A temperature anomaly alarm method and device, a computing server and a storage medium

Abstract

The application provides a temperature anomaly alarm method and device, a computing server and a storage medium, the method comprising: inputting a high-dimensional infrared image into a constructed image processing module, and outputting a time sequence calculation result feature map; performing differential calculation on the time sequence calculation result feature map to generate a change feature map; judging whether there is an abnormal change in the change feature map; and if it is judged that there is an abnormal change in the change feature map, generating alarm information according to the abnormal change. In the technical scheme provided by the embodiment of the application, the infrared imaging device and the image processing module are used to perform real-time monitoring and fault diagnosis on the gas boiler equipment for 24 hours, so that the labor cost is reduced, the abnormal hardware part can be quickly and accurately determined, the accuracy and efficiency of fault diagnosis are improved, and thus the safety and efficiency of the operation of the gas boiler equipment are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of gas equipment technology, specifically to a method, device, computing server, and storage medium for alarming abnormal temperatures. Background Technology

[0002] In centralized heating systems, gas-fired boilers are an important heat source. The safe operation of these boilers is crucial for ensuring normal heating for users. Failures in gas-fired boilers can severely impact heating supply, leading to adverse social consequences and economic losses.

[0003] When a gas-fired boiler malfunctions, the characteristics of its heat-generating medium cause localized temperature radiation changes, which in turn affect the ambient temperature over time. Leveraging this phenomenon, thermal infrared imaging technology is increasingly being applied to the monitoring and fault detection of gas-fired boilers.

[0004] Chinese utility model patent with publication number CN213397432U discloses a "non-contact boiler infrared furnace temperature online measuring device", which uses infrared thermal imaging to monitor the flue gas temperature at the furnace outlet of a thermal power plant boiler. However, specific fault diagnosis requires manual judgment of the monitoring video, resulting in low efficiency and accuracy of fault diagnosis and high labor consumption.

[0005] Chinese utility model patent with publication number CN217786354U discloses "an infrared temperature field measurement system for power plant boilers". It uses an infrared non-contact temperature measurement device to monitor the operation of power plant boilers. It monitors the temperature field of the furnace through the infrared non-contact temperature measurement device, thereby achieving the purpose of alarming when the threshold is exceeded. However, its monitoring range is limited and incomplete, and it mainly relies on manual judgment. It requires a lot of manpower to achieve 24-hour real-time monitoring. Moreover, the judgment may have problems such as misjudgment and untimely judgment, resulting in low efficiency and accuracy of fault diagnosis. Summary of the Invention

[0006] In view of the above problems, embodiments of the present invention provide a temperature anomaly alarm method, device, computing server and storage medium to solve the problem that the fault diagnosis of gas boiler equipment still relies on manual methods in the prior art, resulting in low efficiency and accuracy of fault diagnosis.

[0007] In a first aspect, embodiments of the present invention provide a method for alarming abnormal temperatures, the method comprising:

[0008] The high-dimensional infrared image is input into the constructed image processing module, and the time-series calculation result feature map is output.

[0009] Perform differential calculation on the feature map of the time series calculation result to generate a change feature map;

[0010] Determine whether there are any abnormal changes in the aforementioned feature graph;

[0011] If an abnormal change is detected in the change feature map, the abnormal temperature location is determined based on the abnormal change, and an alarm message is generated.

[0012] In one possible implementation, the image processing module includes a MobileNetV2 model and an image processing neural network model. The image processing module, which takes a high-dimensional infrared image as input and outputs a time-series computational result feature map, includes:

[0013] The MobileNetV2 model is used to extract features from high-dimensional infrared images to obtain multiple basic feature maps.

[0014] Feature fusion is performed on the multiple basic feature maps to obtain a temporal high-dimensional feature map;

[0015] The temporal high-dimensional feature map is input into the image processing neural network model to extract temporal features and obtain the temporal calculation result map.

