Anomaly detection methods, electronic devices, storage media and software products

Abstract

This application discloses an anomaly detection method, electronic device, storage medium, and program product, relating to the field of computer technology. It includes determining a first weight based on a first temperature and a first position detected by multiple temperature detection devices, determining the influence of each temperature detection device on the temperature of multiple preset positions on a first board, and generating a first image of the target point of the temperature anomaly and the positional relationship between the target point and each device based on the multiple first temperatures, the first weight, and the second positions of each device on the board. Therefore, it can effectively avoid the detection blind spot problem caused by the limited deployment positions of temperature sensors in existing technologies, which cannot cover the entire area of ​​the board. It also solves the technical problem of low efficiency, requiring manual inspection of heating conditions in locations where sensors are not deployed. The method achieves the technical effect of automatically reconstructing the temperature distribution across the entire board, accurately locating the correlation between abnormal heating points and devices, and improving the efficiency and accuracy of anomaly detection.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to an anomaly detection method, electronic device, storage medium, and program product. Background Technology

[0002] As a critical hardware component of servers, abnormal heat generation of printed circuit boards (PCBs) is a major factor leading to server failures and service interruptions. To ensure the safe operation of servers, it is necessary to detect abnormal heat generation of PCB boards.

[0003] In related technologies, temperature sensors are typically used for anomaly detection. This involves deploying multiple temperature sensors on the PCB board to collect temperature data at specific points, and then using this data to detect anomalies. However, in this process, if heat is detected in locations on the PCB board where no temperature sensors are deployed, manual inspection and location are required, resulting in low efficiency for anomaly detection. Summary of the Invention

[0004] This application provides an anomaly detection method, electronic device, storage medium, and program product to at least solve the problem of low efficiency in anomaly detection of circuit boards in related technologies.

[0005] This application provides an anomaly detection method, including:

[0006] The system acquires the first temperature detected by multiple temperature detection devices on the first board with temperature anomalies, the first position of each temperature detection device on the first board, and the second position of each component on the first board.

[0007] Based on multiple first temperatures and multiple first positions, a first weight corresponding to each temperature detection device is determined. The first weight is used to indicate the degree of influence of the temperature detected by the temperature detection device on the temperature of multiple preset positions on the first board.

[0008] Based on multiple first temperatures, multiple first weights, and second positions, a first image of the target point of the temperature anomaly is determined, as well as the positional relationship between the target point and each device;

[0009] Based on the first image and positional relationship, the location of the temperature anomaly in the first board is determined.

[0010] This application also provides an anomaly detection device, comprising: an acquisition module, a first determination module, a second determination module, and a processing module, wherein,

[0011] The acquisition module is used to acquire the first temperature detected by multiple temperature detection devices on the first board with temperature anomalies, the first position of each temperature detection device on the first board, and the second position of each device in the first board.

[0012] The first determining module is used to determine the first weight corresponding to each temperature detection device based on multiple first temperatures and multiple first positions. The first weight is used to indicate the degree of influence of the temperature detected by the temperature detection device on the temperature of multiple preset positions on the first board.

[0013] The second determining module is used to determine, based on multiple first temperatures, multiple first weights, and a second position, a first image of the target point of the temperature anomaly and the positional relationship between the target point and each device.

[0014] The processing module is used to determine the location of the temperature anomaly in the first board based on the first image and the positional relationship.

[0015] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for implementing the steps of any of the above-described anomaly detection methods when executing the computer program.

[0016] This application also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of any of the above-described anomaly detection methods.

[0017] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above-described anomaly detection methods.

[0018] The anomaly detection method, electronic device, storage medium, and program product provided in this application can determine the first weight of the influence of each temperature detection device on the temperature of multiple preset positions on a first board based on the first temperature and first position detected by multiple temperature detection devices. Furthermore, based on the multiple first temperatures, the first weight, and the second positions of each device on the board, a first image of the target point of the temperature anomaly and the positional relationship between the target point and each device are generated. Therefore, it can effectively avoid the detection blind spot problem caused by the limited deployment positions of temperature sensors in the prior art, which cannot cover the entire area of ​​the board. It also solves the technical problem of low efficiency, requiring manual inspection of heating conditions in locations where no sensors are deployed. This achieves the technical effect of automatically reconstructing the temperature distribution across the entire board, accurately locating the correlation between abnormal heating points and devices, and improving the efficiency and accuracy of anomaly detection. Attached Figure Description

[0019] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the system architecture provided for an embodiment of this application;

[0021] Figure 2 A flowchart illustrating an anomaly detection method provided in an embodiment of this application;

[0022] Figure 3 A schematic diagram of a neural network structure provided in an embodiment of this application;

[0023] Figure 4 A schematic diagram illustrating a method for determining the positional relationship between a first image, target points, and various devices, provided in an embodiment of this application;

[0024] Figure 5 This is a schematic diagram illustrating a method for determining the location of a temperature anomaly according to an embodiment of this application;

[0025] Figure 6 A flowchart illustrating another anomaly detection method provided in an embodiment of this application;

[0026] Figure 7 This is a schematic diagram of the structure of an anomaly detection device provided in an embodiment of this application;

[0027] Figure 8 A schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, other embodiments obtained by those of ordinary skill in the art without creative effort are all within the protection scope of this application.

[0029] It should be noted that, in the description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. The terms "first," "second," etc., in this application are used to distinguish similar objects and are not used to describe a specific order or sequence.

[0030] In related technologies, temperature sensors are typically used for anomaly detection. This involves deploying multiple temperature sensors on the PCB board to collect temperature data at specific points, and then using this data to detect anomalies. However, in this process, if heat is detected in locations on the PCB board where no temperature sensors are deployed, manual inspection and location are required, resulting in low efficiency for anomaly detection.

