Instrument device fault detection method, device, equipment and storage medium

Abstract

The application discloses an instrument equipment fault detection method, device and equipment and a storage medium, and belongs to the technical field of fault detection. The application carries out angle prediction on a real-time acquired current instrument pointer image, obtains a current pointer angle value, determines a pointer angle change rate set within a preset time range based on an obtained historical pointer angle value and the current pointer angle value, and carries out fault detection through a preset fault detection model according to the angle value change rate set. Compared with the prior art, the application carries out angle prediction on a real-time acquired current instrument pointer image, obtains a current pointer angle value, so that the instrument pointer value can be acquired in real time, and the accuracy is higher. According to the obtained pointer angle value, the accuracy of judging whether the equipment is faulty is higher, and the technical problem that the real-time performance and accuracy cannot be guaranteed when the equipment is judged to be faulty by manually reading the instrument reading and is easily interfered by other factors is avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault detection, in particular to an instrument equipment fault detection method, device, equipment and storage medium. BACKGROUND

[0002] There are a large number of instruments in many fields such as substations to record the on-site conditions and production data. Due to environmental factors, most of the industrial production adopts pointer type instruments. The reading recognition of the pointer type instrument mainly relies on professional personnel to conduct on-site inspection regularly. This work mode is greatly affected by factors such as human eye resolution and staff proficiency, and the detection efficiency is low and the accuracy cannot be effectively guaranteed. In the environment of high radiation, strong noise and the like, it is impossible to guarantee the real-time performance and accuracy by reading the instrument reading to determine whether the equipment fails, which is easily disturbed by other factors.

[0003] The above content is only used to assist in understanding the technical solutions of the present application and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0004] The main purpose of the present application is to provide an instrument equipment fault detection method, device, equipment and storage medium, which aims to solve the technical problems that the existing technology cannot guarantee real-time performance and accuracy by manually reading instrument readings to determine whether the equipment fails, which is easily disturbed by other factors.

[0005] To achieve the above purpose, the present application provides an instrument equipment fault detection method, which comprises the following steps:

[0006] Obtain the current instrument pointer image of the instrument equipment to be detected;

[0007] Predict the pointer angle of the current instrument pointer image by a preset pointer angle prediction model to obtain the current pointer angle value;

[0008] Determine the pointer angle change rate set in the preset time range based on the historical pointer angle value and the current pointer angle value;

[0009] Perform fault detection according to the angle value change rate set by a preset fault detection model.

[0010] Optionally, the step of predicting the pointer angle of the current instrument pointer image by a preset pointer angle prediction model to obtain the current pointer angle value comprises:

[0011] Generate a pointer prediction box in the area where the pointer is located in the current instrument pointer image by a preset pointer angle prediction model;

[0012] Determine the current pointer angle value based on the pointer prediction box.

[0013] Optionally, the step of determining the current pointer angle value based on the prediction frame comprises:

[0014] determining a rotation center of the pointer in the current instrument pointer image according to the positional relationship between the pointer prediction frame and the current instrument pointer image;

[0015] establishing a coordinate system based on the rotation center and the current pointer instrument image, and obtaining a diagonal line vertex coordinate corresponding to the rotation center in the pointer prediction frame;

[0016] determining the pointer angle value according to the diagonal line vertex coordinate and the rotation center coordinate.

[0017] Optionally, after the step of establishing a coordinate system based on the rotation center and the current pointer instrument image, and obtaining a diagonal line vertex coordinate corresponding to the rotation center in the pointer prediction frame, the method further comprises:

[0018] determining a corresponding vertex coordinate quadrant based on the position of the diagonal line vertex coordinate in the coordinate system;

[0019] determining the pointer angle value according to the diagonal line vertex coordinate quadrant and the diagonal line vertex coordinate.

[0020] Optionally, before the step of performing pointer position prediction on the labeled frame image according to the preset pointer position prediction model to obtain a pointer prediction frame, the method further comprises:

[0021] obtaining instrument pointer image dataset samples;

[0022] receiving a labeling instruction input by a user based on the instrument pointer image dataset samples, labeling the instrument pointer image dataset samples based on the labeling instruction to obtain labeled dataset samples;

[0023] performing model training on an initial YOLOv4 model based on the labeled dataset samples to obtain a preset pointer angle prediction model.

[0024] Optionally, before the step of performing fault detection on the angle value change rate set through a preset fault detection model, the method further comprises:

[0025] extracting an angle change rate sample set from the angle change rate set;

[0026] performing model training on an initial LSTM model based on the angle change rate sample set to obtain a preset fault detection model.