[0016] In one possible implementation, before the image processing module that inputs the high-dimensional infrared image and outputs the time-series computation result feature map, it further includes:

[0017] Acquire infrared image data packets sent by an infrared imaging device, wherein the infrared image data packets contain multiple sets of infrared images;

[0018] The multiple sets of infrared images are sorted in chronological order;

[0019] The sorted infrared images are linked in dimensions to generate a high-dimensional infrared image.

[0020] In one possible implementation, the image processing neural network model incorporates a 3-dimensional causal convolution, the size of which is the same as the size of the temporal high-dimensional feature map.

[0021] In one possible implementation, the three dimensions of causal convolution are the first dimension, the second dimension, and the third dimension, respectively.

[0022] The step of inputting the temporal high-dimensional feature map into the image processing neural network model to extract temporal features and obtain the temporal calculation result map includes:

[0023] Feature extraction is performed on the preceding data and the data before it using the first dimension; feature extraction is performed on the current data and the data before it using the second dimension; and feature extraction is performed on the result data and the data before it using the third dimension.

[0024] There is a temporal causal relationship between the antecedent data, the current data, and the result data.

[0025] In one possible implementation, the abnormal change is used to indicate that the temperature change is greater than a preset threshold.

[0026] In one possible implementation, the method further includes:

[0027] If it is determined that there are no abnormal changes in the change feature map, then the step of inputting the high-dimensional infrared image into the image processing module and outputting the time-series calculation result feature map continues.

[0028] Secondly, embodiments of the present invention provide a temperature anomaly alarm device, the temperature anomaly alarm device comprising:

[0029] The processing module is used to input high-dimensional infrared images into the constructed image processing module and output time-series calculation result feature maps;

[0030] The calculation module is used to perform differential calculation on the feature map of the time series calculation result to generate a change feature map;

[0031] The judgment module is used to determine whether there are abnormal changes in the change feature map;

[0032] If the judgment module determines that there is an abnormal change in the change feature map, the alarm module is triggered to determine the abnormal temperature location based on the abnormal change and generate alarm information.

[0033] Thirdly, embodiments of the present invention provide a computing server, the computing server including one or more processors, a memory, and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the computing server, cause the computing server to perform the temperature anomaly alarm method described in the first aspect or any possible implementation of the first aspect.

[0034] Fourthly, embodiments of the present invention provide a computer-readable storage medium comprising a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the temperature anomaly alarm method as described in the first aspect or any possible implementation thereof.

[0035] In the technical solution provided by the embodiments of the present invention, the gas boiler equipment is monitored and diagnosed in real time 24 hours a day by using infrared imaging equipment and image processing module, which reduces labor costs and can quickly and accurately identify abnormal hardware parts, thereby improving the accuracy and efficiency of fault diagnosis and effectively improving the safety and efficiency of gas boiler equipment operation. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating a temperature anomaly alarm method provided in an embodiment of the present invention.

[0037] Figure 2 This is a schematic diagram illustrating the principle of an image processing neural network model provided in an embodiment of the present invention.

[0038] Figure 3 This is a schematic diagram of a temperature anomaly alarm device provided in an embodiment of the present invention.

[0039] Figure 4 This is a schematic diagram of a computing server provided in an embodiment of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer and more understandable, the invention will be further described below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0041] Figure 1 This is a flowchart illustrating a temperature anomaly alarm method provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes:

[0042] Step 101: Input the high-dimensional infrared image into the constructed image processing module and output the time-series calculation result feature map.

[0043] In this embodiment of the invention, infrared images are acquired using infrared imaging devices to monitor the status of a gas-fired boiler in real time. The infrared imaging devices are installed inside the furnace of the gas-fired boiler. To acquire infrared images more comprehensively, multiple infrared imaging devices are installed inside the furnace. Different infrared imaging devices have different fields of view; by superimposing the fields of view of multiple infrared imaging devices, the entire inner wall surface of the furnace can be completely covered, avoiding blind spots caused by pipe obstructions, which would affect the integrity and comprehensiveness of the infrared image data.

