Radiation level monitoring method, system, electronic device, and storage medium

Abstract

The application discloses a radiation level monitoring method and system, an electronic device and a storage medium. The method comprises the following steps: collecting first radiation data through a preset radiation monitoring sensor; performing data dimension reduction processing on the first radiation data through a feature dimension reduction model to obtain second radiation data; the feature dimension reduction model comprises a boundary smoothing-based autoencoder; performing data augmentation on the second radiation data through a preset generative adversarial network model to obtain third radiation data; the preset generative adversarial network model comprises a quantum state-based generative adversarial network model; performing model training on a preset monitoring model through a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model; inputting to-be-analyzed radiation data into the target radiation monitoring model for radiation level analysis to obtain a radiation level monitoring result. The embodiment of the application can effectively alleviate the data privacy risk problem and improve the accuracy of radiation level monitoring. The application can be widely applied to the technical field of monitoring.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of monitoring, in particular to a radiation level monitoring method and system, an electronic device and a storage medium. BACKGROUND

[0002] Radiation monitoring is widely used in various key industries, such as nuclear power plants, environmental protection, medical fields, etc. In the related art, radiation monitoring data usually has high-dimensional and complex characteristics, and in the traditional radiation monitoring method, there are often data privacy risks, and the accuracy of radiation monitoring is low.

[0003] To sum up, the technical problems existing in the related art need to be improved. SUMMARY

[0004] The main purpose of the embodiments of the present application is to provide a radiation level monitoring method and system, an electronic device and a storage medium, which can effectively alleviate the problem of data privacy risk and effectively improve the accuracy of radiation level monitoring.

[0005] To achieve the above purpose, one aspect of the embodiments of the present application provides a radiation level monitoring method, which comprises the following steps:

[0006] acquiring first radiation data through a preset radiation monitoring sensor;

[0007] performing data dimension reduction processing on the first radiation data through a feature dimension reduction model to obtain second radiation data; wherein the feature dimension reduction model comprises a boundary smoothing based autoencoder;

[0008] performing data augmentation on the second radiation data through a preset generative adversarial network model to obtain third radiation data; wherein the preset generative adversarial network model comprises a quantum state based generative adversarial network model;

[0009] training a preset monitoring model through a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model;

[0010] inputting the radiation data to be analyzed into the target radiation monitoring model for radiation level analysis to obtain a radiation level monitoring result.

[0011] In some embodiments, before performing the data dimension reduction processing on the first radiation data through the feature dimension reduction model to obtain the second radiation data, the method further comprises:

[0012] constructing an initial autoencoder;

[0013] inputting preset input data into the initial autoencoder to generate reconstruction data through forward propagation;

[0014] Smooth the reconstruction data by an automatic boundary smoothing algorithm to obtain smooth output data;

[0015] Calculate a weight coefficient according to the preset input data by a preset variance function;

[0016] Calculate a first loss function according to the smooth output data, the preset input data, and the weight coefficient;

[0017] Update parameters of the initial autoencoder according to the first loss function by a gradient descent algorithm to obtain the feature dimension reduction model.

[0018] In some embodiments, the updating parameters of the initial autoencoder according to the first loss function by a gradient descent algorithm to obtain the feature dimension reduction model comprises:

[0019] Calculate a first expected learning rate by a convex hull convergence algorithm;

[0020] Calculate a parameter update amount according to the first expected learning rate and the first loss function;

[0021] Update parameters of the initial autoencoder according to the parameter update amount to obtain the feature dimension reduction model.

[0022] In some embodiments, the data augmentation of the second radiation data by a preset generative adversarial network model to obtain third radiation data comprises:

[0023] Construct an initial generative adversarial network model; wherein the initial generative adversarial network model comprises a generator and a discriminator;

[0024] Receive random noise by the generator to convert the random noise into quantum state data by a quantum state encoding algorithm;

[0025] Generate preset radiation data according to the quantum state data;

[0026] Calculate similarity of the preset radiation data by the discriminator according to a similarity measurement algorithm to obtain an evaluation result;

[0027] Adjust parameters of the initial generative adversarial network model according to the evaluation result to obtain the preset generative adversarial network model.

[0028] In some embodiments, the model training of a preset monitoring model according to the third radiation data by a federated learning architecture to obtain a target radiation monitoring model comprises:

[0029] constructing a center model and a local model; wherein the local model comprises a plurality of extreme learning machine models, and the plurality of extreme learning machine models correspond to the third radiation data;

[0030] training the corresponding extreme learning machine model through the third radiation data respectively to obtain a local update parameter;

[0031] transmitting the local update parameter to the center model to perform an aggregated update on the center model to obtain a target update parameter;

[0032] transmitting the target update parameter to each of the extreme learning machine models respectively to perform a parameter update on the extreme learning machine models to obtain the target radiation monitoring model.

[0033] In some embodiments, the training the corresponding extreme learning machine model through the third radiation data respectively to obtain a local update parameter comprises:

[0034] inputting the third radiation data into the extreme learning machine model to perform a nonlinear mapping to obtain intermediate output data;

[0035] calculating target output data according to the intermediate output data and a weight matrix;

[0036] calculating loss data through a second loss function according to the target output data to determine weight update data through the loss data; wherein the second loss function comprises a gradient penalty coefficient;

[0037] performing a parameter update on the extreme learning machine model through a back propagation algorithm according to the weight update data to obtain the local update parameter.

[0038] In some embodiments, after performing the parameter update on the extreme learning machine model through the back propagation algorithm according to the weight update data to obtain the local update parameter, the method further comprises:

[0039] calculating a second expected learning rate according to the weight update data, the intermediate output data and a basic learning rate;

[0040] adjusting a model training learning rate according to the second expected learning rate.

[0041] To achieve the above object, another aspect of the embodiment of the present application proposes a radiation level monitoring system, which comprises:

[0042] a first module configured to collect first radiation data through a preset radiation monitoring sensor;

[0043] The second module is configured to perform data dimension reduction processing on the first radiation data by using a feature dimension reduction model to obtain second radiation data, wherein the feature dimension reduction model comprises a boundary smoothing-based autoencoder;

[0044] The third module is configured to perform data augmentation on the second radiation data by using a preset generative adversarial network model to obtain third radiation data, wherein the preset generative adversarial network model comprises a quantum state-based generative adversarial network model;

[0045] The fourth module is configured to perform model training on a preset monitoring model by using a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model.

[0046] The fifth module is configured to input to-be-analyzed radiation data into the target radiation monitoring model to perform radiation level analysis and obtain a radiation level monitoring result.

[0047] To achieve the above object, another aspect of the embodiment of the present application provides an electronic device, which comprises:

[0048] at least one processor;

[0049] at least one memory configured to store at least one program;

[0050] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.

[0051] To achieve the above object, another aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the above method.