[0027] Optionally, after the step of predicting the current pointer angle value by the preset pointer angle prediction model, the method further comprises:

[0028] acquiring a current time corresponding to the current pointer angle value;

[0029] when the current time reaches a preset time, performing the step of determining the set of pointer angle change rates within the preset time range based on the historical pointer angle values and the current pointer angle value.

[0030] In addition, to achieve the above object, the present application further provides an instrument device fault detection device, which comprises:

[0031] an image acquisition module, configured to acquire a current instrument pointer image of an instrument device to be detected;

[0032] an angle prediction module, configured to predict a current pointer angle value by predicting a pointer angle of the current instrument pointer image according to a preset pointer angle prediction model;

[0033] a data processing module, configured to determine a set of pointer angle change rates within a preset time range based on historical pointer angle values and the current pointer angle value;

[0034] a fault detection module, configured to perform fault detection by a preset fault detection model according to the set of angle values.

[0035] In addition, to achieve the above object, the present application further provides an instrument device fault detection device, which comprises a memory, a processor and an instrument device fault detection program stored in the memory and executable on the processor, and the instrument device fault detection program is configured to implement the steps of the instrument device fault detection method as described above.

[0036] In addition, to achieve the above object, the present application further provides a storage medium, which stores an instrument device fault detection program, and the instrument device fault detection program is executed by a processor to implement the steps of the instrument device fault detection method as described above.

[0037] The present application obtains the current instrument pointer image of the instrument device to be detected, and predicts the pointer angle of the current instrument pointer image through a preset pointer angle prediction model to obtain a current pointer angle value, determines a pointer angle change rate set in a preset time range based on the obtained historical pointer angle value and the current pointer angle value, and finally detects the fault through a preset fault detection model according to the angle value change rate set. Compared with the manual on-site inspection of the instrument pointer in the prior art, the present application obtains the current instrument pointer image in real time, predicts the angle of the current instrument pointer image, and obtains the current pointer angle value, so that the instrument pointer value can be obtained in real time, and the pointer angle value obtained through the image processing is more accurate. Therefore, the fault detection is performed through the preset fault detection model according to the more accurate real-time pointer angle value, the accuracy of detecting whether the equipment fails is higher, and the technical problem that the real-time performance and accuracy cannot be guaranteed by manually reading the instrument reading to determine whether the equipment fails and is easily disturbed by other factors is avoided. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is a structural schematic diagram of the instrument device fault detection device of the hardware running environment related to the embodiment scheme of the present application.

[0039] Figure 2 is a flowchart of the first embodiment of the instrument device fault detection method of the present application.

[0040] Figure 3 is a flowchart of the second embodiment of the instrument device fault detection method of the present application.

[0041] Figure 4 is an instrument pointer schematic diagram of an embodiment of the instrument device fault detection method of the present application.

[0042] Figure 5 is a structural block diagram of the first embodiment of the instrument device fault detection apparatus of the present application.

[0043] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0044] It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0045] Reference Figure 1 , Figure 1 is a structural schematic diagram of the instrument device fault detection device of the hardware running environment related to the embodiment scheme of the present application.

[0046] As Figure 1As shown in the figure, the instrument device fault detection device can include a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize the connection communication between the components. The user interface 1003 can include a display, an input unit such as a keyboard, and can also include a standard wired interface, a wireless interface. The network interface 1004 can optionally include a standard wired interface, a wireless interface (such as a wireless fidelity (Wi-Fi) interface). The memory 1005 can be a high-speed random access memory (RAM) memory, or a stable non-volatile memory (NVM) such as a disk memory. The memory 1005 can also be a storage device independent of the aforementioned processor 1001.

[0047] Those skilled in the art can understand that Figure 1 The structure shown in the figure does not constitute a limitation on the instrument device fault detection device, and can include more or fewer components than the figure, or combine certain components, or different component arrangements.

[0048] As Figure 1 As shown, the memory 1005 as a storage medium can include an operating system, a network communication module, a user interface module, and an instrument device fault detection program.

[0049] In Figure 1 As shown in the instrument device fault detection device, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the instrument device fault detection device can be arranged in the instrument device fault detection device, and the instrument device fault detection device calls the instrument device fault detection program stored in the memory 1005 through the processor 1001, and executes the instrument device fault detection method provided by the embodiment of the application.