[0044] In this embodiment of the invention, an example is described using four infrared imaging devices installed in the furnace. For instance, the four infrared imaging devices are respectively installed above, below, to the left, and to the right of the furnace. Another example is that the four infrared imaging devices are respectively installed to the upper left, lower left, upper right, and lower right of the furnace. In practical applications, the installation positions of the infrared imaging devices can be adjusted based on the specific conditions inside the furnace; this embodiment of the invention does not limit this.

[0045] Specifically, each infrared imaging device acquires one frame of infrared image per second. Multiple infrared images acquired by multiple infrared imaging devices at the same time are considered as a set of infrared image data. The acquired sets of infrared images are sorted chronologically, and the sorted infrared images are packaged at preset time intervals to generate infrared image data packets, which are then sent to the server. For example, if the preset time interval is 1 minute, then one set of infrared image data is generated per second, and the infrared image data packet contains 60 sets of infrared image data.

[0046] In this embodiment of the invention, the infrared imaging device is either a black-and-white imaging device or a color imaging device. In the black-and-white image generated by the black-and-white imaging device, different gray levels reflect temperature, enabling rapid response in extreme environments. In the color image generated by the color imaging device, the mapping relationship between color and temperature value reflects temperature. Color images enhance visual contrast through pseudo-color, improving the ability to distinguish minute temperature differences and enabling precise temperature monitoring, making them suitable for scenarios with high temperature accuracy requirements. In practical applications, the imaging type of the infrared imaging device can be selected based on the actual situation; this embodiment of the invention does not limit this selection.

[0047] In this embodiment of the invention, an infrared image data packet sent by an infrared imaging device is acquired. The infrared image data packet includes multiple sets of infrared images. An image processing module is deployed on the computing server. Before the infrared image data packet is input into the image processing module, it is preprocessed. The preprocessing process is as follows: the multiple sets of infrared images are sorted in chronological order; the sorted infrared images are linked in dimensions to generate a high-dimensional infrared image. That is, the high-dimensional infrared image is obtained based on multiple single-dimensional infrared images arranged in chronological order; therefore, the high-dimensional infrared image has temporal characteristics in terms of dimensions.

[0048] In this embodiment of the invention, the image processing module includes a MobileNetV2 model and an image processing neural network model. The MobileNetV2 model extracts features from the high-dimensional infrared image to obtain multiple basic feature maps; these basic feature maps are then fused to obtain a temporal high-dimensional feature map; this temporal high-dimensional feature map is then input into the image processing neural network model for temporal feature extraction to obtain a temporal calculation result feature map.

[0049] The MobileNetV2 model introduces an inverse residual structure and a linear bottleneck, which alleviates the vanishing gradient problem while preserving feature integrity and optimizing the efficiency of feature extraction and propagation. This further improves the model's performance and computational efficiency while maintaining its lightweight nature.

[0050] Figure 2 This is a schematic diagram illustrating the principle of an image processing neural network model provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the image processing neural network model incorporates a 3-dimensional causal convolution. The size of the causal convolution is the same as the size of the temporal high-dimensional feature map, i.e., the length and width are the same. The three dimensions of the causal convolution are the first dimension, the second dimension, and the third dimension, respectively. Figure 2 As shown, from left to right, these are the first, second, and third dimensions. During feature extraction, causal convolution only depends on the current data and the data preceding it, not on the data following it. Features are extracted from the preceding data and the data preceding it using the first dimension, from the current data and the data preceding it using the second dimension, and from the result data and the data preceding it using the third dimension. There is a temporal causal relationship between the preceding data, the current data, and the result data.

[0051] For example, a temporal high-dimensional feature map is represented as [data 1, data 2, data 3, data 4, data 5, data 6]. When the current data is data 4, the preceding and resulting data are data 3 and data 5, respectively. In this case, features are extracted from data 1, data 2, and data 3 using the first dimension, from data 1, data 2, data 3, and data 4 using the second dimension, and from data 1, data 2, data 3, data 4, and data 5 using the third dimension. When the resulting data is data 5, data 6 follows data 5. Therefore, relative to data 5, data 6 still belongs to future data and does not participate in the causal convolution feature extraction process. The three dimensions of causal convolution extract features from different data, thus the feature extraction results from the three dimensions can reflect different feature information.

