Infrared-based fault diagnosis method for operational amplifier chip of circuit board of airborne electronic system

Abstract

This invention relates to the field of circuit board fault diagnosis technology, specifically to an infrared-based method for diagnosing operational amplifier chips on airborne electronic system circuit boards, comprising the following steps: acquiring infrared sequence images I of airborne electronic system circuit boards under different operating modes. I and the corresponding sequence temperature matrix T M ; the infrared sequence image I from step S1 I After preprocessing, the data is fed into a semantic segmentation model to obtain infrared segmentation images of key electronic components—operational amplifier chips—in the region of interest. II This invention effectively solves the problem of inaccurate fault detection caused by factors such as changes in the external environment, insignificant temperature differences between fault categories, and component aging in existing technologies by considering the temperature change trend of the temperature sequence and fault category information, and by using infrared images and deep learning algorithms for fault classification. This improves the accuracy and generalization of fault detection.

Description

Technical Field

[0001] This invention relates to the field of circuit board fault diagnosis technology, specifically to a method for diagnosing operational amplifier chips on airborne electronic system circuit boards based on infrared technology. Background Technology

[0002] With the rapid development of avionics technology, the miniaturization and integration of airborne electronic system circuit boards are constantly improving. This has brought challenges to data extraction of components on the circuit boards and circuit fault diagnosis. Traditional contact-based circuit board fault diagnosis methods can no longer meet the increasingly complex needs of airborne electronic system circuit boards for inspection and maintenance. Infrared thermal imaging inspection technology has been widely used due to its advantages such as non-contact, speed, and intuitive detection results.

[0003] For example, CN116429783A discloses a method for diagnosing circuit board faults based on visible light images and infrared thermal image sequences. The method involves acquiring infrared and visible light images of a normal circuit board and a circuit board under test, respectively. After heterogeneous registration of the infrared and visible light images of the two types of circuit boards, the thermal feature vectors of each device on the two types of circuit boards are calculated. Then, the thermal similarity between the thermal feature vectors of each device on the normal circuit board and the thermal feature vectors of the circuit board under test is calculated, and it is determined whether the thermal similarity exceeds the performance failure threshold of the device, thereby determining the fault state of the circuit board under test.

[0004] However, in actual use, the temperature of the circuit board under normal conditions is not constant and may even vary greatly under different external environmental conditions. For some fault types, the temperature difference is only different from the normal condition for a certain period of time. After reaching a steady state, the difference in heating is not significant. In addition, the same type of circuit board may also have large temperature differences under normal working conditions due to the different aging of components. This makes the threshold setting method inaccurate and susceptible to environmental influences, resulting in low fault diagnosis efficiency and poor accuracy, which limits its effectiveness in practical applications. Summary of the Invention

[0005] The purpose of this invention is to provide an infrared-based fault diagnosis method for operational amplifier chips on airborne electronic system circuit boards, in order to solve the problems that the above-mentioned threshold setting method is not accurate enough, is easily affected by the environment, resulting in low fault diagnosis efficiency and poor accuracy, which limits its effectiveness in practical applications.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for fault diagnosis of operational amplifier chips on airborne electronic system circuit boards based on infrared radiation, comprising the following steps:

[0007] S1. Data Acquisition: Acquire infrared sequence images of the airborne electronic system circuit boards under different operating modes. I and the corresponding sequence temperature matrix TM ;

[0008] S2, Data Preprocessing: The infrared sequence image I from step S1... I After preprocessing, the data is fed into a semantic segmentation model to obtain infrared segmentation images of key electronic components—operational amplifier chips—in the region of interest. II ;

[0009] S3. Obtaining the temperature sequence matrix: Combining the temperature sequence matrix T from step S1... M Compared with the infrared segmentation image I of the operational amplifier chip in step S2 II The temperature sequence matrix T of the operational amplifier chip is obtained. I ;

[0010] S4. Temperature State Region Division and Dataset Extraction: Based on the operating time, the temperature state of electronic components—operational amplifier chips—in the region of interest of the airborne electronic system circuit board is divided into a dynamic heating region and a steady-state isothermal region, and the temperature sequence matrix T is extracted. I The data from the portion where the temperature continues to rise is denoted as the dynamic warming zone dataset, T. ID , will dataset T ID Further processing yields multidimensional temperature sequence data T; infrared segmentation images I are extracted from operational amplifier chips. II Images of the stable temperature region are used as the image dataset for the steady-state isothermal region, denoted as I. IS ;

