A method for characterizing the environmental risks of museum collections

Abstract

This invention provides a method for characterizing the risks of the preservation environment for museum collections, comprising the following steps: S1, constructing a risk characterization system and acquiring information on the museum collections and the environment; S2, acquiring environmental variables and environmental risk indicators, and identifying key environmental variables; S3, constructing an environmental risk status classification model and acquiring preset classification control limits; S4, acquiring several real-time environmental variables and determining the classification control limits at the current moment; S5, comparing the current classification control limits with the preset classification control limits to determine the current risk status classification and generate a risk control QR code; S6, displaying the generated QR code on a museum collection terminal device. Through the above steps, this invention can quantify and classify the risks of the preservation environment for museum collections and characterize them quickly and intuitively, enabling managers to promptly and fully grasp the risk status of the preservation environment, facilitating timely risk management, and improving the risk prevention, control, and protection capabilities for museum collections.

Description

Technical Field

[0001] This invention relates to the field of risk characterization technology for the preservation environment of cultural relics, and more specifically to a method for characterizing the risk of the preservation environment of museum collections. Background Technology

[0002] Cultural relics are typically preserved in specific environments, and environmental factors affect their preservation time and quality. For museum collections, long-term storage in enclosed environments exposes them to numerous environmental changes. Therefore, minimizing or mitigating the damaging effects of these environmental factors is crucial for the preventative protection of precious cultural relics.

[0003] Preventing or mitigating environmental damage to museum collections requires monitoring and assessing their preservation environment. However, current monitoring and assessment methods rely on physical and chemical experiments and the experience of cultural relic experts for judgment. This approach is time-consuming, labor-intensive, highly subjective, and fails to characterize the risks of the preservation environment, thus hindering effective risk prevention and protection of museum collections. Summary of the Invention

[0004] To address the problems in the prior art, this invention provides a method for characterizing the risks of the preservation environment of museum collections. This method can quickly and intuitively characterize the risks of the preservation environment of museum collections and improve the ability to prevent, control, and protect risks to museum collections.

[0005] This invention provides a method for characterizing the environmental risks of museum collections, including:

[0006] Step S1: Provide a server, multiple cultural relic terminal devices connected to the server, and a risk control system, and use the cultural relic terminal devices to obtain information on the collection of cultural relics and environmental information;

[0007] Step S2: The information on the cultural relics and the environmental information are input to the risk control system through the server. The system analyzes the environmental variables and environmental risk indicators of the cultural relics in the display case corresponding to each cultural relic terminal device, and determines the key environmental variables of different types of cultural relics.

[0008] Step S3: Based on the environmental risk indicators and key environmental variables of different types of cultural relics, construct an environmental risk status classification model for different types of cultural relics and obtain the preset classification control limits for the environmental risks of different types of cultural relics.

[0009] Step S4: Use the cultural relic terminal device to obtain several real-time environmental variables of the cultural relics in the collection, and obtain the classification control limit at the current moment based on the several real-time environmental variables and the environmental risk status classification model constructed in step S3.

[0010] Step S5: Compare the current level control limit with the preset level control limit to determine the risk status level of the current cultural relic preservation environment and generate the corresponding risk control QR code.

[0011] Step S6: The generated risk control QR code is input into the cultural relic terminal device through the server, and the risk control QR code is displayed on the display interface of the cultural relic terminal device.

[0012] Furthermore, the information on the museum's collection includes the identity information of the artifacts, the conditions under which they are preserved, the management and use of the artifacts, and visitor information.

[0013] Furthermore, the environmental information includes information on cultural relic protection facilities and information on the environment in which cultural relics are preserved.

[0014] Furthermore, the terminal device for cultural relics is a mobile phone, tablet computer, desktop computer, or portable laptop computer.

[0015] Furthermore, step S2 further includes:

[0016] Step S21: Construct a model for the set of environmental variables X and a model for the environmental risk indicator y;

[0017] Step S22: For each environment variable x in the set of environment variables i And standardize the environmental risk indicator y;

[0018] Step S23: Match each standardized environmental variable with the standardized environmental risk indicator. Perform mutual information calculations, sort the results from largest to smallest, and select the top six environmental variables as key environmental variables.