[0052] The embodiments of the present application at least have the following beneficial effects: The present application provides a radiation level monitoring method, system, electronic device and storage medium. The first radiation data is collected by a preset radiation monitoring sensor, and the first radiation data is subjected to data dimension reduction processing by a feature dimension reduction model to obtain second radiation data. In the embodiments of the present application, the feature dimension reduction model includes a boundary smoothing-based autoencoder. The first radiation data is subjected to dimension reduction processing by the boundary smoothing-based autoencoder, thereby effectively reducing data redundancy, improving the robustness of the model, and further improving the accuracy of radiation level monitoring. Further, the second radiation data is subjected to data augmentation by a preset generative adversarial network model to obtain third radiation data. Accordingly, the preset generative adversarial network model in the embodiments of the present application includes a quantum state-based generative adversarial network model. The third radiation data is generated by the quantum state-based generative adversarial network model, which can effectively alleviate the problem of insufficient training samples, so that the model can learn more complex radiation data distribution, thereby improving the generalization ability of the model and effectively improving the accuracy of radiation level monitoring. Then, the preset monitoring model is subjected to model training by a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model, and then the radiation data to be analyzed is input into the target radiation monitoring model for radiation level analysis to obtain a radiation level monitoring result, thereby realizing accurate monitoring of the radiation level. Accordingly, the embodiments of the present application can alleviate the data privacy risk problem by means of federated learning, thereby effectively improving the data privacy protection level. It is easy to understand that the embodiments of the present application can effectively improve the radiation level monitoring accuracy of the target radiation monitoring model obtained by training by means of the boundary smoothing-based autoencoder for data dimension reduction and the quantum state-based generative adversarial network model for data augmentation. At the same time, the federated learning architecture can effectively alleviate the data privacy risk problem. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 FIG. 1 is a flowchart of a radiation level monitoring method provided by the embodiments of the present application;

[0054] Figure 2 FIG. 2 is a flowchart of feature dimension reduction model training provided by the embodiments of the present application;

[0055] Figure 3 FIG. 3 is a schematic diagram of a federated learning architecture provided by the embodiments of the present application;

[0056] Figure 4 FIG. 4 is a structural schematic diagram of a radiation level monitoring system provided by the embodiments of the present application;

[0057] Figure 5 FIG. 5 is a hardware structural schematic diagram of an electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0058] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific examples described herein are only intended to explain the present application and are not intended to limit the present application. When the following description refers to the accompanying drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with embodiments of the present application. They are merely examples of apparatuses and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0059] It can be understood that the terms "first", "second", and the like used in the present application can be used herein to describe various concepts, but unless specifically stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present application, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information. Depending on the context, the word "if" as used herein can be interpreted as "when" or "when" or "in response to determining".

[0060] The terms "at least one", "multiple", "each", "any" and the like used in the present application include one, two or more than two, multiple includes two or more than two, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0062] Before the embodiments of the present application are described in detail, first, some nouns and terms involved in the embodiments of the present application are described, and the nouns and terms involved in the embodiments of the present application are applicable to the following explanations.

[0063] Federated learning architecture: a distributed machine learning architecture whose core is to allow multiple participants to jointly train a shared machine learning model while maintaining their own data privacy and localization.

[0064] Generative adversarial network: a deep learning model composed of a generator and a discriminator. Among them, the generator is responsible for generating realistic data samples, while the discriminator aims to distinguish between the false samples generated by the generator and the real data samples.

[0065] Radiation monitoring is widely used in various key industries, such as nuclear power plants, environmental protection, medical fields, etc. In the related art, radiation monitoring data usually has high dimension and complexity. In the traditional radiation monitoring method, there is often a data privacy risk, and the accuracy of radiation monitoring is low. Exemplarily, the traditional radiation monitoring method depends on a centralized data processing architecture, that is, the monitoring data is transmitted to a central server for analysis and processing. This method has a high risk of data privacy, especially when processing sensitive radiation monitoring data, which may be threatened by data leakage, hacker attacks, etc. In addition, when the feature dimension is high, the traditional radiation monitoring method often uses a simple dimension reduction method or does not perform dimension reduction processing, which may cause information loss or fail to effectively extract important features in the data. At the same time, in actual radiation monitoring, the training data is often insufficient, and the number of samples at different radiation levels is unbalanced, which causes the model to fail to generalize well. In addition, the collected radiation data is often high-latitude, nonlinear complex data, which easily leads to low classification accuracy.

[0066] Therefore, in the embodiments of the present application, a radiation level monitoring method, system, electronic device and storage medium are provided. The scheme acquires first radiation data through a preset radiation monitoring sensor, and performs data dimension reduction processing on the first radiation data through a feature dimension reduction model to obtain second radiation data. In the embodiments of the present application, the feature dimension reduction model includes a boundary smoothing based autoencoder. Then, the present application performs data augmentation on the second radiation data through a preset generative adversarial network model to obtain third radiation data. Correspondingly, the preset generative adversarial network model in the embodiments of the present application includes a quantum state based generative adversarial network model. Then, the present application performs model training on a preset monitoring model according to the third radiation data through a federated learning architecture to obtain a target radiation monitoring model, and then inputs the radiation data to be analyzed into the target radiation monitoring model for radiation level analysis to obtain a radiation level monitoring result, thereby realizing accurate monitoring of the radiation level, effectively alleviating the data privacy risk problem, and effectively improving the accuracy of radiation level monitoring.

[0067] The radiation level monitoring method provided by the embodiments of the present application relates to the technical field of monitoring. The radiation level monitoring method provided by the embodiments of the present application can be applied to a terminal, can be applied to a server, and can also be software running in the terminal or the server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, and the like, but is not limited thereto; the server end can be configured as a stand-alone physical server, can be configured as a server cluster or a distributed system formed by multiple physical servers, can be configured as a cloud server providing basic cloud computing services such as cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, and big data and artificial intelligence platform, and the server can also be a node server in a blockchain network; and the software can be an application that implements the radiation level monitoring method, and the like, but is not limited to the above forms.

[0068] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as a program module. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment in which tasks are performed by remote processing devices connected by a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0069] Figure 1 The method in the radiation level monitoring method provided by the embodiments of the present application is an optional flowchart, Figure 1 The method in the radiation level monitoring method provided by the embodiments of the present application is an optional flowchart,

[0070] Step S110: acquiring first radiation data by a preset radiation monitoring sensor.

[0071] Step S120: performing data dimension reduction processing on the first radiation data by a feature dimension reduction model to obtain second radiation data. The feature dimension reduction model includes a boundary smoothing-based autoencoder.

[0072] Step S130: performing data augmentation on the second radiation data by a preset generative adversarial network model to obtain third radiation data. The preset generative adversarial network model includes a generative adversarial network model based on a quantum state.

[0073] Step S140: performing model training on a preset monitoring model according to the third radiation data by a federated learning architecture to obtain a target radiation monitoring model.

[0074] Step S150: inputting the to-be-analyzed radiation data into the target radiation monitoring model to perform radiation level analysis and obtain a radiation level monitoring result.

[0075] In the working process of the embodiment, the embodiment first collects first radiation data by a preset radiation monitoring sensor. Specifically, in the embodiment, each client collects data by various radiation monitoring sensors. For example, the preset radiation monitoring sensor in the embodiment includes a gamma ray sensor, an X-ray detector, an alpha particle detector, and a beta particle detector. These sensors can monitor different types of radiation levels in the environment in real time, obtain radiation signals through different physical principles, and then convert them into digital signals for processing to obtain the first radiation data. In the embodiment, the collected radiation data is transmitted to a cloud server or a local storage device through a wireless network or a wired network, and the collected data is stored in a standardized format, such as CSV format.