[0011] S5. Dataset Processing and Model Training: Process the multidimensional temperature sequence data from step S4. Image dataset I of steady-state isothermal region IS Take 70% of each and set it as the training set, named T. train and I train The remaining 30% of the data is set as a validation set and named T. val and I val and multidimensional temperature series data The training set is input into a long short-term memory time series feature extraction model that incorporates an attention mechanism to extract classification features, while simultaneously inputting the image dataset of the steady-state isothermal region. IS The training set is input into the infrared thermal image classification feature extraction model to extract classification features;

[0012] S6. Feature Fusion and Model Classification Training: Fusion of the multidimensional temperature sequence data T obtained in S5 and the image dataset I of the steady-state isothermal region. IS The two types of feature information are input into the multilayer perceptron in the model to complete the model's classification training;

[0013] S7. Fault Classification and Diagnosis: Input the known category dataset into the trained model to obtain the classification of component faults in the region of interest of the airborne electronic system circuit board. Based on the classification results, diagnose the faults of operational amplifier chips on the airborne electronic system circuit board.

[0014] Preferably, in step S1, the infrared sequence image I... I and the corresponding sequence temperature matrix T M The data acquisition was achieved using multiple infrared thermal imagers positioned perpendicularly above the circuit board of the airborne electronic system under test. The circuit board of the airborne electronic system under test was connected to the excitation source via a programmable relay switch, and the host computer was connected to both the infrared camera and the programmable relay switch.

[0015] Preferably, in step S2, the infrared sequence image I... I Preprocessing includes, but is not limited to, translation, rotation and mirroring. In step S2, the semantic segmentation model is the DeepLab v3+ semantic segmentation model using the Mobilenet network.

[0016] Preferably, in step S4, the operational amplifier chip's sequence temperature matrix T I Determine the start time t0 of the dynamic heating zone, the start time t1 of the stable thermal equilibrium state, and the stop time t2. Define t0-t1 as the dynamic heating zone and t1-t2 as the steady-state constant temperature zone.

[0017] Preferably, in step S5, T train The specific steps for extracting categorical features from the input to the long short-term memory time series feature extraction model that incorporates an attention mechanism are as follows:

[0018] For training set T train Perform data preprocessing to obtain the dataset T. train The expected value μ and variance σ of each dimension are used to convert each data point into a standardized score, as shown in the following formula:

[0019]

[0020] Among them, z i,t For each sequence point in the i-th dimension after standardization; x i For the temperature data of each sequence point in the i-th dimension of the training set;

[0021] Then the processed T train The input is fed into a long short-term memory time series classification feature extraction model that incorporates attention mechanisms for the extraction of classification features.

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] 1. This invention effectively solves the problem of inaccurate fault detection caused by factors such as changes in the external environment, insignificant temperature differences in fault categories, and component aging in the prior art by considering the temperature change trend and fault category information of the temperature sequence, and by using infrared images and deep learning algorithms for fault classification, thereby improving the accuracy and generalization of fault detection. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating the overall structure of the infrared-based airborne electronic system circuit board operational amplifier chip fault diagnosis method of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] In this embodiment, the power supply module of the airborne weather radar system circuit board is selected as the test target.

[0027] Please see Figure 1 This invention provides a technical solution: a method for fault diagnosis of operational amplifier chips on airborne electronic system circuit boards based on infrared technology, comprising the following steps:

[0028] S1. Data Acquisition: First, an infrared image acquisition system is set up, which includes multiple infrared thermal imagers. The airborne electronic system circuit board under test (BEST) is placed flat on a horizontal plane, with the infrared thermal imager lens perpendicular to the top of the BEST circuit board. The BEST circuit board is connected to the excitation source via a programmable relay switch. The host computer is connected to both the infrared camera and the programmable relay switch. The host computer controls the programmable relay and the infrared camera to open and close synchronously, ensuring that the working state of the BEST circuit board is synchronized with the working state of the infrared camera. That is, in this invention, only the infrared image information of the BEST circuit board in its working state is acquired. When the BEST circuit board operates in different working modes, the infrared image acquisition system acquires the infrared images of the BEST circuit board in different working modes and names them as I. I The acquired infrared image information is then transmitted back to the host computer for storage. The computer processes the collected infrared image information to obtain the temperature information corresponding to each pixel of the airborne electronic system circuit board's infrared image. The processed temperature information is in the form of a two-dimensional data matrix, namely, the sequence temperature matrix corresponding to each pixel of the airborne electronic system circuit board's infrared sequence images under different operating modes, denoted as T. MThe infrared image of the airborne electrical system circuit board is 320×256 pixels, with a temperature accuracy of ±1% for each pixel and a temperature resolution of less than or equal to 0.015K for each pixel.