[0031] To address the aforementioned issues, in this embodiment, when anomaly detection is required, the system acquires the first temperatures detected by multiple temperature detection devices on a first board with temperature anomalies, the first positions of each temperature detection device on the first board, and the second positions of each component on the first board. Based on the multiple first temperatures and multiple first positions, a first weight corresponding to each temperature detection device is determined. The first weight indicates the degree of influence of the temperature detected by the temperature detection device on the temperatures of multiple preset positions on the first board. Based on the multiple first temperatures, multiple first weights, and second positions, a first image of the target point of the temperature anomaly and the positional relationship between the target point and each component are determined. Based on the first image and positional relationship, the location of the temperature anomaly in the first board is determined. In this way, by dynamically associating temperature detection data with board structure information, the temperature state of areas without deployed temperature sensors can be automatically fitted. That is, the temperature distribution of the entire board can be reconstructed based on data from limited temperature sensors, automatically identifying abnormal heating in areas without deployed temperature sensors. This effectively avoids the detection blind spot problem caused by limited sensor coverage in traditional methods, achieving accurate positioning of temperature anomaly points on the board and component correlation analysis. This improves the efficiency of anomaly detection and enhances the accuracy and reliability of fault determination.

[0032] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] This section describes the specific application environment architecture or hardware architecture upon which the anomaly detection method depends. (References) Figure 1 , Figure 1 This is a schematic diagram of the system architecture provided for an embodiment of this application. Please refer to [link / reference]. Figure 1 This includes electronic devices and circuit boards. The electronic devices can be any device with on-device computing capabilities, such as servers or terminal devices. The circuit boards include multiple temperature detection devices. The electronic devices can acquire relevant data from the multiple temperature detection devices on the circuit board, process this data, and obtain images containing temperature data and location data at various locations on the circuit board, thereby completing anomaly detection.

[0034] Figure 2This is a flowchart illustrating an anomaly detection method provided in an embodiment of this application, as shown below. Figure 2 As shown, embodiments of this application provide an anomaly detection method, which is described in detail below:

[0035] S201. Obtain the first temperature detected by multiple temperature detection devices on the first board with temperature anomalies, the first position of each temperature detection device on the first board, and the second position of each device in the first board.

[0036] The execution subject of this application embodiment can be an electronic device or an anomaly detection device installed in the electronic device. The anomaly detection device can be implemented by software or by a combination of software and hardware.

[0037] The first board can be the PCB board that triggers a temperature alarm. That is, the electronic device can identify the first board through the alarm message. The alarm message is used to indicate that the current PCB board is overheating abnormally. The alarm message records the alarm information corresponding to the fault or abnormality generated by the current PCB board. In this way, the electronic device can identify the PCB board that triggers the alarm message as the first board with abnormal temperature.

[0038] The temperature detection device is used to detect the temperature at specific points on the first board. For example, the temperature monitoring device can be a high-precision temperature sensor.

[0039] The first temperature can be the temperature data collected by each temperature detection device on the first board. That is, the first temperature is used to indicate the temperature at the corresponding position on the first board. For example, the first temperature can be 30°C.

[0040] In some embodiments, the first board includes multiple temperature detection devices. For any one of the temperature detection devices, the temperature of key points in the first board can be periodically collected and the collected temperature signal can be sent to the electronic device. The electronic device obtains the temperature data collected by each temperature detection device, i.e., the first temperature, based on the received temperature signal.

[0041] The first position can be the location of each temperature detection device on the first board, that is, the first position is used to indicate the location of each key component on the first board. For example, the first position can be the location of the Central Processing Unit (CPU) or the location of the Graphics Processing Unit (GPU).

[0042] In some embodiments, the electronic device can obtain a preset configuration file, which is a pre-set file used to store the location information of temperature detection devices in the board; and parse the preset configuration file to obtain the location information of each temperature detection device on the first board, i.e., the first position.

[0043] Optionally, the electronic device can obtain the identifier of the first board, which indicates the type of the board; and query a preset association library based on the identifier of the first board to obtain the location information of each temperature detection device corresponding to the identifier of the first board, i.e., the first position. The preset association library is pre-set and is used to indicate the mapping relationship between the identifier of the board and the first position.

[0044] The second position can be the deployment position of each component (e.g., CPU, GPU, power supply, etc.) in the first board. The second position includes the position of each component and the positional relationship between each component. The second position can be the structural parameter of the first board, that is, the second position is used to indicate the physical characteristics of the first board.

[0045] In some embodiments, the electronic device can obtain the second position of each device in the first board according to a preset parameter library. The preset parameter library is used to store the device positions of each board. Specifically, the electronic device can match the device structure information corresponding to the type of the first board in the preset parameter library according to the board type, i.e., the second position.

[0046] In some embodiments, the first and second positions can be represented by a board coordinate system. The origin of the coordinate system can be set at a specific corner of the board (e.g., the lower left corner), and the coordinate unit can be millimeters, depending on the application scenario and accuracy requirements. After acquiring this position information, the electronic device establishes a spatial mapping relationship between the temperature detection points and the device layout, providing basic data support for subsequent temperature distribution reconstruction and anomaly localization.

[0047] S202. Determine the first weight corresponding to each temperature detection device based on multiple first temperatures and multiple first positions.

[0048] The first weight is used to indicate the degree of influence of the temperature detected by the temperature detection device on the temperatures of multiple preset locations on the first board. For example, the larger the first weight, the greater the influence of the temperature detected by the temperature detection device on the temperatures of multiple preset locations on the first board; the smaller the first weight, the smaller the influence of the temperature detected by the temperature detection device on the temperatures of multiple preset locations on the first board.

[0049] The preset location can be the location of the point on the first board that needs to be detected for anomalies. That is, the preset location can include the first location of multiple temperature detection devices, as well as the location of points where no temperature detection devices are deployed.

[0050] In some embodiments, the electronic device can use a grid division method to determine multiple preset positions on the first board. Specifically, the electronic device can perform a global grid division on the effective area of ​​the first board (including the deployment area of ​​each device and the area where abnormal heat may occur) according to the size and layout of the first board, and determine the center point or vertex of each grid as the preset position. In this way, the entire area of ​​the first board can be completely covered, ensuring the comprehensiveness and accuracy of abnormal detection at each preset position of the first board.

