number recognition result, and the target pointer segmentation result comprises: Determining the measuring range of the pointer instrument according to the scale digital recognition result; Determining the graduation value of the pointer instrument according to the target scale segmentation result that meets the set reliability score condition and the measuring range of the pointer instrument; Determining the target scale pointed to by the pointer according to the target scale segmentation result that meets the set reliability score condition and the target pointer recognition result; According to the graduation value of the pointer instrument and the target scale, a reading corresponding to the target scale is determined as the reading of the pointer instrument.

15. A data center inspection system, It is characterized in that include: Autonomous mobile devices and server devices; The autonomous mobile device is provided with an image acquisition device; The autonomous mobile device is used to move in the data center, and during the movement, control the image acquisition device to acquire images of the pointer instrument in the data center to obtain a pointer instrument image including a dial image of the pointer instrument; and provide the pointer instrument image to the server device; The server device is used to execute the steps in the method according to any one of claims 1-14.

16. An electronic device, It is characterized in that include: A memory and a processor; wherein the memory is used to store a computer program; The processor is coupled to the memory and configured to execute the computer program to perform the steps of the method according to any one of claims 1 to 14.

17. A computer-readable storage medium storing computer instructions, It is characterized in that When the computer instructions are executed by one or more processors, the one or more processors are caused to execute the steps in the method according to any one of claims 1 to 14.

Images