[0029] S2. Data Preprocessing: Use Labelme annotation software to process the infrared sequence images of the airborne electronic system circuit board acquired in step S1. I The key electronic components—operational amplifier chips—in the region of interest were labeled. After image labeling, image preprocessing was performed using image transformation methods such as translation, rotation, and mirror flipping. The preprocessed infrared sequence images of the airborne electronic system circuit board were then processed. I The improved DeepLab v3+ semantic segmentation model is fed into the infrared segmentation image of the key electronic components in the region of interest. II The improved DeepLab v3+ semantic segmentation model consists of two parts: an encoder and a decoder.

[0030] The encoder section should include the following components: the backbone network MobileNet, the hollow space pyramid pooling module, the convolutional block attention module, and the upsampling module;

[0031] The decoder should include the following components: a 1×1 convolution module, a feature fusion module, and an upsampling module.

[0032] This invention is primarily applied to the field of infrared imaging of circuit boards. Therefore, it replaces the Xception backbone network of the original DeepLab v3+ semantic segmentation model with the lighter Mobilenet network to improve model speed and reduce algorithm complexity. The hollow spatial pyramid pooling module can adjust the sampling ratio and receptive field, capturing target and related contextual information at multiple scales. The convolutional block attention module can highlight important features and suppress unimportant features (primarily referring to highlighting the segmentation of target components and suppressing background segmentation), thereby improving the model's segmentation performance.

[0033] S3, combined with the sequence temperature matrix T M Infrared segmentation image of key electronic components in the region of interest – operational amplifier chips I II The temperature sequence matrix T of the key electronic component in the region of interest—operational amplifier chips—is obtained. I The temperature sequence matrix T I The specific cargo volume zone method is as follows:

[0034] Infrared segmentation image I of the key electronic component—operational amplifier chip—in the region of interest in step S2. IIEdge coordinates are obtained from the edges of each electronic component. The coordinates of all pixels inside the edge are obtained by traversing the point set in the contour. A matrix M of size 320×256 is constructed, with the coordinates at the above positions set to 1 and all other positions set to 0. Matrix M is then compared with the sequence temperature matrix T. M Element-wise multiplication yields the temperature sequence matrix T of the key electronic component—the operational amplifier chip—in the region of interest. I .

[0035] T I (i,j)=M(i,j)*T M (i,j)

[0036] Where i and j represent the row and column of matrix M, respectively.

[0037] S4. Temperature State Region Division and Dataset Extraction: Based on the operating time, the temperature state of the electronic components—operational amplifier chips—in the region of interest of the airborne electronic system circuit board in step S3 is divided into a dynamic heating region and a steady-state isothermal region. The temperature sequence matrix T is extracted. I The data from the portion where the temperature continues to rise is denoted as the dynamic warming zone dataset, T. ID ; Extracting infrared segmentation images from operational amplifier chips I II Images of the stable temperature region are used as the image dataset for the steady-state isothermal region, denoted as I. IS The specific steps are as follows:

[0038] By analyzing the sequence temperature matrix T of the operational amplifier chip I Determine the start time t0 of the dynamic heating zone, the start time t1 of the stable thermal equilibrium state, and the stop time t2. Define t0-t1 as the dynamic heating zone and t1-t2 as the steady-state constant temperature zone.