[0076] For example, the data attributes of the collected first radiation data in the embodiment include a timestamp (Ta) for recording the specific time of data collection, a sensor position (Pa) for marking the geographic location of the data collection point, a gamma ray radiation value (Ra) for the measurement result of the gamma ray, a unit of μSv / h, an X-ray radiation value (Xa) for the measurement result of the X-ray, a unit of μSv / h, an alpha particle radiation value (α a ) for the measurement result of the alpha particle, a unit of Bq, a beta particle radiation value (β a ) for the measurement result of the beta particle, a unit of Bq, a temperature (Ta) for the ambient temperature, a unit of Celsius, a humidity (Ha) for the ambient humidity, a unit of percentage, an atmospheric pressure (Pa) for the ambient atmospheric pressure, a unit of hPa, and a wind speed (Va) for the ambient wind speed, a unit of m / s. In some embodiments of the application, the number of data attributes is usually more than 10, and the number of data attributes may reach dozens or even hundreds.

[0077] Further, the embodiment of the present application obtains second radiation data by performing data dimension reduction processing on the first radiation data through a feature dimension reduction model. Specifically, the feature dimension reduction model in the embodiment of the present application includes a self-encoder based on boundary smoothing. Accordingly, since the first radiation data collected by the embodiment of the present application has a high data dimension, it may lead to information loss or failure to effectively extract important features in the data. Therefore, the embodiment of the present application performs dimension reduction processing on the high-dimension data (first radiation data) through the self-encoder based on boundary smoothing to obtain second radiation data, effectively alleviating the information loss problem that may be caused by the traditional dimension reduction method, and enhancing the robustness of the model to outliers and boundary effects through the boundary smoothing strategy, thereby improving the processing capability of the monitoring data. Then, the embodiment of the present application performs data augmentation on the second radiation data through a preset generative adversarial network model to obtain third radiation data. Specifically, the preset generative adversarial network model in the embodiment of the present application includes a quantum state-based generative adversarial network model. Since insufficient training data and unbalanced sample quantity may easily lead to poor generalization of the model. Therefore, the embodiment of the present application obtains third radiation data by performing data augmentation on the second radiation data through the quantum state-based generative adversarial network model, thereby enriching the training samples and effectively improving the generalization capability of the model, so that it can more accurately identify different types of radiation levels. The quantum state-based generative adversarial network model is a quantum machine learning model that combines quantum computing and generative adversarial networks. The embodiment of the present application generates high-quality and diversified synthetic radiation data, i.e., third radiation data, through the quantum state-based generative adversarial network model, thereby effectively alleviating the problem of insufficient training samples in actual application and enhancing the generalization capability of the model.

[0078] Further, the embodiment of the present application trains a preset monitoring model through a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model. Specifically, after each client collects corresponding first radiation data through a corresponding preset radiation monitoring sensor, the corresponding third radiation data is obtained through data dimension reduction processing and data augmentation. In order to protect the data privacy of each client and alleviate the risk of data leakage, the embodiment of the present application trains the preset monitoring model through the federated learning architecture, trains the model independently at each client and only exchanges model parameters, thereby effectively improving the data protection level. Finally, the embodiment of the present application inputs the radiation data to be analyzed into the target radiation monitoring model for radiation level analysis to obtain a radiation level monitoring result. Specifically, the radiation data to be analyzed in the embodiment of the present application refers to radiation data that needs to be analyzed for radiation level monitoring. Accordingly, the embodiment of the present application performs radiation level analysis on the radiation data to be analyzed through the target radiation monitoring model trained to obtain a radiation level monitoring result, such as "low radiation", "medium radiation" and "high radiation".

[0079] It should be noted that, in this embodiment of the invention, the collected data is first labeled before the quantum state-based generative adversarial network algorithm generates samples. The labeling method in this embodiment is manual labeling, and the labeled categories include: "low radiation," "medium radiation," and "high radiation," for a total of three categories.

[0080] In some embodiments of the present invention, before performing data dimensionality reduction processing on the first radiation data using a feature dimensionality reduction model to obtain the second radiation data, the radiation level monitoring method provided in the embodiments of the present invention further includes, but is not limited to, the following steps:

[0081] Construct the initial autoencoder.

[0082] The preset input data is input into the initial autoencoder and reconstructed data is generated through forward propagation.

[0083] The reconstructed data is smoothed using an automatic boundary smoothing algorithm to obtain smoothed output data.

[0084] The weight coefficients are calculated based on the preset input data using a preset variance function.

[0085] The first loss function is calculated based on the smoothed output data, the preset input data, and the weights.

[0086] Based on the first loss function, the parameters of the initial autoencoder are updated using the gradient descent algorithm to obtain the feature dimensionality reduction model.

[0087] In this specific embodiment, before performing data dimensionality reduction processing using a feature dimensionality reduction model, an autoencoder based on boundary smoothing is first trained, such as... Figure 2 As shown. Specifically, in this embodiment of the invention, an initial autoencoder is first constructed. For example, in this embodiment of the invention, an initial autoencoder is constructed and the corresponding autoencoder parameters are initialized, as shown in the following equation (1):

[0088]

[0089] Where, in the formula Represents the encoder weights, Indicates the encoder bias; Indicates the weights of the decoder. This represents the decoder bias; randn() represents the random function; d r,in d is the dimension of the encoder input layer. r,mid d is the dimension of the encoder's intermediate layer. r,out It is the dimension of the encoder output layer; zeros() is a function that generates a vector of all zeros.

[0090] Next, in this embodiment of the invention, the preset input data is input into the initial autoencoder to generate reconstructed data through forward propagation, and then the reconstructed data is smoothed using an automatic boundary smoothing algorithm to obtain smoothed output data. Here, the preset input data refers to the data used for training the feature reduction model. Accordingly, in order to handle outliers and boundary effects during the dimensionality reduction process, this embodiment of the invention adopts an automatic boundary smoothing strategy, which enhances the robustness of the model by dynamically adjusting boundary values. The smoothing function is used to smooth the encoder output (reconstructed data), as shown in equation (2) below:

[0091]

[0092] Where, in the formula It is the output of the encoder, τ r () is a smoothing function, δ r It is the smoothing intensity parameter; θ r It is an automatically adjusted threshold parameter used to control the degree of smoothing and improve adaptability to input data boundaries.

[0093] It should be noted that the smoothing intensity parameter in this embodiment of the invention is determined using an adaptive method based on local data density, and the calculation method is shown in the following equation (3):

[0094]

[0095] Where, in the formula calculate In its neighborhood Kernel density estimation in the model can accurately handle data in both high-density and sparse regions, improving the robustness and adaptability of the overall model. Represents the neighborhood; k r () is the kernel density estimation function; mean() is the average value function; ∈ rea This is a preset small constant. For example, in the embodiments of the present invention, ∈ rea Set to 0.001.

[0096] Next, in this embodiment of the invention, weight coefficients are calculated using a preset equation function based on preset input data. Accordingly, in this embodiment of the invention, the weight coefficients of the autoencoder are used to adjust the contribution of each dimension of features to the error. These weights are dynamically allocated according to the importance of the features, and the calculation method is shown in equation (4) below:

[0097]

[0098] Where, in the formula It is the feature variance of the input data of the i-th dimension autoencoder; These are the features of the input data of the i-th dimension autoencoder; ∈ reis a small constant to avoid division by zero; Var() is the variance function.