[0050] The embodiment of the application provides an instrument device fault detection method, which refers to Figure 2 , Figure 2 The flowchart of the first embodiment of the instrument device fault detection method of the application.

[0051] In this embodiment, the instrument device fault detection method includes the following steps:

[0052] Step S10: acquire a current instrument pointer image of an instrument device to be detected;

[0053] It should be noted that the execution subject of the method in the embodiment is the instrument device fault detection device, wherein the instrument device fault detection device can be an electronic device such as a personal computer or a server, and can also be other devices that can achieve the same or similar functions, which are not limited in the embodiment. In the embodiment, a server is taken as an example for illustration.

[0054] It should be noted that the current instrument pointer image can be an instrument pointer image of a device to be detected for fault detection acquired by an image acquisition device, that is, the current instrument pointer image can be an instrument pointer image of a device to be detected acquired at a preset frequency. For example, if the preset frequency is 100 HZ, 100 instrument pointer images of the device to be detected need to be acquired within 1 s, which are not limited in the embodiment.

[0055] The device to be detected can be a device with a pointer instrument in a substation that needs to be detected for fault detection, such as a voltmeter, a barometer, a transformer, etc. in the substation, which are not limited in the embodiment.

[0056] Step S20: predicting a pointer angle of the current instrument pointer image by using a preset pointer angle prediction model to obtain a current pointer angle value;

[0057] It should be noted that the preset pointer angle prediction model is a model for predicting a pointer angle of a current instrument pointer image to obtain a current pointer angle value. The preset image processing model can be an angle prediction program based on deep learning, or other models with an angle prediction program, which are not limited in the embodiment.

[0058] Further, in order to obtain the preset pointer angle prediction model, before step S20, the method further comprises:

[0059] acquiring instrument pointer image dataset samples;

[0060] receiving a labeling instruction input by a user based on the instrument pointer image dataset samples, labeling the instrument pointer image dataset samples based on the labeling instruction to obtain labeled dataset samples;

[0061] training an initial YOLOv4 model based on the labeled dataset samples to obtain the preset pointer angle prediction model.

[0062] It should be noted that after obtaining the instrument pointer image dataset sample, a reminder information can be generated to remind the user to perform instrument detection, the user makes an operation instruction according to the reminder information, the server receives the operation instruction of the user, analyzes the operation instruction to obtain a labeling instruction, and labels the instrument pointer image dataset sample based on the labeling instruction to obtain a labeled dataset sample.

[0063] It can be understood that the labeled dataset sample can be obtained by recording the center of the instrument pointer in the instrument pointer image as one vertex of the labeling box, and recording the other end of the instrument pointer as the other vertex of the labeling box, and generating a rectangular box with the two vertices as the labeling box. The image sample with the labeling box in the instrument pointer image dataset sample is recorded as the labeled dataset sample.

[0064] It should be noted that after obtaining the labeled dataset sample, the labeled dataset sample is trained by an initial neural network model to obtain a preset pointer angle prediction model. The neural network model can be a yolov4 neural network in a You Only Look Once (YOLO) target detection network, or other neural network models with pointer angle prediction function, which are not limited in the embodiment.

[0065] Step S30: determining a pointer angle change rate set in a preset time range based on the historical pointer angle value and the current pointer angle value;

[0066] It should be noted that the historical pointer angle value is the pointer angle value collected in the past period of time. After collecting the current pointer angle value each time, the current pointer angle value can be stored in a pre-allocated storage area for subsequent data processing. The pre-allocated storage area can be a server cloud or a local server, which is not limited in the embodiment.

[0067] It can be understood that since the current instrument pointer image is collected according to the preset frequency, each current instrument pointer image has a time interval, that is, each time the pointer angle of the current instrument pointer image is predicted to obtain the current pointer angle value, there is a time interval. Through the plurality of time intervals and the plurality of pointer angle values, the pointer angle change rate in each time interval can be determined to obtain the pointer angle change rate set.

[0068] Step S40: performing fault detection according to the angle value change rate set through a preset fault detection model.

[0069] It should be noted that the preset fault detection model can be a model for judging whether the corresponding instrument equipment fails according to the obtained angle rate set, the preset image processing model can be an angle prediction program based on deep learning, or other models with angle prediction program, which is not limited in the embodiment.

[0070] Further, in order to obtain the preset fault detection model, before step S40, comprising:

[0071] extracting the angle rate sample set in the angle rate set;

[0072] training the initial LSTM model based on the angle rate sample set to obtain the preset fault detection model.