[0052] In this embodiment of the invention, the image processing neural network model shifts rightward step by step according to the temporal order in the temporal high-dimensional feature map, and performs feature extraction through causal convolution to obtain the temporal calculation result feature map. Because the temporal high-dimensional feature map is generated based on continuous infrared images with temporal features, there are temporal causal relationships between the dimensions of the temporal high-dimensional feature map. The role of temporal convolution is to extract the change process features generated when the temporal features change, and finally output the temporal calculation result feature map.

[0053] In infrared images, different gray levels or colors reflect temperature values; that is, there is a mapping relationship between the gray levels or colors of an infrared image and temperature values. After performing a series of feature extractions on a high-dimensional infrared image, the resulting time-series computational feature map also shows a certain mapping relationship between gray levels or colors and temperature values.

[0054] Step 102: Perform differential calculation on the feature map of the time series calculation result to generate a change feature map.

[0055] In this step, performing differential calculation on the feature map of the time series calculation result means subtracting the pixel values ​​at corresponding positions in the feature map of the time series calculation result to obtain the pixel difference, so as to weaken the similar parts and highlight the changed parts, which facilitates the identification and judgment of subsequent abnormal changes.

[0056] Step 103: Determine if there are any abnormal changes in the change feature graph; if abnormal changes are found in the change feature graph, proceed to step 104; if no abnormal changes are found in the change feature graph, continue to step 101.

[0057] In this step, abnormal changes are used to indicate that the temperature change exceeds a preset threshold. If abnormal changes are present in the change characteristic graph, it indicates that the temperature change trend of the gas boiler equipment is unstable over time, which may indicate a malfunction. In this case, the abnormal temperature location is identified based on the abnormal changes, and an alarm message is generated. If no abnormal changes are present in the change characteristic graph, it indicates that the temperature change trend of the gas boiler equipment is relatively stable over time, and the temperatures of various hardware components are relatively stable with no abnormalities. In this case, no alarm message needs to be generated.

[0058] Step 104: Identify the location of the temperature anomaly based on the abnormal changes and generate an alarm message.

[0059] In this embodiment of the invention, the structure of the gas-fired boiler equipment is relatively complex, and the hardware component causing abnormal temperature changes may be external or internal. When the component causing the abnormal temperature change is internal hardware and cannot be directly observed, the component with the abnormal temperature can be inferred through the change feature map; when the component with the abnormal temperature is external hardware and can be directly observed, the component with the abnormal temperature can be quickly determined through the change feature map.

[0060] In this embodiment of the invention, feature extraction is performed on the temporal high-dimensional feature map using three-dimensional causal convolution. On one hand, change process features in the time dimension are extracted; on the other hand, logical change features with temporal continuity and causal relationships are extracted. Extracting change process features in the time dimension is to determine whether abnormal changes exist during temperature changes. During the operation of a gas-fired boiler, the temperature of the equipment itself is not constant but rather undergoes steady-state changes (i.e., the change trend is relatively stable). The change process features extracted in this embodiment not only reflect temperature changes but also abnormal temperature changes caused by abnormal faults in the gas-fired boiler within the steady-state changes, providing important reference value in identifying the location of temperature anomalies. Extracting logical change features involves combining past, present, and future temperature changes to extract the logical relationships causing abnormal temperature changes, and deducing the internal hardware causing the abnormal temperature changes from these logical relationships. In the technical solution provided by this embodiment, regardless of whether the temperature anomaly is internal or external hardware, the location of the temperature anomaly can be quickly and accurately identified, allowing for timely maintenance of the faulty part and improving fault diagnosis and maintenance efficiency.

[0061] In the technical solution provided by the embodiments of the present invention, the gas boiler equipment is monitored and diagnosed in real time 24 hours a day by using infrared imaging equipment and image processing module, which reduces labor costs and can quickly and accurately identify abnormal hardware parts, thereby improving the accuracy and efficiency of fault diagnosis and effectively improving the safety and efficiency of gas boiler equipment operation.