[0039] Extracting the temperature sequence matrix T of the key electronic component—the operational amplifier chip—in the region of interest. I The temperature matrix sequence data during the time interval t0-t1 (from the start of circuit board operation to stable thermal equilibrium) is used as the dynamic heating zone dataset and named T. ID This represents the change in the temperature matrix of the operational amplifier chip (op-amp) within the region of interest (ROI) during the time interval from the start of circuit board operation to stable thermal equilibrium. The dynamic heating region sequence temperature information refers to the temperature information of the key electronic component—the op-amp chip—within the region of interest during this time interval, characterized by significant changes in the sequence temperature information of the key electronic component within this region over time. The dynamic heating region sequence temperature information data T... ID By performing averaging and maximizing operations, the mean sequence data T of the key electronic component—the operational amplifier chip—in the region of interest during the dynamic heating phase is obtained.D-mean and the maximum value sequence data T D-Max To further expand the feature information, the temperature change rate T′ of the mean sequence and the maximum sequence were then extracted respectively. D-mean and T′ D-Max The mean T D-mean Maximum value T D-Max The mean temperature change rate T′ D-mean The rate of temperature change T′ of the maximum value sequence D-Max By performing merging and splicing processing, multi-dimensional temperature sequence data T of the key electronic component in the region of interest—the operational amplifier chip—is obtained. This data characterizes the temperature variation of the target operational amplifier chip under various operating modes (normal mode and different fault modes).

[0040]

[0041] I. Infrared segmentation image of key electronic components—operational amplifier chips—within the region of interest. II The infrared image sequence during the time interval t1-t2 (from when the circuit board is in a stable working state to when it stops working) is named I. IS This represents the infrared image changes of the key electronic component—the operational amplifier chip—in the region of interest within the interval from the stable thermal equilibrium state of the circuit board to its stop working state. The sequential temperature information of the key electronic component—the operational amplifier chip—in this interval changes only slightly over time.

[0042] S5. Dataset Processing and Model Training: Process the dataset T... ID Further processing yields multidimensional temperature series data T. Image dataset I of steady-state isothermal region IS Set the same index and retrieve them separately. with I IS 70% of it is set as the training set and named T train and I train The rest with I IS 30% of the data is set as the validation set and named T. val and I val T train Inputting the data into a long short-term memory time series feature extraction model that incorporates an attention mechanism extracts classification features; I train The input is fed into the improved Transformer-Encoder network V8 infrared thermal image classification feature extraction model to extract classification features. The specific steps are as follows:

[0043] First, the training set T... train Perform data preprocessing to obtain the dataset T. trainThe expected value μ and variance σ of each dimension (different features, such as the maximum value feature and the rate of change of the maximum value feature) are used to convert each data point into a standardized score, as shown in the formula. The validation set is not processed.

[0044]

[0045] Among them, z i,t For each sequence point in the i-th dimension after standardization; x i The temperature data for each sequence point in the i-th dimension of the training set.

[0046] Then the processed T train The input is fed into the Long Short-Term Memory (LSTM) Time Series Feature Extraction Model Integrated with Attention Mechanism (LSTM&ATt) for feature extraction. The LSTM&ATt model consists of two cascaded LSTM networks that integrate an attention mechanism. In this network, a time query and a set of key-value pairs are mapped to an output. The query, key, and value are matrices obtained from the feature extraction of the LSTM networks. The output is defined as a weighted sum of the fully represented values. In the LSTM&ATt model, the first network extracts the basic relationships between temperature series data, while the second network expands the relationship details to obtain more complex relationships hidden within the data. The mathematical expression for each LSTM network integrating the attention mechanism is shown in the following equation:

[0047] I q =f LSTM-Q (z)

[0048] I k =f LSTM-K (z)

[0049] I v =f LSTM-V (z)

[0050] Among them, f LSTM-Q f LSTM-K f LSTM-V These represent the LSTM networks used to obtain the query, key, and value, respectively, and z represents the input to the training set; I q I k and I vThese represent vector matrices consisting of query, key, and value, respectively.

[0051] Finally, the output of the LSTM network with the attention mechanism is obtained as follows:

[0052]

[0053] Among them, f softmax This represents the Softmax function; This represents the transpose of the vector matrix consisting of keys.

[0054] Among them, I train The input is fed into an infrared thermal image classification feature extraction model based on a Transformer-Encoder network to complete the classification feature extraction. Compared with the recognition algorithm, the Transformer-Encoder network detection algorithm can focus on the object in the region of interest in the image while ignoring background information. This fully utilizes the thermal influence between multiple operational amplifier chips in the region of interest during the circuit board's operation and incorporates the features generated by the mutual influence of various components under different fault conditions into the recognition network, thereby improving the accuracy of classification.