[0051] Optionally, the electronic device can determine multiple preset locations based on the key areas of the first board. Specifically, the electronic device needs to determine the key heat-generating areas on the first board and perform further meshing based on the key heat-generating areas. For example, for the CPU heat-generating area, the mesh size of the area can be reduced, or the mesh can be further divided within each mesh to ensure that the heat-generating locations in the area are covered. Furthermore, the electronic device can sparsely set multiple preset locations based on non-critical heat-generating areas. For example, multiple preset locations can be set within the CPU area, and several preset locations can be set in the gap area between the CPU and the GPU to ensure that the heat-generating status of the device itself and adjacent areas is the focus.

[0052] In some embodiments, the electronic device may determine a first weight corresponding to each temperature detection device based on a plurality of first temperatures and a plurality of first positions by: determining a first position of the temperature detection device and a first distance between it and each preset position; and determining a first weight corresponding to the temperature detection device based on the plurality of first distances and the first temperature detected by the temperature detection device.

[0053] The first distance is used to indicate the spatial relationship between the temperature detection device and each preset position.

[0054] In some embodiments, the electronic device can determine the first distance using coordinate system mapping. Specifically, the electronic device can establish a coordinate system with a specific corner point (e.g., the lower left corner) of the first board as the origin, and map the first position and preset position of each temperature detection device to this coordinate system to obtain corresponding coordinate data. Based on this coordinate data, the first distance is determined. Optionally, the electronic device can determine the first position of the temperature detection device and the first distance between each preset position based on the grid size and each preset position.

[0055] In some embodiments, the electronic device may determine the degree of influence of each temperature detection device on the temperature of each preset location, i.e., the first weight, based on the distance relationship between the temperature detection device and the preset location, and the temperature relationship between the temperature detection device and the preset location.

[0056] In some embodiments, the electronic device repeats the above steps to calculate the first distance between the temperature detection device and each preset position on the board in sequence. For other temperature detection devices, the same first distance determination process is performed to obtain the first distance between each temperature detection device and each preset position.

[0057] For example, if the device includes two temperature detection devices, namely temperature detection device A and temperature detection device B, and the first board includes three preset positions, namely preset position A of temperature detection device A, preset position B and preset position C of temperature detection device B, then the electronic device can determine the first distance of temperature detection device A to preset position A, preset position B and preset position C, and the first distance of temperature detection device B to preset position A, preset position B and preset position C.

[0058] In some embodiments, the electronic device may determine a first weight corresponding to a temperature detection device based on a plurality of first distances and a first temperature detected by a temperature detection device in the following manner: for any preset position; determine a first sub-weight corresponding to the temperature detection device based on the first distance corresponding to the preset position; determine a second sub-weight based on the first temperature; and determine a first weight corresponding to the temperature detection device at the preset position based on the first sub-weight and the second sub-weight.

[0059] The first sub-weight is used to indicate the correlation between the first distance and the temperature at the preset location. For example, the larger the first sub-weight, the stronger the correlation between the first distance and the temperature at the preset location; the smaller the first sub-weight, the weaker the correlation.

[0060] The second sub-weight is used to indicate the correlation between the first temperature and the temperature at the preset location. For example, a larger second sub-weight indicates a strong correlation between the first temperature and the temperature at the preset location, while a smaller second sub-weight indicates a weak correlation.

[0061] For example, if the first sub-weight corresponding to the temperature detection device is greater than the second sub-weight, the electronic device can determine that, in the current state, the correlation between the first distance and the temperature at the preset position is greater than the correlation between the first temperature and the temperature at the preset position.

[0062] In some embodiments, the first sub-weight and the second sub-weight can be preset values, and their sum is 1. Specifically, the electronic device can determine the first sub-weight and the second sub-weight based on the determination method that the closer the distance and the higher the temperature, the greater the corresponding sub-weight ratio. Optionally, the first sub-weight and the second sub-weight can be preset values, and their values ​​are both in the range of [0, 1].

[0063] For example, if the first board includes a temperature detection device A, and the first board includes two preset positions, namely preset position A and preset position B of the temperature detection device A. Preset position A and preset position B are relatively close in a first distance, and the first temperature A of the temperature detection device A is relatively low. This indicates that the position of the temperature detection device A has a greater influence on preset position B than the first temperature A has on preset position B; that is, the first sub-weight is greater than the second sub-weight. Therefore, the electronic device can set the first sub-weight of the temperature detection device A on preset position B to 0.8 and the second sub-weight to 0.2.

[0064] In some embodiments, the electronic device may determine the first weight corresponding to the temperature detection device at a preset position based on the following implementation: determining the weighting coefficients corresponding to the first temperature and the first distance; determining the first weight based on the weighting coefficient of the first distance, the first sub-weight, the weighting coefficient of the first temperature, and the second sub-weight.

[0065] The weighting coefficients for the first temperature and the first distance are preset values. The weighting coefficient for the first temperature is used to indicate the degree of influence of the first temperature on the preset position, and the weighting coefficient for the first distance is used to indicate the degree of influence of the first distance on the preset position. For example, if the degree of influence of the first temperature on the preset position is greater than the degree of influence of the first distance on the preset position, then the weighting coefficient for the first temperature is greater than the weighting coefficient for the first distance.

[0066] For example, the first board includes two temperature detection devices, namely temperature detection device A and temperature detection device B, and the first board includes three preset positions: preset position A of temperature detection device A, preset position B of temperature detection device B, and preset position C. The first sub-weight of temperature detection device A to preset position C is 0.8 and the second sub-weight is 0.2, and the first sub-weight of temperature detection device B to preset position C is 0.6 and the second sub-weight is 0.4.

[0067] If the influence of temperature on distance is significantly greater for temperature detection device A and preset position C, then the weighting coefficient of the first temperature can be determined to be 0.7, the weighting coefficient of the first distance can be 0.3, and the first weight of temperature detection device A on preset position C can be 0.8*0.7+0.2*0.3=0.62.