[0099] Further, the embodiment of the present application calculates a first loss function according to the smoothed output data, the preset input data and the weight coefficient, and updates the parameters of the initial encoder through a gradient descent algorithm according to the first loss function, so as to obtain the feature dimension reduction model. In the training process of the autoencoder, the data is processed through forward propagation, and then the parameters of the autoencoder are updated through error back propagation using the gradient descent method. Correspondingly, in this process, the calculation method of the first loss function of the autoencoder in the embodiment of the present application is as shown in the following formula (5):

[0100]

[0101] wherein, L r is the first loss function of the autoencoder; x r is the input data of the autoencoder, is the output reconstructed by the decoder, Sig() is the Sigmoid activation function; is the L2 norm; is the output of the smoothed encoder; is the weight coefficient of the i-dimensional autoencoder; n r is the vector dimension corresponding to the sample input into the autoencoder in the current batch. Correspondingly, the calculation formula of the output reconstructed by the decoder in the embodiment of the present application is as shown in the following formula (6):

[0102]

[0103] Correspondingly, the training of the autoencoder in the embodiment of the present application is completed by continuously optimizing the first loss function in the iteration process. The calculation formula of the parameter update amount of the parameter updated each time is as shown in the following formula (7):

[0104]

[0105] wherein, in the formula is the weight parameter update amount of the encoder; is the bias parameter update amount of the encoder; is the weight parameter update amount of the decoder; is the bias parameter update amount of the decoder; is the learning rate of the autoencoder in the t-th iteration.

[0106] Further, the updating method of the parameters of the autoencoder in the embodiment of the present application is as shown in the following formula (8):

[0107]

[0108] wherein, in the formula, ← represents a parameter updating operation.

[0109] It is easy to understand that the embodiments of the present application obtain the feature dimension reduction model by repeating the above steps until the preset stopping iteration condition is met, i.e., the model training is completed. For example, the preset stopping iteration condition in the embodiments of the present application is that the preset maximum iteration number is reached, such as the preset maximum iteration number is set to 1000 times.

[0110] In some embodiments of the present application, the initial autoencoder is updated in parameters according to the gradient descent algorithm based on the first loss function to obtain the feature dimension reduction model, including but not limited to the following steps:

[0111] The first expected learning rate is calculated by the convex hull convergence algorithm.

[0112] The parameter updating amount is calculated according to the first expected learning rate and the first loss function.

[0113] The initial autoencoder is updated in parameters according to the parameter updating amount to obtain the feature dimension reduction model.

[0114] In the specific embodiments, in order to make the network converge quickly, the embodiments of the present application optimize the training process of the autoencoder by the convex hull convergence strategy. Specifically, the embodiments of the present application first calculate the first expected learning rate by the convex hull convergence algorithm, calculate the parameter updating amount according to the first expected learning rate and the first loss function, and then update the initial autoencoder in parameters according to the parameter updating amount to obtain the feature dimension reduction model. Accordingly, the embodiments of the present application adjust the learning step and the weight updating strategy to make the network converge quickly, and the learning rate is dynamically adjusted according to the data distribution characteristics in the current iteration, as shown in the following formula (9):

[0115]

[0116] wherein, in the formula, is the learning rate of the initial autoencoder; is the learning rate of the autoencoder in the tth iteration; γ r is the learning rate decay factor of the autoencoder, is the average diameter of the convex hull in the tth iteration, is the diameter of the initial convex hull, and the learning rate decreases with the decrease of the diameter of the convex hull to prevent instability caused by too large step size in the later learning period.

[0117] Accordingly, the embodiments of the present application substitute the calculated first expected learning rate into the above formula (7) to calculate the corresponding parameter updating amount, and then update the initial autoencoder in parameters to obtain the feature dimension reduction model.

[0118] In some embodiments of the present application, the second radiation data is data augmented by a preset generative adversarial network model to obtain third radiation data, including but not limited to the following steps:

[0119] An initial generative adversarial network model is constructed. The initial generative adversarial network model includes a generator and a discriminator.

[0120] The generator receives random noise to convert the random noise into quantum state data by a quantum state encoding algorithm.

[0121] The preset radiation data is generated according to the quantum state data.

[0122] The discriminator calculates the similarity of the preset radiation data according to a similarity measurement algorithm to obtain an evaluation result.

[0123] The initial generative adversarial network model is adjusted according to the evaluation result to obtain a preset generative adversarial network model.

[0124] In the specific embodiment, the embodiment of the present application first trains a quantum state-based generative adversarial network model before data augmentation. Specifically, the embodiment of the present application first constructs an initial generative adversarial network model. The initial generative adversarial network model in the embodiment of the present application includes a generator and a discriminator. Accordingly, the generator in the embodiment of the present application is used to generate synthetic data close to real radiation data, and the discriminator is used to distinguish between generated data and real data. For example, the embodiment of the present application defines the initialization parameters of the generator G and the discriminator D of the quantum state-based generative adversarial network. The parameters of the generator G are represented by b G , which are initialized as random values, as shown in the following formula (10):

[0125] b G = rand(θ G ) (10)

[0126] In the formula, θ G represents the parameter dimension of the generator.

[0127] At the same time, the parameters of the discriminator D in the embodiment of the present application are represented by b D , as shown in the following formula (11):

[0128] b D = rand(θ D ) (11)

[0129] In the formula, θ D represents the parameter dimension of the discriminator.

[0130] Then, the embodiment of the present application receives random noise through the generator to convert the random noise into quantum state data through a quantum state encoding algorithm. Accordingly, in the embodiment of the present application, the generator receives random noise as input, and converts the noise into data representation with high-dimensional characteristics through a quantum state encoding process, to enhance the representation ability of data through the principle of quantum computing. For example, in the embodiment of the present application, the input noise vector is represented by b z The quantum state-based process is shown in the following formula (12):

[0131]

[0132] wherein, in the formula represents a quantum state encoding function, which encodes classical information onto a quantum state. b z is a noise vector randomly sampled from a Gaussian distribution.

[0133] Accordingly, in the embodiment of the present application, the quantum state encoding function encodes classical information b z onto a quantum state, and uses quantum gate operations to achieve the encoding. For example, in the embodiment of the present application, a quantum rotation gate R(θ) and a quantum superposition state are used, where θ is the rotation angle, which can be calculated from the elements of the input vector b z , as shown in the following formula (13):

[0134]

[0135] wherein, in the formula is the i-th element of the vector b z , and θ i is the rotation angle of the corresponding quantum rotation gate. In the embodiment of the present application, the input noise vector is randomly sampled from a standard normal distribution, and the rotation angle θ i of each element is set to

[0136] Meanwhile, in the embodiment of the present application, the quantum state encoding function is a series of quantum rotation operations, as shown in the following formula (14):

[0137]

[0138] wherein, in the formula is a quantum rotation operation.

[0139] Further, the embodiment of the present application generates preset radiation data according to quantum state data. Wherein, the embodiment of the present application will be used to generate synthesized radiation data based on quantum state noise, which is not only diversified but also high quality, which can help the model to learn more complex data distribution. For example, the embodiment of the present application generates the vector b encoded by quantum state encoding gen The generator G generates the synthesized data b gen The way is as shown in the following formula (15):

[0140] b encoded =G(b G ; b G ) (15)

[0141] Wherein, G(·; b G ) is a parameterized generator function, and the parameter is b real .