[0073] It should be noted that the preset fault detection model contains the mapping relationship between the angle value rate set and whether the fault occurs, and the mapping relationship needs to be trained by statistical data, and the mapping relationship in the preset fault detection model can be updated according to the data result after training, therefore, the initial neural network model needs to select a functional neural network with memory information for a certain time, for example: long short-term memory recurrent convolutional network (Long Short-Term Memory, LSTM) and the like.

[0074] It can be understood that after obtaining the angle rate sample set, the initial neural network model is trained according to the angle rate sample set to obtain the preset fault detection model; the initial neural network model can be a long short-term memory recurrent convolutional network (Long Short-Term Memory, LSTM), or other neural network model with fault detection function, which is not limited in the embodiment.

[0075] It should be noted that when the angle value rate set is detected according to the preset fault detection model, if the detection result is that the current instrument equipment to be detected has a fault, a warning information can be generated and sent to the client, the client can be a WeChat applet, an APP and the like, the application program has the interface of data storage, display, operation instruction receiving and the like; the warning information can be one or more of the reminding information such as text reminder, vibration reminder, light flickering reminder, which is not limited in the embodiment.

[0076] The embodiment obtains a current instrument pointer image of an instrument device to be detected, predicts a pointer angle of the current instrument pointer image through a preset pointer angle prediction model, obtains a current pointer angle value, determines a pointer angle change rate set in a preset time range based on a historical pointer angle value and the current pointer angle value, and finally performs fault detection through a preset fault detection model according to the angle value change rate set. Compared with the prior art of manually checking an instrument pointer in the field, the embodiment can obtain a current instrument pointer image in real time, predict an angle of the current instrument pointer image, and obtain a current pointer angle value, so that the instrument pointer value can be obtained in real time, and the pointer angle value obtained through image processing is more accurate. Therefore, the fault detection model is used to perform fault detection according to the more accurate real-time pointer angle value, so that the accuracy of detecting whether a device fails is higher, and the technical problem that the real-time performance and accuracy cannot be guaranteed when the device failure is determined by manually reading an instrument reading and is easily disturbed by other factors is avoided.

[0077] Reference Figure 3 , Figure 3 FIG. 2 is a flowchart of a second embodiment of the instrument device fault detection method.

[0078] Based on the first embodiment, in the embodiment, the step S20 comprises:

[0079] Step S201: generating a pointer prediction box in a region where a pointer is located in the current instrument pointer image through a preset pointer angle prediction model.

[0080] It should be noted that the current instrument pointer image is subjected to the preset pointer angle prediction model, so that a position of a pointer region in the current instrument pointer image generates a rectangular box in quadrature, which is recorded as a prediction box. The nearest vertex position to the bottom 1 / 2 of the picture is determined according to four vertex positions of the prediction box, and the predicted rotation center is positioned.

[0081] It should be noted that when the current instrument pointer image is subjected to pointer angle prediction, the generated prediction box may deviate from the actual pointer region. In the embodiment, an intersection over union (IOU) threshold value can be set. The IOU threshold value is used to determine whether the position and size of the prediction box deviate too much from the actual pointer region. If the intersection over union ratio of the prediction box is greater than or equal to the IOU threshold value, the prediction box is considered to be valid. If the intersection over union ratio of the prediction box is less than the IOU threshold value, the prediction box is considered to be invalid.

[0082] Step S202: determining a current pointer angle value based on the pointer prediction box.

[0083] It should be noted that, when the pointer prediction frame is obtained, the rotation center of the pointer in the current instrument pointer image can be determined according to the positional relationship between the pointer prediction frame and the current instrument pointer image, a coordinate system is established based on the rotation center and the current instrument pointer image, the diagonal vertex coordinates corresponding to the rotation center in the pointer prediction frame are obtained, and the pointer angle value is determined according to the diagonal vertex coordinates and the rotation center coordinates.

[0084] It can be understood that, in the process of determining the pointer angle value through the diagonal vertex coordinates and the rotation center coordinates, the reference Figure 4 When the pointer is in the position as shown in the figure, it is possible to select the wrong pointer angle value according to the diagonal vertex coordinates and the rotation center coordinates to determine the pointer angle value, for example: the prediction frame as shown in the figure is generated according to the current instrument pointer image through the preset pointer angle prediction model, and according to the rotation direction of the instrument pointer, the diagonal vertex coordinates and the rotation center coordinates, it can be determined that the current pointer angle value should be 225°, for the server, it is possible to output two current pointer angle values of 225° and 135°, which affects the subsequent data processing and fault judgment result.