[0062] Figure 3 This is a schematic diagram of a temperature anomaly alarm device provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the temperature anomaly alarm device includes: a processing module 11, a calculation module 12, a judgment module 13, and an alarm module 14. The processing module 11 is used to input a high-dimensional infrared image into the constructed image processing module and output a time-series calculation result feature map. The calculation module 12 is used to perform differential calculations on the time-series calculation result feature map to generate a change feature map. The judgment module 13 is used to determine whether there are abnormal changes in the change feature map. If the judgment module 13 determines that there are abnormal changes in the change feature map, it triggers the alarm module 14 to determine the temperature anomaly location based on the abnormal changes and generate alarm information.

[0063] In this embodiment of the invention, the image processing module includes a MobileNetV2 model and an image processing neural network model. The processing module 11 includes a first extraction submodule, a fusion submodule, and a second extraction submodule. The first extraction submodule is used to extract features from a high-dimensional infrared image using the MobileNetV2 model to obtain multiple basic feature maps; the fusion submodule is used to fuse the features of the multiple basic feature maps to obtain a temporal high-dimensional feature map; the second extraction submodule is used to input the temporal high-dimensional feature map into the image processing neural network model for temporal feature extraction to obtain a temporal calculation result map.

[0064] In this embodiment of the invention, the temperature anomaly alarm device further includes an acquisition module 15, a sorting module 16, and a generation module 17. The acquisition module 15 is used to acquire infrared image data packets sent by the infrared imaging device, wherein the infrared image data packets contain multiple sets of infrared images; the sorting module 16 is used to sort the multiple sets of infrared images in chronological order; and the generation module 17 is used to link the sorted infrared images in dimensions to generate a high-dimensional infrared image.

[0065] In this embodiment of the invention, a 3-dimensional causal convolution is constructed in the image processing neural network model, and the size of the causal convolution is the same as the size of the temporal high-dimensional feature map.

[0066] In this embodiment of the invention, the three dimensions of causal convolution are the first dimension, the second dimension, and the third dimension. The second extraction submodule is specifically used to extract features from the preceding data and the data before it using the first dimension, to extract features from the current data and the data before it using the second dimension, and to extract features from the result data and the data before it using the third dimension; there is a temporal causal relationship between the preceding data, the current data, and the result data.

[0067] In this embodiment of the invention, abnormal changes are used to indicate that the temperature change is greater than a preset threshold.

[0068] In this embodiment of the invention, if the judgment module 13 determines that there is no abnormal change in the change feature map, the processing module 11 is triggered to continue executing the step of inputting the high-dimensional infrared image into the image processing module and outputting the time-series calculation result feature map.

[0069] In the technical solution provided by the embodiments of the present invention, the gas boiler equipment is monitored and diagnosed in real time 24 hours a day by using infrared imaging equipment and image processing module, which reduces labor costs and can quickly and accurately identify abnormal hardware parts, thereby improving the accuracy and efficiency of fault diagnosis and effectively improving the safety and efficiency of gas boiler equipment operation.

[0070] This invention provides a computer-readable storage medium that includes a stored program. When the program runs, it controls the device containing the computer-readable storage medium to execute the steps of the above-described temperature anomaly alarm method. For a detailed description, please refer to the embodiments of the above-described temperature anomaly alarm method.

[0071] Figure 4 A schematic diagram of a computing server provided in an embodiment of the present invention, as shown below. Figure 4 As shown, the computing server includes a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program 303, it implements the steps in the various method embodiments described above. Alternatively, when the processor 301 executes the computer program 303, it implements the functions of each module / unit in the various device embodiments described above.

[0072] The computing server 3 may include, but is not limited to, a processor 301 and a memory 302. Those skilled in the art will understand that... Figure 4 This is merely an example of computing server 3 and does not constitute a limitation on computing server 3. It may include more or fewer components than shown, or different components.

[0073] The processor 301 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. In this embodiment of the invention, the processor can be the controller described in the above embodiments.

[0074] The memory 302 can be an internal storage unit of the computing server 3, such as a hard disk or memory of the computing server 3. The memory 302 can also be an external storage device of the computing server 3, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the computing server 3. The memory 302 can also include both internal and external storage units of the computing server 3. The memory 302 is used to store computer programs and other programs and data required by the computing server.