[0142] Further, the embodiment of the present application calculates the similarity of the preset radiation data by the discriminator according to the similarity measurement algorithm to obtain the evaluation result. Wherein, the discriminator in the embodiment of the present application realizes more accurate measurement of the similarity between data by the similarity measurement method when evaluating the generated data and the real data, so as to provide more effective gradient feedback to the generator. For example, the embodiment of the present application calculates the similarity between the generated data and the real data by the measurement method based on quantum computing As shown in the following formula (16):

[0143]

[0144] Wherein, b represents the real radiation data, is an improved similarity measurement function.

[0145] It should be noted that the similarity measurement function in the embodiment of the present application is embodied by calculating the Euclidean distance between the generated data and the real data, and the embodiment of the present application takes the Euclidean distance as an example, and the calculation method is as shown in the following formula (17):

[0146]

[0147] Wherein, b and b respectively represent the i-th element of the generated data and the real data vector.

[0148] Correspondingly, the generator and the discriminator in the embodiments of the present application are trained in a loop iteration under the framework of confrontation, and the generator is continuously optimized to generate more accurate data, while the recognition ability of the discriminator is improved, until the generated data is difficult to distinguish from the real data. For example, in the process of loop iteration, the discriminator D and the generator G in the embodiments of the present application are optimized through the first loss function, as shown in the following formula (18):

[0149]

[0150] Actually, D() in the formula is a discriminator function, which can output the probability of data being real data. is the generator loss, is the discriminator loss.

[0151] In the training process, the embodiments of the present application periodically evaluate the performance of the model, and specifically evaluate the quality and diversity of the generated data. Then, based on the evaluation results, the training strategy is further adjusted, such as adjusting the quantum state encoding parameters, optimizing the similarity measurement method, etc., to ensure that the model can effectively perform data augmentation. For example, the performance evaluation of the model in the embodiments of the present application is performed by calculating the distance between the generated data and the real data, as shown in the following formula (19):

[0152] Δ=dist(b gen ,b real ) (19)

[0153] Wherein, dist(,) in the formula is the Euclidean distance measurement function, which is used to evaluate the quality and diversity of the generated data.

[0154] Finally, the embodiments of the present application adjust the parameters of the initial generative adversarial network model according to the evaluation results to obtain a preset generative adversarial network model. Among them, based on the evaluation results, the embodiments of the present application adjust the parameters of the generator and the discriminator, as shown in the following formula (20):

[0155]

[0156] Wherein, η is the learning rate, and are the gradients of the first loss function of the generator and the discriminator, respectively. In one embodiment, the learning rate η is set to 0.001.

[0157] At the same time, in the embodiments of the present application, the gradient update of the parameter b G of the generator G is specified using the chain rule, as shown in the following formula (21):

[0158]

[0159] Wherein, in the formula denotes the loss a partial derivative of the generator output with respect to is a partial derivative of the generator output with respect to its parameters.

[0160] In addition, in the embodiments of the present application, the parameter b of the discriminator D is updated in the following manner: D The gradient update is shown in the following formula (22):

[0161]

[0162] wherein, in the formula is the loss a partial derivative of the discriminator output with respect to is a partial derivative of the discriminator output with respect to its parameters.

[0163] It is easy to understand that, in order to ensure the quality and diversity of the generated data, the model needs to be trained for a sufficient number of iterations, for example, the number of training iterations is set to 10,000.

[0164] In some embodiments of the present application, the preset monitoring model is trained through a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model, including but not limited to the following steps:

[0165] A center model and a local model are constructed. The local model includes a plurality of extreme learning machine models, and the plurality of extreme learning machine models correspond to the third radiation data.

[0166] The corresponding extreme learning machine models are trained respectively through the third radiation data to obtain local update parameters.

[0167] The local update parameters are transmitted to the center model to perform aggregated updating on the center model to obtain target update parameters.

[0168] The target update parameters are respectively transmitted to each extreme learning machine model to perform parameter updating on the extreme learning machine model to obtain a target radiation monitoring model.

[0169] In the embodiment, the embodiment first constructs a center model and a local model. Specifically, the local model in the embodiment refers to a model independently trained on a local data set by each client, and each client is provided with a corresponding local model. Correspondingly, the center model refers to a model shared by all participants (clients). In the embodiment, a local model is first constructed in each client, and a center model is constructed in the corresponding center server. Correspondingly, the local model of each client in the embodiment includes an extreme learning machine model, and each extreme learning machine model corresponds to the third radiation data one by one. For example, each client collects first radiation data through a corresponding radiation monitoring sensor, and obtains corresponding third radiation data through data dimension reduction and data augmentation. Then, the embodiment trains the corresponding extreme learning machine model through the third radiation data to obtain local update parameters, and then transmits the local update parameters to the center model to update the center model to obtain target update parameters. Specifically, the embodiment inputs the third radiation data corresponding to each client into the extreme learning machine model to train the model and obtain the local update parameters of the local model after training. Then, the embodiment transmits the local update parameters corresponding to each client to the center model to update the center model. As shown in FIG. 8, in the embodiment, the local model is trained on the data of each client, and the local model and the center model interact, that is, the update of the model parameters flows. Figure 3 fed θ fed represents the model parameters updated after local training, and θ ​​The exchange between the local model and the center model is in a manner allowing the center model to aggregate updates and allowing the local model to receive new, aggregated parameters, i.e., target update parameters. Further, embodiments of the present application transmit the target update parameters to each extreme learning machine model respectively to perform parameter update on the extreme learning machine model to obtain a target radiation monitoring model. Specifically, embodiments of the present application train the local model to obtain local update parameters through multiple iterations and send the local update parameters to the center model to perform global update, and then send the updated parameters (target update parameters) back to the local model to perform parameter update on the local model to obtain the target radiation monitoring model. In embodiments of the present application, through the federated learning architecture, the privacy of data can be protected, and in the training process, the original data is not shared between clients or with the central server, but only model parameters are exchanged for update, i.e., the center model synchronizes the parameters to each client by assigning θ fed to each client, effectively mitigating the problem of data privacy risk.

[0170] In some embodiments of the present application, the corresponding extreme learning machine model is trained by the third radiation data to obtain local update parameters, including but not limited to the following steps:

[0171] The third radiation data is input into the extreme learning machine model to perform nonlinear mapping to obtain intermediate output data.

[0172] The target output data is calculated according to the intermediate output data and the weight matrix.

[0173] The loss data is calculated according to the target output data through a second loss function to determine the weight update data through the loss data. The second loss function includes a gradient penalty coefficient.

[0174] The extreme learning machine model is updated by the back propagation algorithm according to the weight update data to obtain the local update parameters.