[0068] If the distance between temperature detection device B and preset position C has a slightly greater impact than temperature, then the weighting factor for the first temperature can be determined to be 0.3, the weighting factor for the first distance can be 0.7, and the first weight of temperature detection device B for preset position C can be 0.6*0.3+0.4*0.7=0.46.

[0069] S203. Based on multiple first temperatures, multiple first weights, and second positions, determine the first image of the target point of the temperature anomaly, as well as the positional relationship between the target point and each device.

[0070] The first image is used to determine the points in the first board where the temperature is abnormal. That is, the first image can include points with normal temperature and points with abnormal temperature, i.e., target points. In this way, the electronic device can accurately detect abnormalities in the board based on the first image.

[0071] In some embodiments, for any preset position, the electronic device can determine the temperature of the preset position based on the first temperature of each temperature detection device corresponding to the preset position and the first weight of each temperature detection device, repeat the above operation to obtain the temperature of each preset position on the first board, and then obtain the first image of the target point of temperature anomaly and the positional relationship between the target point and each device.

[0072] For example, if the first board includes two temperature detection devices, namely temperature detection device A and temperature detection device B, and the first board includes three preset positions, preset position A of temperature detection device A, preset position B of temperature detection device B and preset position C, if for preset position C, the first temperature of temperature detection device A is 30°C and the corresponding first weight is 0.5, and the first temperature of temperature detection device B is 50°C and the corresponding first weight is 0.8, then the temperature of preset position C can be expressed as 30°C*0.5+50°C*0.8=55°C.

[0073] In some embodiments, the electronic device can determine whether the temperature at each preset location is abnormal based on a preset threshold. If the temperature at the preset location is greater than or equal to the preset threshold, it is determined that the temperature at the preset location is abnormal. If the temperature at the preset location is less than the preset threshold, it is determined that the temperature at the preset location is not abnormal. In this way, a first image of the target point with abnormal temperature can be obtained.

[0074] In some embodiments, the electronic device can determine the positional relationship between the target point and each device based on the preset position and the connection or contact relationship between the devices in the first board. For example, if the preset position is next to the CPU position in the first board, the target point can be determined to be at the CPU position. This can help the user to locate the specific position of the board in a timely manner and improve the accuracy and efficiency of temperature anomaly detection of the first board.

[0075] S204. Based on the first image and positional relationship, determine the location of the temperature anomaly in the first board.

[0076] The location of the abnormal temperature is the point where the first board malfunctioned.

[0077] In some embodiments, after the electronic device obtains the first image and positional relationship, the electronic device can filter target points that do not cause the first board to malfunction based on at least one abnormal target point and positional relationship, and determine the location of the temperature abnormality in the first board.

[0078] In some embodiments, the electronic device can perform anomaly detection on the first board based on a pre-trained neural network model, including determining the first weights corresponding to each temperature detection device, determining the first image of the target point of the temperature anomaly, and the positional relationship between the target point and each device. Specifically, the electronic device can input data of the first board into the neural network model, including the first temperature detected by multiple temperature detection devices on the first board with the temperature anomaly, the first position of each temperature detection device on the first board, and the second position of each device in the first board. Based on the input data, the neural network model can determine the first weights corresponding to each temperature detection device, and output a temperature distribution map (first image) of the first board based on multiple first weights and the first temperature, combined with the structural parameters of the first board, and display the location of the temperature anomaly to the user.

[0079] During training, the electronic device can acquire temperature and location data of each temperature detection device on multiple boards with normal temperatures, temperature and location data of each temperature detection device on multiple boards with abnormal temperatures, and temperature distribution maps of each board; perform image feature processing on multiple temperature distribution maps to obtain feature-processed temperature distribution maps; perform fitting processing on multiple temperature data, multiple location data, and multiple feature-processed temperature distribution maps to obtain fitted temperature distribution maps; and train the neural network model based on the fitted temperature distribution maps.

[0080] The training dataset for the model includes thermal imaging temperature distribution maps of normal and abnormal circuit boards, as well as data and heating conditions from temperature detection devices on the boards. The heating conditions of abnormal circuit boards include the board's temperature data and the location of abnormal heating, while the heating conditions of normal circuit boards include the temperature data at each detection location. The thermal imaging temperature distribution maps are obtained by scanning the surface of the circuit boards using devices such as thermal imaging scanners.

[0081] Specifically, the image feature processing involves: after the electronic device acquires the dataset, performing consistency processing on the spatial resolution, brightness, and contrast of the thermal imaging temperature distribution map in the dataset, retaining only the main features of the temperature distribution as much as possible, such as the brightness (color depth) of each region in the thermal imaging temperature distribution map, and abnormal areas of the image. At the same time, the location and data of the temperature sensor on the corresponding board are acquired to ensure that the data of each board in the dataset includes the temperature distribution map with retained main features, the location of the detection device on the board, and temperature data information.

[0082] The processed dataset is then input into a neural network model for training. This allows the model to obtain the mapping relationship between the location of the temperature detection device on the board, temperature data, and temperature distribution. The model can then fit the temperature distribution map of the board based on the location of the temperature detection device and the data. During training, the model is trained and its parameters are adjusted through multiple iterations. Appropriate training parameters, such as the learning rate and the number of iterations, are set. Based on the difference between the model output and the true label, the weights and biases of each part of the model are adjusted in real time, so that the model output gradually approaches the feature value of abnormal heating, ultimately achieving accurate identification of subtle features.

[0083] Based on the above process, electronic devices can train a neural network model using the thermal imaging temperature distribution map of the faulty board and the temperature sensor data on the board as a dataset. The thermal imaging temperature distribution map is fitted to the temperature sensor data, enabling the model to output the overall temperature distribution map of the board and locate abnormal heat points. When it is necessary to detect the heating status of a target board, the temperature data (first temperature) collected by the temperature detection device of the first board is input into the trained neural network model for analysis, realizing the detection and location of abnormal heat points. Using a neural network model to automatically identify board temperature data improves detection efficiency. During model training, the thermal imaging temperature distribution map in the dataset is further processed for consistency and fitted to the temperature data on the board. The model accuracy is improved based on the specific board shape, ensuring model performance. The neural network model significantly reduces reliance on manual inspection, avoids subjective errors, and achieves high-precision identification of the board's heating status.