[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0076] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A temperature abnormality alarm method characterized by comprising: The method includes: The high-dimensional infrared image is input into the constructed image processing module, and the time-series calculation result feature map is output. Perform differential calculation on the feature map of the time series calculation result to generate a change feature map; Determine whether there are any abnormal changes in the aforementioned feature graph; If an abnormal change is detected in the change feature map, the abnormal temperature location is determined based on the abnormal change, and an alarm message is generated. The image processing module includes a MobileNetV2 model and an image processing neural network model. The image processing module, which takes a high-dimensional infrared image as input, outputs a time-series computation result feature map, including: The MobileNetV2 model is used to extract features from high-dimensional infrared images to obtain multiple basic feature maps. Feature fusion is performed on the multiple basic feature maps to obtain a temporal high-dimensional feature map; The temporal high-dimensional feature map is input into the image processing neural network model to extract temporal features and obtain the temporal calculation result map. Before the image processing module, which inputs the high-dimensional infrared image, outputs the time-series calculation result feature map, it also includes: Acquire infrared image data packets sent by an infrared imaging device, wherein the infrared image data packets contain multiple sets of infrared images; The multiple sets of infrared images are sorted in chronological order; The sorted infrared images are linked in dimensions to generate a high-dimensional infrared image.

2. The method of claim 1, wherein, The image processing neural network model incorporates a 3-dimensional causal convolution, the size of which is the same as the size of the temporal high-dimensional feature map.

3. The method of claim 2, wherein, The three dimensions of causal convolution are the first dimension, the second dimension, and the third dimension; The step of inputting the temporal high-dimensional feature map into the image processing neural network model to extract temporal features and obtain the temporal calculation result map includes: Feature extraction is performed on the preceding data and the data before it using the first dimension; feature extraction is performed on the current data and the data before it using the second dimension; and feature extraction is performed on the result data and the data before it using the third dimension. There is a temporal causal relationship between the antecedent data, the current data, and the result data.

4. The method of claim 1, wherein, The abnormal change is used to indicate that the temperature change is greater than a preset threshold.

5. The method of claim 1, wherein, The method further includes: If it is determined that there are no abnormal changes in the change feature map, then the step of inputting the high-dimensional infrared image into the image processing module and outputting the time-series calculation result feature map continues.

6. A temperature abnormality alarm device characterized by comprising: The abnormal temperature alarm device includes: The processing module is used to input high-dimensional infrared images into the constructed image processing module and output time-series calculation result feature maps; The calculation module is used to perform differential calculation on the feature map of the time series calculation result to generate a change feature map; The judgment module is used to determine whether there are abnormal changes in the change feature map; If the judgment module determines that there is an abnormal change in the change feature map, the alarm module is triggered to determine the abnormal temperature location based on the abnormal change and generate alarm information. The image processing module includes a MobileNetV2 model and an image processing neural network model. The processing module includes a first extraction submodule, a fusion submodule, and a second extraction submodule. The first extraction submodule is used to extract features from high-dimensional infrared images using the MobileNetV2 model to obtain multiple basic feature maps; The fusion submodule is used to fuse multiple basic feature maps to obtain a temporal high-dimensional feature map; The second extraction submodule is used to input the temporal high-dimensional feature map into the image processing neural network model, perform temporal feature extraction, and obtain the temporal calculation result map; The temperature anomaly alarm device also includes an acquisition module, a sorting module, and a generation module; The acquisition module is used to acquire infrared image data packets sent by the infrared imaging device. The infrared image data packets contain multiple sets of infrared images. The sorting module is used to sort multiple sets of infrared images in chronological order. The generation module is used to link the sorted infrared images in different dimensions to generate high-dimensional infrared images.

7. A computing server, comprising: The computing server includes one or more processors, a memory, and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the computing server, cause the computing server to perform the temperature anomaly alarm method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the temperature anomaly alarm method as described in any one of claims 1-5.

Images