[0055] S6. Feature Fusion and Model Classification Training: Fusion of the multidimensional temperature sequence data T obtained in S5 and the image dataset I of the steady-state isothermal region. IS The two types of feature information are input into the multilayer perceptron in the model to complete the model's classification training.

[0056] S7. Fault Classification and Diagnosis: Input the known category dataset into the trained model to obtain the classification of component faults in the region of interest of the airborne electronic system circuit board. Based on the classification results, diagnose the faults of operational amplifier chips on the airborne electronic system circuit board.

[0057] This invention, after making in-depth improvements and optimizations to address the shortcomings of existing technologies, brings significant beneficial effects, as detailed below:

[0058] Reducing the risk of secondary damage to faulty aircraft parts: In existing technologies, traditional injection-type testing methods, when applied to airborne circuit boards, may cause secondary damage to faulty parts due to their complex structure and high cost. However, the non-destructive testing technology used in this invention, particularly infrared technology, can accurately locate faults without extensive disassembly of complex airborne electronic systems, effectively avoiding the dangers and secondary damage that may be caused by high-voltage, high-current operating environments.

[0059] Simplifying the testing process and reducing testing difficulty: Existing technologies often require detailed analysis of the circuit board's operating mechanism when detecting circuit board faults, which not only increases the difficulty of testing but also raises the professional requirements for technicians. This invention, however, by collecting infrared temperature data from the circuit board, can intuitively extract the temperature characteristics of different fault modes to complete fault diagnosis, eliminating the need for in-depth analysis of the circuit board's operating mechanism and thus reducing the difficulty of testing.

[0060] Improving the generalizability and applicability of detection methods: Traditional circuit testing methods in the prior art often require the purchase of corresponding diagnostic equipment according to different models and systems, which increases maintenance costs and reduces the generalizability of the detection methods. However, the infrared detection technology used in this invention only requires an infrared circuit board data acquisition system composed of a CNC computer and a thermal imager to complete the fault diagnosis of most circuit board operational amplifier chips, which not only reduces maintenance costs but also improves the generalizability and applicability of the detection methods.

[0061] Improving the accuracy and efficiency of fault detection: Existing technologies often face challenges in fault detection accuracy when dealing with complex external environments (such as the effects of light and room temperature) and varying degrees of aging of circuit board components. This invention, by considering the temperature variation trends of temperature sequences and the fault category information hidden within them, and combining algorithms such as deep learning, can more accurately extract fault information from circuit boards, thereby improving the accuracy and efficiency of fault detection.

[0062] In summary, this invention overcomes the limitations of traditional methods and improves diagnostic accuracy by combining visible light images with infrared thermal image sequences. By constructing an infrared image acquisition system, infrared sequence images and temperature matrices of circuit boards under different operating modes are acquired, and then deep learning algorithms are used for classification training. This method is particularly suitable for fault diagnosis in complex scenarios such as airborne electronic system circuit boards, achieving non-contact, rapid, and intuitive detection results.

[0063] This invention details a method for infrared image processing and temperature state analysis of key electronic components. First, an improved model is used to achieve precise segmentation of infrared images to obtain the component's temperature information. Second, the operating state is divided into different regions, and temperature sequence data and infrared image data are extracted separately for each region. Furthermore, this paper proposes a method for processing the sequential temperature information of electronic components, constructing a rich dataset by calculating multidimensional temperature data and extracting infrared image sequences.

[0064] For feature extraction, this paper employs an LSTM network with an attention mechanism and a Transformer-Encoder network. These models can extract the basic relationships between temperature sequence data and reveal more complex relationships, thereby effectively expanding feature information and improving classification accuracy. Experimental results show that this method can focus on the region of interest, ignore background information, and achieve efficient and accurate fault diagnosis.

[0065] Compared to traditional detection methods, this approach reduces the difficulty of detection, improves generalization, and avoids the dangers of secondary damage and high-voltage, high-current environments. This technology is of great significance for improving the efficiency and accuracy of fault diagnosis in airborne electronic system circuit boards, and is expected to reduce maintenance costs and enhance system safety. Therefore, this method has broad application prospects in the fault detection of airborne electronic system circuit boards.