[0085] Further, in order to accurately determine the current pointer angle value, after step S202, the method further comprises:

[0086] determining the corresponding vertex coordinate quadrant based on the position of the diagonal vertex coordinates in the coordinate system;

[0087] determining the pointer angle value according to the diagonal vertex coordinate quadrant and the diagonal vertex coordinates.

[0088] It should be noted that, according to the diagonal vertex coordinates in the coordinate system, the corresponding vertex coordinate quadrant is determined, and in different quadrants, the confirmation method of the current pointer angle value can be different, for example: in the second and third quadrants, the smaller value of the pointer angle value is selected as the current pointer angle value; in the first and fourth quadrants, the larger value of the pointer angle value is selected as the current pointer angle value.

[0089] It should be noted that, since there are multiple pointer angle change rates in the pointer angle change rate set, it is necessary to accumulate the pointer angle values within the preset time, so as to accumulate the pointer angle change rate within the preset time; the preset time can be 1s, 5s, etc., and the pointer angle change rate within the preset time can also be set according to the fault judgment requirement, which is not limited in the embodiment.

[0090] It can be understood that when the current pointer angle value is acquired, the corresponding current time can be recorded, and when the current time reaches the preset moment, the step of determining the pointer angle change rate set in the preset time range based on the historical pointer angle value and the current pointer angle value is performed.

[0091] The embodiment obtains a current instrument pointer image of the instrument device to be detected, generates a pointer prediction box in the current instrument pointer image through a preset pointer angle prediction model, obtains a current pointer angle value through the pointer prediction box, determines a pointer angle change rate set in a preset time range based on the obtained historical pointer angle value and the current pointer angle value, and finally performs fault detection through a preset fault detection model according to the angle value change rate set. Compared with the manual on-site inspection of the instrument pointer in the prior art, the embodiment obtains a current instrument pointer image in real time, generates a pointer prediction box in the current instrument pointer image, and obtains a current pointer angle value through the pointer prediction box, so that the instrument pointer value can be more accurately obtained. The pointer angle value obtained through image processing is more accurate, and therefore, the fault detection model is used to perform fault detection according to the more accurate real-time pointer angle value, so that the accuracy of detecting whether the equipment fails is higher, and the technical problem that the real-time performance and accuracy cannot be guaranteed when the equipment failure is determined by manually reading the instrument reading and is easily disturbed by other factors is avoided.

[0092] In addition, the embodiment of the present application also provides a storage medium, and the storage medium stores an instrument device fault detection program. When the instrument device fault detection program is executed by a processor, the steps of the instrument device fault detection method described above are implemented.

[0093] Since the storage medium adopts all the technical solutions of the above-mentioned embodiments, at least all the beneficial effects brought by the technical solutions of the above-mentioned embodiments are achieved, which will not be repeated here.

[0094] Reference Figure 5 , Figure 5 The figure is a structural block diagram of the first embodiment of the instrument device fault detection device of the present application.

[0095] As Figure 5 shown, the instrument device fault detection device provided by the embodiment of the present application comprises:

[0096] An image acquisition module is configured to acquire a current instrument pointer image of an instrument device to be detected.

[0097] It should be noted that the execution subject of the method in the embodiment is the instrument equipment fault detection device, wherein the instrument equipment fault detection device can be an electronic device such as a personal computer or a server, and can also be other devices that can achieve the same or similar functions, which are not limited in the embodiment. In the embodiment, a server is taken as an example for illustration.

[0098] It should be noted that the current instrument pointer image can be an instrument pointer image of a to-be-detected device that needs to be detected for fault detection, that is, the current instrument pointer image can be an instrument pointer image of a to-be-detected device collected at a preset frequency. For example, if the preset frequency is 100 HZ, 100 instrument pointer images of the to-be-detected device need to be collected within 1 s, which is not limited in the embodiment.

[0099] The to-be-detected device can be a device with a pointer instrument in a substation that needs to be detected for fault detection, such as a voltmeter, a barometer, a transformer, etc. in the substation, which is not limited in the embodiment.

[0100] The angle prediction module is configured to predict a pointer angle of the current instrument pointer image by using a preset pointer angle prediction model to obtain a current pointer angle value.

[0101] It should be noted that the preset pointer angle prediction model is a model for predicting a pointer angle of a current instrument pointer image to obtain a current pointer angle value. The preset image processing model can be an angle prediction program based on deep learning, or can be other models with an angle prediction program, which are not limited in the embodiment.