[0084] Below, through Figure 3 The structure of the neural network will be explained. Figure 3This diagram illustrates a neural network structure provided in an embodiment of this application. The neural network structure can be a convolutional neural network architecture, including an input layer, convolutional layers, pooling layers, fully connected layers, and an output layer. This network is an encoder-decoder structure, taking a 64×64 single-channel image as input. The image sequentially passes through a first convolutional layer C1 (32 5×5 convolutional kernels, stride 1), a first pooling layer P1 (2×2 max pooling, stride 1), a second convolutional layer C2 (32 5×5 convolutional kernels, stride 1), and a second pooling layer P2 (2×2 max pooling, stride 1) for multi-level feature extraction and compression. After passing through a fully connected layer and undergoing dropout regularization to prevent overfitting, the image passes through a fully connected layer containing three neurons and then outputs a complete temperature distribution reconstruction image through a Softmax activation function. This network achieves accurate mapping from sparse sensor data to global temperature distribution through end-to-end learning, possessing multi-scale feature capture capabilities and anti-overfitting characteristics, effectively identifying abnormal heating areas and locating fault points.

[0085] Based on this, it can be concluded that electronic devices, by employing a convolutional neural network architecture, can extract image features from temperature distribution maps and combine them with feature information from the location and data of the board's temperature sensors. This allows the model's output to gradually approximate the feature values ​​of abnormal heating, ultimately achieving accurate identification of subtle features. This learning ability of neural networks can accurately capture minute changes in the board's temperature distribution and make timely judgments, outputting the location information of abnormal heating points on the board.

[0086] By utilizing a neural network model to extract and learn features from the thermal imaging temperature distribution map of the circuit board, and continuously adjusting the parameters of the neural network, the model establishes a mapping relationship between the temperature distribution and the location and data of the temperature sensors on the board. This ensures that the diagnostic accuracy of the model reaches the expected target. Furthermore, by preprocessing the training samples to obtain the feature information of the temperature data, the feature information of abnormal circuit board temperature data is input into the trained model, and finally, the model's temperature distribution map is fitted to output the corresponding hardware circuit board status category. This clearly determines whether the hardware circuit board has abnormal heating and the possible type of abnormality, providing strong support for timely hardware maintenance and fault diagnosis. Simultaneously, newly emerging abnormal heating data from circuit boards is periodically input into the model to continuously optimize the model structure and parameters, making the model more applicable to different situations and complex circuit boards, enhancing the applicability of the neural network model to complex circuit boards, and consistently maintaining a high resolution and location accuracy.

[0087] exist Figure 2In the illustrated embodiment, when anomaly detection is required, the system acquires the first temperature detected by multiple temperature detection devices on the first board with the temperature anomaly, the first position of each temperature detection device on the first board, and the second position of each component on the first board. Based on the multiple first temperatures and multiple first positions, a first weight corresponding to each temperature detection device is determined. The first weight indicates the degree of influence of the temperature detected by the temperature detection device on the temperature of multiple preset positions on the first board. Based on the multiple first temperatures, multiple first weights, and second positions, a first image of the target point of the temperature anomaly and the positional relationship between the target point and each component are determined. Based on the first image and positional relationship, the location of the temperature anomaly in the first board is determined. In this way, by dynamically associating temperature detection data with board structure information, the temperature state of areas without deployed temperature sensors can be automatically fitted, that is, the temperature distribution of the entire board can be reconstructed based on the data of limited temperature sensors. This can automatically identify abnormal heating in areas without deployed temperature sensors, effectively avoiding the detection blind spot problem caused by the limited sensor coverage of traditional methods. It achieves accurate positioning of temperature anomaly points on the board and component correlation analysis, improving the efficiency of anomaly detection and enhancing the accuracy and reliability of fault judgment.

[0088] Based on any of the above embodiments, the method for determining a first image of a target point with temperature anomalies and the positional relationship between the target point and each device according to multiple first temperatures, multiple first weights and a second position in the above method will be described in detail.

[0089] Figure 4 This is a schematic diagram illustrating a method for determining the positional relationship between a first image, target points, and various devices, as provided in an embodiment of this application. Please refer to... Figure 4 ,include:

[0090] S401. Based on multiple first temperatures and multiple first weights, determine the first image of the target point of the temperature anomaly.

[0091] In some embodiments, the electronic device can obtain temperature data for each preset location based on the first temperature of each temperature detection device and the first weight of each preset location, and determine the target point of temperature anomaly based on the temperature data of each preset location, thereby obtaining a first image including the target point of temperature anomaly.

[0092] In some embodiments, the electronic device may determine a first image of a target point with a temperature anomaly based on a plurality of first temperatures and a plurality of first weights: determining a plurality of third locations in a preset location; determining a second temperature of each third location based on the plurality of first temperatures and the plurality of first weights; and determining a first image based on the first temperature of each first location and the second temperature of each third location.

[0093] The third position is the preset position where no temperature detection device is configured. For example, the first board includes two preset positions, namely preset position A and preset position B. The position of temperature detection device A is preset position A, and preset position B is the third position, that is, the preset position where no temperature detection device is deployed.

[0094] The second temperature can be the temperature data at the third location on the first board.

[0095] In some embodiments, the electronic device may determine the second temperature of each third position based on a plurality of first temperatures and a plurality of first weights in the following manner: for any third position; in each of the first weights, determine the sub-weight corresponding to the third position; determine the weighted sum of the plurality of first temperatures and the plurality of sub-weights corresponding to the third position to obtain the second temperature of the third position.

[0096] The sub-weight can be the first weight of each temperature detection device for the third position. For example, if there are temperature detection devices A and B, to determine the sub-weight corresponding to the third position A, it is necessary to obtain the first weight of temperature detection device A for the third position A and the first weight of temperature detection device B for the third position A, i.e., the sub-weight.