[0175] In the embodiment, the embodiment of the application first inputs the third radiation data into the extreme learning machine model to perform nonlinear mapping to obtain intermediate output data, and calculates target output data according to the intermediate output data and a weight matrix. Specifically, the local model of each client in the embodiment of the application adopts an extreme learning machine algorithm based on gradient penalty as a radiation level monitoring model. Accordingly, the extreme learning machine in the embodiment of the application includes an input layer, a hidden layer and an output layer, the input layer receives collected data, the hidden layer performs nonlinear conversion on features using an activation function, and is connected to the output layer through a weight matrix, and the output data of the output layer, that is, the target output data, is calculated through the weight matrix and the intermediate output data. Illustratively, the embodiment of the application first initializes the parameters of each extreme learning machine model, the initialization mode is random initialization, and the initialized parameters are subject to a normal distribution with a mean of 0 and a variance of a unit matrix. Then, the input layer of the extreme learning machine model in the embodiment of the application receives input data X r , and the main task of the hidden layer is to perform nonlinear mapping on the input data X r to obtain an intermediate output H u , that is, intermediate output data. For example, the weight of the hidden layer of the extreme learning machine model in the embodiment of the application is W u , and the bias is b u , and the hidden layer output is represented as shown in the following formula (23):

[0176] H u =Re(W u X u +b u )⊙A u (23)

[0177] Wherein, H u is the output of the hidden layer; Re() is a ReLU activation function; is an element multiplication; A u is a local sensitivity mapping factor.

[0178] Correspondingly, the local sensitivity mapping factor in the embodiment of the application enables the model to adaptively enhance the attention to local patterns, avoids the global features from dominating the learning process of the model, thereby improving the classification accuracy, and the calculation formula is shown in the following formula (24):

[0179]

[0180] Wherein, F u (H u ) is a local response function on the output of the hidden layer, such as a local feature response of local maximum (i.e. max pooling); λ uee is a coefficient for adjusting the local sensitivity, controlling the degree of influence of local features on activation. For example, λuee Set to 0.01.

[0181] Further, in this embodiment of the invention, loss data is calculated using the second loss function based on the target output data. Weight update data is then determined based on the loss, and the parameters of the extreme learning machine model are updated using the backpropagation algorithm based on the weight update data to obtain the local update parameters. Specifically, in this embodiment of the invention, the second loss function includes a gradient penalty coefficient. During forward propagation, each gradient calculated in this embodiment of the invention requires gradient penalty. The gradient penalty ensures that the model's gradient remains within a reasonable range during each parameter update, thus mitigating training instability caused by excessively large or small gradients. For example, in this embodiment of the invention, a gradient penalty term (gradient penalty coefficient) is used in the first loss function to constrain the first loss function, ensuring that each parameter update maintains the stability and effectiveness of the gradient. The calculation formula is shown in equation (25) below:

[0182]

[0183] Wherein, λ u β is the first gradient penalty coefficient; u γ is the second gradient penalty coefficient; u This is the third gradient penalty coefficient; This is a gradient penalty term; This represents the gradient of the hidden layer output of the Extreme Learning Machine with respect to the weights.

[0184] For example, after the nonlinear transformation of the hidden layer, the final result is passed to the output layer of the classifier. In the output layer of the extreme learning machine, the input signal is mapped by the Softmax function to generate the classification result, i.e., the target output data, as shown in the following equation (26):

[0185]

[0186] Where, in the formula

[0187] The classification results are from the Extreme Learning Machine; W u b is the weight matrix of the output layer; u For the bias term of the output layer; H u The output of the hidden layer is θ; Sof() is the Softmax function, which maps a linear combination of the inputs to a probability distribution, ultimately yielding the probability of each class; u The coefficient used to control the strength of the quadratic term; (H) u ) 2The square term of the hidden layer output is used to enhance the fitting ability of the classification model to complex data patterns, especially in high-dimensional radiation monitoring data, and to better capture the nonlinear characteristics of the data.

[0188] Then, the first loss function of the extreme learning machine model is calculated, which includes a constant error calculation term and a gradient penalty term. The gradient penalty term is used to make the classifier better learn the deep features in the data during each optimization process and reduce possible bias in the training process. The calculation formula is shown in the following formula (27):

[0189]

[0190] Wherein, in the formula is the first loss function of the extreme learning machine model; is the first loss function of the cross-entropy, which is used to measure the difference between the predicted value and the true label; is the gradient penalty term; λ u is the gradient penalty coefficient, which controls the contribution of the gradient penalty to the first loss function; is the local feature enhancement term; A u (i) is the i-th element of the local sensitivity mapping factor; μ uev is the weight coefficient of the local feature enhancement term, which is used to control the influence of the enhancement term on the final loss. For example, in the embodiment of the present application, μ uev is set to 0.2.

[0191] Next, the back propagation and parameter update of the extreme learning machine model are performed. The calculated error will update the weight parameters of the network through the back propagation algorithm, and the parameters of the model are updated through the back propagation algorithm. The gradient calculation adopts the chain rule. The parameter update method of the extreme learning machine model is shown in the following formula (28):

[0192]

[0193] Wherein, in the formula ΔW u is the weight update amount of the extreme learning machine model; Δb u is the bias update amount of the extreme learning machine model; is the gradient of the first loss function of the extreme learning machine model with respect to the weight; W u (t) and b u (t) are the weight and bias of the t-th iteration, respectively; W u (t+1) and b u (t+1) are the weight and bias of the t+1-th iteration, respectively.

[0194] It should be noted that the embodiment of the present application repeats the above steps until the preset stopping iteration condition is met, that is, the model training is completed. Accordingly, the preset stopping iteration condition in the embodiment of the present application is to reach the preset maximum iteration number, for example, the preset maximum iteration number is set to 1000 times. Accordingly, by continuously updating the weight, the extreme learning machine model can continuously optimize in the training process, and finally achieve the optimal classification effect. Wherein, after the extreme learning machine model training in the embodiment of the present application is completed, the radiation level monitoring result is the classification result of the extreme learning machine For example, in the embodiment of the present application, the classification categories include: "low radiation", "medium radiation" and "high radiation", a total of 3 categories.

[0195] In some embodiments of the present application, after performing parameter updating on the extreme learning machine model according to the weight update data through the back propagation algorithm to obtain the local update parameter, the radiation level monitoring method provided by the embodiment of the present application further includes but is not limited to the following steps:

[0196] According to the weight update data, the intermediate output data and the basic learning rate, a second expected learning rate is calculated.

[0197] According to the second expected learning rate, the model training learning rate is adjusted.

[0198] In this specific embodiment, the embodiment of the present application adjusts the learning rate adaptively at each weight update, so that the model can converge faster. Wherein, the embodiment of the present application calculates a second expected learning rate according to the weight update data, the intermediate output data and the basic learning rate, and adjusts the model training learning rate according to the second expected learning rate. Specifically, the basic learning rate in the embodiment of the present application refers to the basic learning rate of the extreme learning machine model, and the weight update data refers to the change amount of the weight. Accordingly, the weight update of the extreme learning machine model in the embodiment of the present application follows an adaptive convergence strategy, which automatically adjusts the learning rate according to the change of the current gradient. For example, when the model training is relatively stable, the learning rate will automatically decrease, thereby speeding up the convergence process, or when the model has a large gradient change, the learning rate will be appropriately increased to help the model overcome the local optimal solution and further improve the classification accuracy. Wherein, the learning rate calculation formula of the extreme learning machine model in the embodiment of the present application is shown in the following formula (29):

[0199]

[0200] Wherein, in the formula, η u (t+1) is the learning rate of the extreme learning machine model of the t+1 iteration; η base is the basic learning rate of the extreme learning machine model; γ u is the first adjustment coefficient of the learning rate of the extreme learning machine model, and δ uis a second adjustment coefficient of the learning rate of the extreme learning machine model; AW u (t) is a change amount of the weight of the tth iteration; H u (t) is an output of the hidden layer of the tth iteration. For example, in the embodiment of the present application, γ u is set to 0.2, δ u is set to 0.3.