[0066] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 process, method, article, or apparatus.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for fault diagnosis of operational amplifier chips on circuit boards of airborne electronic systems based on infrared technology, characterized in that: Includes the following steps: S1. Data Acquisition: Acquire infrared sequence images of the airborne electronic system circuit boards under different operating modes. and the corresponding sequence temperature matrix ; S2, Data Preprocessing: The infrared sequence images from step S1... After preprocessing, the images are fed into a semantic segmentation model to obtain infrared segmented images of key electronic components—operational amplifier chips—in the region of interest. ; S3. Obtaining the temperature sequence matrix: Combining the temperature sequence matrix from step S1... Infrared segmentation image of the operational amplifier chip in step S2 The temperature sequence matrix of the operational amplifier chip is obtained. ; S4. Temperature State Region Division and Dataset Extraction: Based on the operating time, the temperature state of the electronic components—operational amplifier chips—in the region of interest of the airborne electronic system circuit board is divided into a dynamic heating region and a steady-state isothermal region, and the temperature sequence matrix is ​​extracted. The data from the portion where the temperature continues to rise is denoted as the dynamic warming zone dataset. , the dataset Further processing yields multidimensional temperature sequence data T; infrared segmentation images of operational amplifier chips are extracted. Images of the stable temperature region are used as the image dataset of the steady-state isothermal region, denoted as . ; S5. Dataset Processing and Model Training: Process the multidimensional temperature sequence data T and the image dataset of the steady-state isothermal region from step S4. Take 70% of each and set it as the training set, named and ; The remaining 30% of the data was set as the validation set, named as follows: and The training set of multidimensional temperature sequence data T is input into a long short-term memory time series feature extraction model that incorporates an attention mechanism to extract classification features. Simultaneously, the image dataset of the steady-state isothermal region is used. The training set is input into the infrared thermal image classification feature extraction model to extract classification features; S6. Feature Fusion and Model Classification Training: Fusion of the multidimensional temperature sequence data obtained in S5 Image datasets of steady-state isothermal regions The two types of feature information are input into the multilayer perceptron in the model to complete the model's classification training; S7. Fault Classification and Diagnosis: Input the known category dataset into the trained model to obtain the classification of component faults in the region of interest of the airborne electronic system circuit board. Based on the classification results, diagnose the faults of operational amplifier chips on the airborne electronic system circuit board.

2. The method for fault diagnosis of operational amplifier chips on airborne electronic system circuit boards based on infrared as described in claim 1, characterized in that: In step S1, the infrared sequence image is processed. and the corresponding sequence temperature matrix The data acquisition was achieved using multiple infrared thermal imagers positioned perpendicularly above the circuit board of the airborne electronic system under test. The circuit board of the airborne electronic system under test was connected to the excitation source via a programmable relay switch, and the host computer was connected to both the infrared camera and the programmable relay switch.

3. The method for fault diagnosis of operational amplifier chips on airborne electronic system circuit boards based on infrared as described in claim 1, characterized in that: In step S2, the infrared sequence image is processed. Preprocessing includes, but is not limited to, translation, rotation and mirroring. In step S2, the semantic segmentation model is the DeepLab v3+ semantic segmentation model using the Mobilenet network.

4. The method for fault diagnosis of operational amplifier chips on airborne electronic system circuit boards based on infrared as described in claim 1, characterized in that: The operational amplifier chip's sequence temperature matrix in step S4 Determine the start time of the dynamic heating zone The start time of a stable thermal equilibrium state Stop working time ,Will Defined as a dynamic heating zone, Defined as the steady-state isothermal region.

5. The method for fault diagnosis of operational amplifier chips on airborne electronic system circuit boards based on infrared as described in claim 1, characterized in that: In step S5, The specific steps for extracting categorical features from the input to the long short-term memory time series feature extraction model that incorporates an attention mechanism are as follows: For the training set Perform data preprocessing to obtain the dataset Expectations in various dimensions and variance Each data point is converted into a standard score, as shown in the formula: in, For the standardized process of the first Temperature data for each sequence point in each dimension; For the training set Temperature data for each sequence point in each dimension; After processing The input is fed into a long short-term memory time series classification feature extraction model that incorporates attention mechanisms for the extraction of classification features.

Images