[0097] In some embodiments, after determining the sub-weight corresponding to the third position, the electronic device can multiply the first temperature and the corresponding sub-weight, and add the multiple multiplication results together to obtain the second temperature at the third position.

[0098] For example, if the two sub-weights corresponding to the third position A are sub-weight A and sub-weight B, sub-weight A is 0.6 and sub-weight B is 0.5, where sub-weight A is the first weight of the temperature detection device A corresponding to the third position A and the first temperature is 30°C, and sub-weight B is the first weight of the temperature detection device B corresponding to the third position A and the first temperature is 50°C, then the second temperature of the third position can be expressed as 30°C*0.6+50°C*0.5=43°C.

[0099] S402. Determine the positional relationship between the target point and each device based on the first image and the second position.

[0100] Positional relationships can refer to the relationship between the target point and the location of various devices. For example, the target point may be associated with the CPU location, and the target point may be located next to the CPU location.

[0101] In some embodiments, the electronic device may determine the positional relationship between the target point and each device based on the first image and the second position in the following manner: based on the second position, determine the coverage area corresponding to each device in the first board, and determine the coverage area where the target point is located; determine that there is a positional association between the target point and the device corresponding to the coverage area where the target point is located.

[0102] The coverage area can be the area affected by the components in the first board.

[0103] In some embodiments, the electronic device can determine the coverage area of ​​each device in the first board based on the thermally affected area of ​​each device. Specifically, the electronic device can obtain the thermal characteristic parameters (e.g., power consumption) of each device, and based on the thermal characteristic parameters of each device, expand outward from the device location to obtain the coverage area of ​​each device in the first board. Among these, the thermally affected area of ​​high-power devices (e.g., CPU) is larger, while the thermally affected area of ​​low-power devices is smaller.

[0104] Optionally, the electronic device can learn the actual thermal impact pattern of each device through machine learning algorithms based on the spatial distribution relationship between temperature anomalies and devices in historical anomaly data, and dynamically adjust the shape and size of the coverage area. For areas where anomaly associations frequently occur, the coverage range of the corresponding devices can be appropriately expanded.

[0105] In some embodiments, the electronic device may employ bounding box detection to determine whether a device exists within the coverage area where the target point is located, based on the spatial relationship between the coverage area and various devices. If the target point is located within the coverage area of ​​a device, it is determined that there is a positional association between the target point and the device corresponding to the coverage area where the target point is located.

[0106] If there are no devices in the coverage area where the target point is located, and it is not within the coverage area of ​​any device, the electronic device can determine the distance between the target point and each device. Based on the distance threshold, if the distance between the target point and the device is less than the distance threshold, it is determined that there is a positional relationship between the target point and the device corresponding to the coverage area where the target point is located. The distance threshold can be dynamically adjusted according to the device type and thermal characteristics.

[0107] If a target point is located in the overlapping area of ​​multiple devices, the electronic device records the positional relationship between the target point and related devices. Based on the thermal characteristics and historical data of each device, a higher weight is assigned to the primary associated device, and the positional relationship of the target point is determined based on the device with the highest weight (strongest association). Through this method, the electronic device can accurately establish the positional relationship between temperature anomaly target points and board devices, providing a reliable spatial basis for subsequent fault diagnosis and maintenance decisions.

[0108] exist Figure 4In the illustrated embodiment, since a first image of the full-area temperature distribution of the board can be reconstructed based on multiple first temperatures and multiple first weights, and the positional association between the target point and the specific device can be determined based on the coverage area of ​​each device, the detection blind spot problem caused by the limited deployment of temperature sensors can be effectively avoided. This solves the technical problems of traditional methods being unable to accurately locate the correspondence between abnormal heat points and specific devices and relying on manual inspection with low efficiency. While realizing the visual reconstruction of the full-area temperature of the board, it significantly improves the accuracy of the association and location of abnormal heat points and faulty devices, thereby providing reliable technical support for the rapid fault diagnosis and accurate repair of server boards.

[0109] Based on any of the above embodiments, the method by which the electronic device determines the location of the temperature anomaly in the first board based on the first image and the positional relationship in the above processing method will be described in detail.

[0110] Figure 5 This is a schematic diagram illustrating a method for determining the location of a temperature anomaly according to an embodiment of this application. Please refer to... Figure 5 ,include:

[0111] S501. Obtain the operating status of the first board.

[0112] The operating status is used to indicate the load condition of the first board. For example, the operating status can be a high load operating status or a low load operating status.

[0113] In some embodiments, the electronic device can collect the operating data of the first board in real time, such as CPU load and memory usage, and can also obtain operating parameters such as the operating environment temperature of the first board, thereby determining the operating status of the first board.

[0114] S502. Determine the third threshold based on the operating status.

[0115] The third threshold is used to determine whether the temperature at the location of the temperature anomaly is normal.

[0116] In some embodiments, after the neural network model outputs the first image, a third threshold is determined based on the operating state (load and ambient temperature) of the first board. For example, under high load conditions, i.e., when the first board is operating under high load, the electronic device can appropriately increase the third threshold to prevent normal temperature rise from being misjudged as a fault; under low load conditions, i.e., when the first board is operating under low load, the electronic device can appropriately decrease the third threshold to improve the detection sensitivity of minor anomalies.

[0117] S503. Based on the temperature of each preset position in the first image and the third threshold, determine the target point whose temperature is greater than or equal to the third threshold.

[0118] In some embodiments, the electronic device can compare the temperature data corresponding to the target point with the abnormal temperature in the first image with a third threshold indicating the load condition. If the temperature of a target point is less than the third threshold, it indicates that the target point is in a normal heating state in the first board and will not cause the first board to malfunction, so it needs to be filtered out. If the temperature of a target point is greater than or equal to the third threshold, it indicates that the target point is in an abnormal heating state in the first board and may cause the first board to malfunction, so the target point is determined to be a point that needs to be investigated.

[0119] S504. Based on the target point and its location relationship with the temperature being greater than or equal to the third threshold, determine the location of the temperature anomaly in the first board.