[0102] Further, in order to obtain the preset pointer angle prediction model, before step S20, the method further includes:

[0103] Obtaining instrument pointer image dataset samples;

[0104] Receiving a labeling instruction input by a user based on the instrument pointer image dataset samples, labeling the instrument pointer image dataset samples based on the labeling instruction, and obtaining labeled dataset samples;

[0105] Model training is performed on the initial YOLOv4 model based on the labeled dataset samples to obtain a preset pointer angle prediction model.

[0106] It should be noted that after obtaining the instrument pointer image dataset samples, a reminder information can be generated to remind the user to perform instrument detection. The user makes an operation instruction according to the reminder information. The server analyzes the operation instruction to obtain a labeling instruction, and labels the instrument pointer image dataset samples based on the labeling instruction to obtain labeled dataset samples.

[0107] It can be understood that the obtaining of the annotated dataset sample can be that the rotation center of the instrument pointer in the instrument pointer image is recorded as one vertex of the annotation box, and in addition, the other end of the instrument pointer is recorded as another vertex of the annotation box, to generate a rectangular box orthogonal to the two vertices, recorded as an annotation box; the image sample in which the annotation box exists in the instrument pointer image dataset sample is recorded as an annotated dataset sample.

[0108] It should be noted that after obtaining the annotated dataset sample, the annotated dataset sample is trained by an initial neural network model to obtain a preset pointer angle prediction model. The neural network model can be a yolov4 neural network in a You Only Look Once (YOLO) target detection network, and can also be other neural network models with a pointer angle prediction function, which are not limited by the embodiment.

[0109] The data processing module is configured to determine a set of pointer angle change rates in a preset time range based on the historical pointer angle value and the current pointer angle value.

[0110] It should be noted that the historical pointer angle value is a pointer angle value collected in the past period of time, and after each current pointer angle value is collected, the current pointer angle value can be stored in a pre-allocated storage area for subsequent data processing. The pre-allocated storage area can be a server cloud or a local server, and the embodiment is not limited thereto.

[0111] It can be understood that since the current instrument pointer image is collected according to the preset frequency, each current instrument pointer image has a time interval, that is, each time the pointer angle of the current instrument pointer image is predicted to obtain the current pointer angle value, there is a time interval. Through a plurality of time intervals and a plurality of pointer angle values, the pointer angle change rate in each time interval can be determined to obtain a set of pointer angle change rates.

[0112] The fault detection module is configured to perform fault detection according to the set of angle change rates by a preset fault detection model.

[0113] It should be noted that the preset fault detection model can be a model for judging whether the corresponding instrument device fails according to the obtained angle change rate set. The preset image processing model can be a deep learning-based angle prediction program, or other models with an angle prediction program, which are not limited by the embodiment.

[0114] Further, in order to obtain the preset fault detection model, before step S40, the method further comprises:

[0115] extracting a set of angle change rate samples in the set of angle change rates;

[0116] model training is performed on the initial LSTM model based on the set of angle change rate samples to obtain a preset fault detection model.

[0117] It is worth noting that the preset fault detection model contains a mapping relationship between the set of angle value change rates and whether a fault occurs. The mapping relationship needs to be model trained through statistical data, and the mapping relationship in the preset fault detection model can be updated according to the data results after training. Therefore, the initial neural network model needs to select a functional neural network that has memory for a certain time constant information, for example, a long short-term memory (LSTM) recurrent convolutional network.

[0118] It can be understood that after obtaining the set of angle change rate samples, the initial neural network model is trained according to the set of angle change rate samples to obtain a preset fault detection model. The initial neural network model can be a long short-term memory (LSTM) recurrent convolutional network, or other neural network models with fault detection functions. The present embodiment does not limit this.

[0119] It should be noted that when the set of angle value change rates is detected for faults according to the preset fault detection model, if the detection result is that the current instrument device to be detected has a fault, a warning information can be generated and sent to a client. The client can be a WeChat applet, an APP, or other application programs. The application program has an interface with functions of data storage, display, operation instruction receiving, etc. The warning information can be one or more of text reminders, vibration reminders, light flickering reminders, etc. The present embodiment does not limit this.