[0201] It is easy to understand that when processing the radiation level monitoring data, the model often faces the problem of high-dimensional data, and the gradient disappearance phenomenon is easy to occur. Therefore, in the embodiment of the present application, the adaptive gradient disappearance alleviation strategy is adopted, that is, the training step is adjusted according to the change of the current gradient, which effectively alleviates the gradient disappearance problem, makes the training process more stable, and the adjustment mode is as shown in the following formula (30):

[0202]

[0203] wherein, in the formula, η' is a learning rate of the extreme learning machine model of the t+1th iteration after adjustment; λ u is a learning rate of the extreme learning machine model of the t+1th iteration after adjustment; λ uas is a third adjustment coefficient of the learning rate of the extreme learning machine model, μ uas is a fourth adjustment coefficient of the learning rate of the extreme learning machine model; α uas is a fifth adjustment coefficient of the learning rate of the extreme learning machine model; β uas is a sixth adjustment coefficient of the learning rate of the extreme learning machine model. For example, in the embodiment of the present application, λ uas is set to 0.2, μ uas is set to 0.2, α uas is set to 0.1, and β uas is set to 0.1.

[0204] It is easy to understand that the embodiments of the present application jointly train multiple dispersed client models without exchanging local data by adopting a federated learning architecture, thereby ensuring data privacy while fully utilizing distributed data for model optimization. Compared with traditional centralized learning methods, the problem of data privacy leakage caused by uploading sensitive radiation monitoring data to a central server is effectively alleviated through federated learning, which has significant privacy protection advantages. Meanwhile, for the high-dimensional features that may exist in the radiation monitoring data, the embodiments of the present application use a boundary smoothing-based autoencoder for feature dimension reduction. The encoder-decoder structure of the autoencoder compresses high-dimensional data into low-dimensional representation, and the smoothing processing alleviates the influence of boundary effects and outliers, enhances the robustness of the model to noise, effectively alleviates the problem of high dimensionality of data, and improves the stability and accuracy of data processing. In addition, the embodiments of the present application alleviate the problem of insufficient training samples by using a quantum state-based generative adversarial network. The quantum state encoding uses the advantages of quantum computing to simulate more complex radiation data distribution and generate diversified high-quality synthetic data, effectively expanding the size of the training data set and improving the generalization ability of the model. Especially in the case of sample scarcity, the accuracy of the model can be significantly improved. Accordingly, the traditional extreme learning machine is prone to gradient vanishing or explosion problems during training, leading to unstable training. In order to alleviate this problem, the embodiments of the present application use a gradient penalty term in the first loss function of the extreme learning machine, which normalizes the gradient range and ensures the stability of the training process, so that the model can maintain higher training efficiency and accuracy when processing complex radiation data. At the same time, the extreme learning machine in the embodiments of the present application also combines adaptive learning rate and local sensitivity mapping mechanism, so that the model can dynamically adjust the learning rate according to the change of the current gradient, thereby speeding up the convergence process and alleviating the problem of unstable training caused by excessively large or small gradient. In addition, the local sensitivity mapping can enhance the model's attention to local patterns, avoiding the dominance of global features, thereby improving the classification accuracy.

[0205] Please refer to Figure 4 The embodiments of the present application also provide a radiation level monitoring system, which can implement the above-mentioned radiation level monitoring method. The system comprises:

[0206] The first module 210 is configured to collect first radiation data through a preset radiation monitoring sensor.

[0207] The second module 220 is configured to perform data dimension reduction processing on the first radiation data through a feature dimension reduction model to obtain second radiation data. The feature dimension reduction model comprises a boundary smoothing-based autoencoder.

[0208] The third module 230 is configured to perform data augmentation on the second radiation data by using a preset generative adversarial network model to obtain third radiation data. The preset generative adversarial network model includes a generative adversarial network model based on a quantum state.

[0209] The fourth module 240 is configured to perform model training on a preset monitoring model by using a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model.

[0210] The fifth module 250 is configured to input the to-be-analyzed radiation data into the target radiation monitoring model to perform radiation level analysis and obtain a radiation level monitoring result.

[0211] It can be understood that the content in the above method embodiments is applicable to the present system embodiments, the present system embodiments specifically implement the functions of the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0212] Embodiments of the present application also provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor implements the above radiation level monitoring method when executing the computer program. The electronic device can be any intelligent terminal, such as a tablet computer or a vehicle-mounted computer.

[0213] It can be understood that the content in the above method embodiments is applicable to the present device embodiments, the present device embodiments specifically implement the functions of the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0214] Please refer to Figure 5 , Figure 5 The hardware structure of the electronic device of another embodiment is illustrated, and the electronic device includes:

[0215] The processor 310 can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, and is configured to execute related programs to implement the technical solutions provided by the embodiments of the present application.

[0216] The memory 320 can be implemented in the form of a read only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 320 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present specification are implemented by software or firmware, the related program codes are stored in the memory 320 and are called and executed by the processor 310 to implement the radiation level monitoring method of the embodiments of the present application.

[0217] The input / output interface 330 is configured to realize information input and output.

[0218] The communication interface 340 is configured to realize the communication interaction between the device and other devices. The communication can be realized by a wired manner (for example, a USB, a network cable, etc.) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, etc.).

[0219] The bus 350 is configured to transmit information between various components (for example, the processor 310, the memory 320, the input / output interface 330, and the communication interface 340) of the device.

[0220] The processor 310, the memory 320, the input / output interface 330, and the communication interface 340 are connected to each other through the bus 350 to realize the communication connection between the device.

[0221] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the above-mentioned radiation level monitoring method.

[0222] It can be understood that the contents in the above-mentioned method embodiments are applicable to the present storage medium embodiments. The present storage medium embodiments specifically implement the same functions as the above-mentioned method embodiments, and achieve the same beneficial effects as the above-mentioned method embodiments.

[0223] The memory is a non-transitory computer readable storage medium, which can be used to store non-transitory software programs and non-transitory computer executable programs. In addition, the memory can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory remotely arranged relative to the processor. These remote memories can be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0224] The embodiments described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0225] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than shown in the figures, or combine certain steps, or different steps.

[0226] The device embodiments described above are merely illustrative, and units described as separate components can or can not be physically separated, that is, can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0227] Those skilled in the art can understand that all or some of the steps in the above disclosed method, the functions of the modules / units in the system and the device can be implemented as software, firmware, hardware and their appropriate combinations.

[0228] The terms "first", "second", "third", "fourth" and the like (if any) in the specification of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0229] It should be understood that, in the application, "at least one" refers to one or more, and "multiple" refers to two or more. "And / or" is used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, "A and / or B" can represent three cases of only A, only B and A and B existing at the same time, wherein A and B can be singular or plural. The character " / " generally represents an "or" relationship between the front and rear associated objects. "At least one of the following" or the like means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can represent a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b and c can be single or multiple.

[0230] In several embodiments provided in the application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the above units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed mutual units can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0231] The units described above as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0232] In addition, each functional unit in each embodiment of the application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0233] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in other words, the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program storage media.