[0120] In some embodiments, after identifying the target points in the first image that affect the operation of the first board, the electronic device can visually mark the locations of the target points on the first image based on the positional relationships of the target points, and display the specific troubleshooting locations to the user. For example, the location of the target point (e.g., CPU location), the name of the associated device, and the type of temperature anomaly (e.g., device overheating) can be obtained, providing a precise basis for subsequent troubleshooting.

[0121] exist Figure 5 In the illustrated embodiment, since the real-time operating status of the first board can be obtained, the anomaly judgment threshold can be dynamically adjusted based on the load. Under high load conditions, the threshold can be appropriately relaxed to avoid false alarms, while under low load conditions, the threshold can be tightened to improve detection sensitivity. Combined with the temperature distribution features in the first image and the correlation between the target point and the device location, multi-dimensional comprehensive analysis can be achieved. Therefore, it can effectively avoid the misjudgment problem caused by the inability of fixed threshold detection methods to adapt to dynamic operating conditions, and solve the technical problem of the impact on the positioning accuracy in traditional anomaly detection due to the failure to consider the actual operating status. While improving the positioning accuracy of abnormal hot spots, it significantly reduces the false alarm rate, thereby providing reliable technical support for the intelligent operation and maintenance and predictive maintenance of server boards.

[0122] Figure 6 A flowchart illustrating another anomaly detection method provided in this application embodiment is shown below. Figure 6 As shown, embodiments of this application provide an anomaly detection method, which is described in detail below:

[0123] S601. Obtain the dataset, which includes thermal imaging temperature distribution maps of normal and abnormal boards, heating status of normal and abnormal boards, and onboard temperature sensor data of normal and abnormal boards.

[0124] S602. Perform consistency processing on the thermal imaging temperature distribution map in the dataset.

[0125] S603. Build a convolutional neural network model, defining the model's layer structure, parameter settings, activation functions, etc.

[0126] In some embodiments, the core advantage of the neural network model lies in its powerful pattern learning capability. Through deep training on a large amount of temperature distribution and sensor data from normal and abnormal overheating boards, it constructs a precise mapping relationship between temperature distribution characteristics and sensor locations and data. This training data can cover the heating conditions of most boards in the server, and can include temperature samples from various typical abnormal overheating scenarios, as well as baseline data for the corresponding normal conditions of the boards.

[0127] S604. Based on the preprocessed thermal imaging temperature distribution map and the onboard temperature sensor data, a neural network model is trained and the model parameters are adjusted so that the model can fit the thermal imaging temperature distribution and temperature sensor data of a board.

[0128] S605. The model, after training, can fit the temperature distribution map of the output board and locate abnormal hot spots after inputting onboard temperature sensor data.

[0129] S606 accurately locates abnormal heat points on the board, assisting users in quickly troubleshooting board faults.

[0130] In this embodiment, a mapping relationship is constructed between the thermal imaging temperature distribution map of the circuit board and the location and temperature data of the circuit board temperature sensor. This enables the detection and location of abnormal heating points on the circuit board by simply inputting the location and temperature data of the circuit board temperature sensor. In this process, the use of a convolutional neural network model to identify the thermal imaging map of abnormal heating on the circuit board achieves high-precision fault detection and location. The model's powerful feature extraction capability can accurately capture subtle features related to abnormal heating in the thermal imaging map. Compared with traditional methods based on experience or manual observation, this significantly reduces the chances of misjudgment and missed judgment. By inputting the circuit board temperature sensor data into the trained model, the model can complete analysis and identification in a short time, quickly locating circuit boards that may have abnormal heating. This greatly improves the speed compared to manual inspection, increasing the efficiency of circuit board anomaly detection. Furthermore, it significantly reduces reliance on manual inspection. In traditional hardware maintenance, professionals need to spend a lot of time inspecting each circuit board individually, which is not only time-consuming and labor-intensive but also requires specialized knowledge and experience. Once the neural network model is trained, it can automatically perform detection, requiring only a small number of personnel for model maintenance and preliminary verification of results, thus freeing up a significant amount of manpower.

[0131] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method.

[0132] Figure 7 This is a schematic diagram of an anomaly detection device provided in an embodiment of this application. Figure 7 As shown, embodiments of this application also provide an anomaly detection device 70, including: an acquisition module 71, a first determination module 72, a second determination module 73, and a processing module 74, wherein,

[0133] The acquisition module 71 is used to acquire the first temperature detected by multiple temperature detection devices on the first board with temperature anomalies, the first position of each temperature detection device on the first board, and the second position of each device in the first board.

[0134] The first determining module 72 is used to determine the first weight corresponding to each temperature detection device based on multiple first temperatures and multiple first positions. The first weight is used to indicate the degree of influence of the temperature detected by the temperature detection device on the temperature of multiple preset positions on the first board.

[0135] The second determining module 73 is used to determine, based on multiple first temperatures, multiple first weights and second positions, a first image of the target point of the temperature anomaly and the positional relationship between the target point and each device.

[0136] Processing module 74 is used to determine the location of the temperature anomaly in the first board based on the first image and positional relationship.

[0137] For a description of the features in the embodiment corresponding to the anomaly detection device, please refer to the relevant description in the embodiment corresponding to the anomaly detection method, which will not be repeated here.

[0138] In one possible implementation, the first determining module 72 is specifically used for:

[0139] Determine the first position of the temperature detection device and the first distance between it and each preset position;

[0140] Based on the first temperature detected by multiple first distance and temperature detection devices, the first weight corresponding to the temperature detection device is determined.

[0141] In one possible implementation, the first determining module 72 is specifically used for:

[0142] The first sub-weight corresponding to the temperature detection device is determined based on the first distance corresponding to the preset position;

[0143] Based on the first temperature, a second sub-weight is determined. The second sub-weight is used to indicate the correlation between the first temperature and the temperature at a preset location.

[0144] Based on the first sub-weight and the second sub-weight, determine the first weight corresponding to the temperature detection device at the preset position.