[0120] The embodiment obtains a current instrument pointer image of an instrument device to be detected, predicts a pointer angle of the current instrument pointer image through a preset pointer angle prediction model, obtains a current pointer angle value, determines a pointer angle change rate set in a preset time range based on a historical pointer angle value and the current pointer angle value, and finally performs fault detection through a preset fault detection model according to the angle value change rate set. Compared with the prior art of manually checking an instrument pointer in the field, the embodiment can obtain a current instrument pointer image in real time, predict an angle of the current instrument pointer image, and obtain a current pointer angle value, so that the instrument pointer value can be obtained in real time, and the pointer angle value obtained through image processing is more accurate. Therefore, the fault detection model is used to perform fault detection according to the more accurate real-time pointer angle value, so that the accuracy of detecting whether a device fails is higher, and the technical problem that the real-time performance and accuracy cannot be guaranteed when a device failure is determined by manually reading an instrument reading and is easily disturbed by other factors is avoided.

[0121] In an embodiment, the angle prediction module 20 is further configured to generate a pointer prediction box in a region where a pointer is located in the current instrument pointer image through a preset pointer angle prediction model; and determine a current pointer angle value based on the pointer prediction box.

[0122] In an embodiment, the angle prediction module 20 is further configured to determine a rotation center of the pointer in the current instrument pointer image according to a positional relationship between the pointer prediction box and the current instrument pointer image; establish a coordinate system based on the rotation center and the current instrument pointer image; and obtain a diagonal line vertex coordinate corresponding to the rotation center in the pointer prediction box; and determine a pointer angle value according to the diagonal line vertex coordinate and the rotation center coordinate.

[0123] In an embodiment, the angle prediction module 20 is further configured to determine a corresponding vertex coordinate quadrant based on a position of the diagonal line vertex coordinate in the coordinate system; and determine a pointer angle value according to the diagonal line vertex coordinate quadrant and the diagonal line vertex coordinate.

[0124] In an embodiment, the angle prediction module 20 is further configured to obtain an instrument pointer image dataset sample; receive a labeling instruction input by a user based on the instrument pointer image dataset sample; label the instrument pointer image dataset sample based on the labeling instruction to obtain a labeled dataset sample; and perform model training on an initial YOLOv4 model based on the labeled dataset sample to obtain a preset pointer angle prediction model.

[0125] In an embodiment, the fault detection module 40 is further configured to extract a set of angle change rate samples from the set of angle change rates; and perform model training on an initial LSTM model based on the set of angle change rate samples to obtain a preset fault detection model.

[0126] In an embodiment, the instrument device fault detection apparatus further includes a time accumulation module configured to obtain a current time corresponding to the current pointer angle value; and perform the step of determining the set of pointer angle change rates within the preset time range based on the historical pointer angle values and the current pointer angle value when the current time reaches a preset time.

[0127] It should be understood that the above is only illustrative, and does not constitute any limitation on the technical solutions of the present application. In specific applications, those skilled in the art can set up according to the needs, and the present application does not limit this.

[0128] It should be noted that the above-described workflow is only illustrative and does not limit the scope of protection of the present application. In actual applications, those skilled in the art can select part or all of them to achieve the purpose of the present embodiment scheme according to actual needs, which is not limited here.

[0129] In addition, technical details not described in detail in the present embodiment can be referred to the instrument device fault detection method provided by any embodiment of the present application, which will not be described here.

[0130] In addition, it should be noted that in this paper, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, article or system. Without more limitations, the element defined by the statement "includes a" does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0131] The above-mentioned embodiment numbers of the present application are only for description, not representing the advantages and disadvantages of the embodiments.

[0132] Those skilled in the art can clearly understand the above-mentioned embodiment method can be realized by means of software and the necessary general hardware platform, of course, can also be through hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application essentially or say the part of the prior art contribution can be embodied in the form of software products, the computer software product is stored in a storage medium (such as read only memory (Read Only Memory, ROM) / RAM, disk, optical disk), including a number of instructions to make a terminal device (may be a mobile phone, computer, server, or network equipment, etc.) executes the method described in various embodiments of the present application.