[0234] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A method of monitoring a level of radiation, characterized by, The method comprises the following steps: acquiring first radiation data through a preset radiation monitoring sensor; performing data dimension reduction processing on the first radiation data through a feature dimension reduction model to obtain second radiation data; wherein the feature dimension reduction model comprises a boundary smoothing-based autoencoder; performing data augmentation on the second radiation data through a preset generative adversarial network model to obtain third radiation data; wherein the preset generative adversarial network model comprises a quantum state-based generative adversarial network model; training a preset monitoring model through a federated learning architecture according to the third radiation data to obtain a target radiation monitoring model; inputting to-be-analyzed radiation data into the target radiation monitoring model for radiation level analysis to obtain a radiation level monitoring result; wherein the training of the preset monitoring model through the federated learning architecture according to the third radiation data to obtain the target radiation monitoring model comprises: constructing a center model and a local model; wherein the local model comprises a plurality of extreme learning machine models, and the plurality of extreme learning machine models correspond to the third radiation data; training the corresponding extreme learning machine models through the third radiation data to obtain local update parameters; transmitting the local update parameters to the center model to perform aggregated updating on the center model to obtain target update parameters; transmitting the target update parameters to each of the extreme learning machine models to perform parameter updating on the extreme learning machine models to obtain the target radiation monitoring model; wherein the training of the corresponding extreme learning machine models through the third radiation data to obtain the local update parameters comprises: inputting the third radiation data into the extreme learning machine model for nonlinear mapping to obtain intermediate output data; calculating target output data according to the intermediate output data and a weight matrix; calculating loss data through a second loss function according to the target output data to determine weight update data through the loss data; wherein the second loss function comprises a gradient penalty coefficient; performing parameter updating on the extreme learning machine model through a backpropagation algorithm according to the weight update data to obtain the local update parameters; wherein the hidden layer output representation of the extreme learning machine model comprises a local sensitivity mapping factor, as shown in the following formula: ； wherein, in the formula is the output of the hidden layer, is a local response function on the output of the hidden layer, is a coefficient for adjusting the local sensitivity; wherein the calculation formula of the gradient penalty coefficient is as shown in the following formula: ； wherein, in the formula is a first gradient penalty coefficient; is a second gradient penalty coefficient; is a third gradient penalty coefficient; is a weight of the hidden layer, is a bias of the hidden layer; is a gradient penalty term; is a gradient of the extreme learning machine hidden layer output with respect to the weight.

2. The method of claim 1, wherein, Before performing the data dimension reduction processing on the first radiation data through the feature dimension reduction model to obtain the second radiation data, the method further comprises: constructing an initial autoencoder; inputting preset input data into the initial autoencoder to generate reconstruction data through forward propagation; performing smoothing processing on the reconstruction data through an automatic boundary smoothing algorithm to obtain smoothed output data; calculating a weight coefficient through a preset variance function according to the preset input data; calculating a first loss function according to the smoothed output data, the preset input data, and the weight coefficient; Updating parameters of the initial auto-encoder according to the first loss function by a gradient descent algorithm to obtain the feature dimension reduction model.

3. The method of claim 2, wherein, The updating parameters of the initial auto-encoder according to the first loss function by a gradient descent algorithm to obtain the feature dimension reduction model comprises: calculating a first expected learning rate by a convex hull convergence algorithm; calculating a parameter update amount according to the first expected learning rate and the first loss function; updating parameters of the initial auto-encoder according to the parameter update amount to obtain the feature dimension reduction model.

4. The method of claim 1, wherein, The data augmentation of the second radiation data by the preset generative adversarial network model comprises: constructing an initial generative adversarial network model; wherein the initial generative adversarial network model comprises a generator and a discriminator; receiving random noise by the generator to convert the random noise into quantum state data by a quantum state encoding algorithm; generating preset radiation data according to the quantum state data; calculating the similarity of the preset radiation data by the discriminator according to a similarity measurement algorithm to obtain an evaluation result; adjusting parameters of the initial generative adversarial network model according to the evaluation result to obtain the preset generative adversarial network model.

5. The method of claim 1, wherein, After performing the parameter updating of the extreme learning machine model according to the weight update data by the back propagation algorithm to obtain the local update parameter, the method further comprises: calculating a second expected learning rate according to the weight update data, the intermediate output data and a basic learning rate; adjusting the model training learning rate according to the second expected learning rate.

6. A radiation level monitoring system characterized by, The system comprises: a first module configured to collect first radiation data by a preset radiation monitoring sensor; a second module configured to perform data dimension reduction processing on the first radiation data by a feature dimension reduction model to obtain second radiation data; wherein the feature dimension reduction model comprises a boundary smoothing-based auto-encoder; a third module configured to perform data augmentation on the second radiation data by a preset generative adversarial network model to obtain third radiation data; wherein the preset generative adversarial network model comprises a quantum state-based generative adversarial network model; a fourth module configured to perform model training on a preset monitoring model according to the third radiation data by a federated learning architecture to obtain a target radiation monitoring model; a fifth module configured to input to-be-analyzed radiation data into the target radiation monitoring model for radiation level analysis to obtain a radiation level monitoring result; The model training of the preset monitoring model according to the third radiation data by the federated learning architecture to obtain the target radiation monitoring model comprises: constructing a center model and a local model; wherein the local model comprises a plurality of extreme learning machine models, and the plurality of extreme learning machine models correspond to the third radiation data; training the corresponding extreme learning machine models respectively by the third radiation data to obtain local update parameters; transmitting the local update parameters to the center model to perform aggregated updating on the center model to obtain target update parameters; and transmitting the target update parameters to the target radiation monitoring model to perform model training on the target radiation monitoring model. transmit the target update parameters to each of the extreme learning machine models respectively to perform parameter update on the extreme learning machine models, and obtain the target radiation monitoring model; wherein the training of the corresponding extreme learning machine model by the third radiation data to obtain local update parameters comprises: inputting the third radiation data into the extreme learning machine model to perform nonlinear mapping, and obtaining intermediate output data; calculating target output data according to the intermediate output data and a weight matrix; calculating loss data by a second loss function according to the target output data, so as to determine weight update data by the loss data; wherein the second loss function comprises a gradient penalty coefficient; performing parameter update on the extreme learning machine model by a back propagation algorithm according to the weight update data, and obtaining the local update parameters; wherein the hidden layer output representation of the extreme learning machine model comprises a local sensitivity mapping factor, and is expressed as follows: ； wherein in the formula is the output of the hidden layer, is a local response function on the output of the hidden layer, is a coefficient for adjusting the local sensitivity. wherein the calculation formula of the gradient penalty coefficient is expressed as follows: ； wherein, in the formula is a first gradient penalty coefficient; is a second gradient penalty coefficient; is a third gradient penalty coefficient; is a weight of the hidden layer, is a bias of the hidden layer; is a gradient penalty term; is a gradient of the extreme learning machine hidden layer output with respect to the weight.

7. An electronic device, comprising: comprise: at least one processor; at least one memory for storing at least one program; when the at least one program is executed by the at least one processor, the at least one processor implements the method in any one of claims 1-5.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium comprising: The computer program is executed by the processor to implement the method in any one of claims 1 to 5.

Images