[0145] In one possible implementation, the first determining module 72 is specifically used for:

[0146] Determine the weighting coefficients corresponding to the first temperature and the first distance. The weighting coefficients for the first temperature and the first distance are preset values.

[0147] The first weight is determined based on the weighting coefficient of the first distance, the first sub-weight, the weighting coefficient of the first temperature, and the second sub-weight.

[0148] In one possible implementation, the second determining module 73 is specifically used for:

[0149] The first image of the target point of the temperature anomaly is determined based on multiple first temperatures and multiple first weights;

[0150] Based on the first image and the second location, determine the positional relationship between the target point and each device.

[0151] In one possible implementation, the second determining module 73 is specifically used for:

[0152] In the preset positions, multiple third positions are determined, and the third positions are preset positions where no temperature detection device is configured;

[0153] The second temperature at each third position is determined based on multiple first temperatures and multiple first weights;

[0154] The first image is determined based on the first temperature at each first position and the second temperature at each third position.

[0155] In one possible implementation, the second determining module 73 is specifically used for:

[0156] In each of the first weights, determine the sub-weight corresponding to the third position;

[0157] The weighted sum of multiple sub-weights corresponding to multiple first temperatures and third positions is determined to obtain the second temperature at the third position.

[0158] In one possible implementation, the second determining module 73 is specifically used for:

[0159] Based on the second position, determine the coverage area corresponding to each device in the first board, and determine the coverage area where the target point is located;

[0160] It is determined that there is a locational correlation between the target point and the device corresponding to the coverage area where the target point is located.

[0161] In one possible implementation, the processing module 74 is specifically used for:

[0162] Obtain the operating status of the first board, which indicates the load status of the first board;

[0163] Based on the operating status, a third threshold is determined. The third threshold is used to determine whether the temperature at the location of the temperature anomaly is normal.

[0164] Based on the temperature at each preset location in the first image and the third threshold, determine the target point whose temperature is greater than or equal to the third threshold;

[0165] Based on the target points with temperatures greater than or equal to the third threshold and their location relationships, the location of temperature anomalies in the first board is determined.

[0166] Figure 8 A schematic diagram of the structure of the electronic device provided in this application. Figure 8 As shown, the electronic device 80 provided in this embodiment includes at least one processor 81 and a memory 82. Optionally, the electronic device 80 further includes a communication component 83. The processor 81, memory 82, and communication component 83 are connected via a bus.

[0167] In the specific implementation process, at least one processor 81 executes computer execution instructions stored in memory 82, causing at least one processor 81 to execute the above-described anomaly detection method embodiment.

[0168] The specific implementation process of processor 81 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0169] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0170] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0171] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0172] Embodiments of this application also provide a computer-readable storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above-described anomaly detection method embodiments at runtime.

[0173] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0174] Embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the steps in any of the above-described anomaly detection method embodiments.

[0175] Embodiments of this application also provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in any of the above-described anomaly detection method embodiments.

[0176] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for specific applications, but such implementations should not be considered beyond the scope of this application.

[0177] The above provides a detailed description of an anomaly detection method provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this application.

Claims

1. An anomaly detection method characterized by, The method comprises: obtaining a first temperature detected by a plurality of temperature detection devices on a first board card, a first position of each temperature detection device on the first board card, and a second position of each device in the first board card; for any one temperature detection device, determining a first distance between the first position of the temperature detection device and each preset position, the first distance indicating a spatial position relationship between the temperature detection device and each preset position; for any one preset position, determining a first sub-weight corresponding to the temperature detection device according to the first distance corresponding to the preset position; determining a second sub-weight according to the first temperature, the second sub-weight indicating an association relationship between the first temperature and a temperature of the preset position; determining a first weight corresponding to the temperature detection device at the preset position according to the first sub-weight and the second sub-weight, the first weight indicating an influence degree of a temperature detected by the temperature detection device on a temperature of a plurality of preset positions on the first board card; determining a first image of a target point of a temperature anomaly according to the plurality of first temperatures and the plurality of first weights; determining a position relationship between the target point and the devices according to the first image and the second position; determining a position of the temperature anomaly in the first board card according to the first image and the position relationship.

2. The method of claim 1, wherein, The method further comprises: determining a weighting coefficient corresponding to the first temperature and the first distance, the weighting coefficient of the first temperature and the weighting coefficient of the first distance being preset values; determining the first weight according to the weighting coefficient of the first distance, the first sub-weight, the weighting coefficient of the first temperature, and the second sub-weight.

3. The method of claim 1, wherein, The method further comprises: determining a plurality of third positions in the preset positions, the third positions being preset positions without temperature detection devices; determining a second temperature of each third position according to the plurality of first temperatures and the plurality of first weights; determining the first image according to the first temperature of each first position and the second temperature of each third position.

4. The method of claim 3, wherein, The method further comprises: determining a sub-weight corresponding to the third position in the first weights; determining a weighted sum of the plurality of first temperatures and the plurality of sub-weights corresponding to the third position to obtain the second temperature of the third position.

5. The method of claim 1, wherein, The method further comprises: determining a coverage area corresponding to the devices in the first board card according to the second position, and determining a coverage area in which the target point is located; determining that the target point and the device corresponding to the coverage area in which the target point is located have a position correlation relationship.

6. The method of claim 1, wherein, The method further comprises: acquire a running state of the first board card, the running state being used to indicate a load condition of the first board card; determine a third threshold according to the running state, the third threshold being used to determine whether a temperature of a position of the temperature anomaly is normal; determine a target point whose temperature is greater than or equal to the third threshold according to the temperature of each preset position in the first image and the third threshold; determine the position of the temperature anomaly in the first board card according to the target point whose temperature is greater than or equal to the third threshold and the position relationship.

7. An electronic device, comprising: comprise: a memory, used to store a computer program; a processor, used to execute the computer program to realize steps of the anomaly detection method according to any one of claims 1 to 6.

8. A computer program product, characterised in that, comprise a computer program, the computer program being executed by a processor to realize the anomaly detection method according to any one of claims 1 to 6.

Images