[0133] The above is only the preferred embodiment of the present application, not therefore limit the patent scope of the present application, any equivalent structure or equivalent flow transformation made by using the content of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. An instrument equipment failure detection method characterized by, The instrument equipment fault detection method comprises: acquiring a current instrument pointer image of an instrument equipment to be detected; predicting a pointer angle of the current instrument pointer image by a preset pointer angle prediction model to obtain a current pointer angle value; determining a set of pointer angle change rates within a preset time range based on a historical pointer angle value and the current pointer angle value; performing fault detection by a preset fault detection model according to the set of angle value change rates; the step of predicting the pointer angle of the current instrument pointer image by the preset pointer angle prediction model to obtain the current pointer angle value comprises: generating a pointer prediction box at a region where a pointer is located in the current instrument pointer image by the preset pointer angle prediction model; determining a rotation center of the pointer in the current instrument pointer image according to a positional relationship between the pointer prediction box and the current instrument pointer image, specifically comprising: determining a minimum distance between four vertex positions of the pointer prediction box and a position at 1 / 2 of the bottom of the current instrument pointer image according to the positional relationship between the pointer prediction box and the current instrument pointer image, and taking the vertex position corresponding to the minimum distance as the rotation center; establishing a coordinate system based on the rotation center and the current pointer instrument image, and acquiring diagonal line vertex coordinates corresponding to the rotation center in the pointer prediction box; determining the current pointer angle value according to the diagonal line vertex coordinates and the rotation center coordinates.

2. The meter device fault detection method of claim 1, wherein, After the step of establishing the coordinate system based on the rotation center and the current pointer instrument image, and acquiring the diagonal line vertex coordinates corresponding to the rotation center in the pointer prediction box, the method further comprises: determining a corresponding vertex coordinate quadrant based on the position of the diagonal line vertex coordinates in the coordinate system; determining the pointer angle value according to the diagonal line vertex coordinate quadrant and the diagonal line vertex coordinates.

3. The meter device fault detection method of claim 1, wherein, Before the step of generating the pointer prediction box at the region where the pointer is located in the current instrument pointer image by the preset pointer angle prediction model, the method further comprises: acquiring instrument pointer image dataset samples; receiving a labeling instruction input by a user based on the instrument pointer image dataset samples, labeling the instrument pointer image dataset samples based on the labeling instruction, and obtaining labeled dataset samples; training an initial YOLOv4 model based on the labeled dataset samples to obtain a preset pointer angle prediction model.

4. The meter device fault detection method of any one of claims 1-3, wherein, Before the step of performing fault detection by the preset fault detection model according to the set of angle value change rates, the method further comprises: extracting a set of angle change rate samples in the set of angle change rates; training an initial LSTM model based on the set of angle change rate samples to obtain a preset fault detection model.

5. The meter device fault detection method of any one of claims 1-3, wherein, After the step of predicting the pointer angle of the current instrument pointer image by the preset pointer angle prediction model to obtain the current pointer angle value, the method further comprises: acquiring a current time corresponding to the current pointer angle value; When the current time reaches a preset moment, the step of determining a set of pointer angle change rates within a preset time range based on the historical pointer angle value and the current pointer angle value is performed.

6. An apparatus for detecting faults in metering devices, characterized by The instrument device fault detection apparatus comprises: An image acquisition module is configured to acquire a current instrument pointer image of an instrument device to be detected. An angle prediction module is configured to perform pointer angle prediction on the current instrument pointer image according to a preset pointer angle prediction model to obtain a current pointer angle value. A data processing module is configured to determine a set of pointer angle change rates within a preset time range based on a historical pointer angle value and the current pointer angle value. A fault detection module is configured to perform fault detection by a preset fault detection model according to the set of angle values. The angle prediction module is further configured to generate a pointer prediction box in a region where a pointer is located in the current instrument pointer image by a preset pointer angle prediction model. A rotation center of the pointer in the current instrument pointer image is determined according to a positional relationship between the pointer prediction box and the current instrument pointer image. A coordinate system is established based on the rotation center and the current pointer instrument image, and a diagonal line vertex coordinate corresponding to the rotation center in the pointer prediction box is acquired. A current pointer angle value is determined according to the diagonal line vertex coordinate and the rotation center coordinate. The angle prediction module is further configured to determine a minimum distance between four vertex positions of the pointer prediction box and a position at 1 / 2 of a bottom of the current instrument pointer image according to a positional relationship between the pointer prediction box and the current instrument pointer image, and take the vertex position corresponding to the minimum distance as the rotation center.

7. An instrument equipment failure detection apparatus, characterized by, The instrument device fault detection apparatus comprises a memory, a processor, and an instrument device fault detection program stored on the memory and executable on the processor, the instrument device fault detection program being configured to implement the instrument device fault detection method according to any one of claims 1 to 5.

8. A storage medium, characterized by The storage medium stores an instrument device fault detection program, and the instrument device fault detection program is executed by the processor to implement the instrument device fault detection method according to any one of claims 1 to 5.

Images