[0019] Furthermore, the model for the set of environmental variables X and the model for the environmental risk index y satisfy the following relationship:

[0020] X = [x1, x2, ..., x i ,…,x m ]∈R n×m

[0021] y∈R n

[0022] Where m is the total number of environmental variables, n is the number of samples obtained from sampling the preservation environment of the artifacts in the collection, i is the ordinal number of all environmental variables, and x i Let i represent the i-th environmental variable in the set of environmental variables, where i = 1, 2, 3, ..., m, and R is the set of real numbers.

[0023] Furthermore, the sampled values ​​of the standardized environmental variables satisfy the following relationship:

[0024]

[0025]

[0026] Where n is the number of samples obtained by sampling the preservation environment of the museum's collection. This represents the t-th sampled value of the i-th environmental variable after standardization. Represents the i-th environment variable x i The mean of all sampled values, s i Let x represent the standard deviation of all sampled values ​​of the i-th environmental variable. i Let x represent the i-th environment variable in the set of environment variables. it Represents the i-th environment variable x i The t-th sample value;

[0027] The standardized environmental variable consists of all sampled values ​​of the i-th environmental variable.

[0028] Furthermore, the i-th standardized environmental variable With standardized environmental risk indicators The mutual information satisfies the following relation:

[0029]

[0030] in, Represents the marginal entropy. Represents conditional entropy. express and The joint probability density, express marginal probability density, express The marginal probability density.

[0031] Furthermore, step S3 further includes:

[0032] Step S31: Construct a training model based on a time-series correlation neural network as an environmental risk status classification model;

[0033] Step S32: Construct a training set X′ for key environmental variables and a training set y′ for environmental risk indicators. For each key environmental variable x′ in the training set... j Standardize the environmental risk indicator y′;

[0034] Step S33: Standardize the key environmental variable x′ jThe environmental risk indicator y′ is input into the training model constructed in step S31 for training, and the key environmental variable x′ is continuously used. j By fitting the risk index y′, the neural network learns the key environmental variable x′. j The relationship with the risk indicator y′ is used to extract the key environmental variable x′. j Data feature h;

[0035] Step S34: Based on the data characteristics h of the key environmental variable X′, derive the preset graded control limit T for environmental risk. 2 .

[0036] Furthermore, the preset hierarchical control limit T 2 The following relationship must be satisfied:

[0037]

[0038]

[0039] Where N represents the number of samples, h i This represents the features obtained for each sample point during the offline process. This represents the average value of a characteristic.

[0040] Furthermore, step S4 further includes:

[0041] Step S41: Select key real-time environment variables from the acquired real-time environment variables;

[0042] Step S42: Standardize the real-time key environmental variables;

[0043] Step S43: Input the standardized real-time key environmental variables into the environmental risk status classification model constructed in step S3 to obtain the classification control limits at the current moment.

[0044] Furthermore, tiered control limits The following relationship must be satisfied:

[0045] T fact 2 =h(X) online ) T Σ -1 h(X online )

[0046]

[0047] Where h(X) online ) represents the features of each sample point during the online process.

[0048] Through the above steps, this invention can quantify and classify the risks of the preservation environment of museum collections and quickly and intuitively characterize them, enabling managers to grasp the risk status of the preservation environment of museum collections in a timely and comprehensive manner, which is conducive to promoting timely risk management and improving the risk prevention and protection capabilities of museum collections. Attached Figure Description

[0049] Figure 1 This is a flowchart of the method for characterizing the environmental risks of museum artifacts according to the present invention.

[0050] Figure 2 This is a block diagram of the structure of a time-dependent neural network. Detailed Implementation

[0051] The preferred embodiments of the present invention are given below with reference to the accompanying drawings and described in detail.

[0052] like Figure 1 As shown, the method for characterizing the environmental risks of museum artifacts according to the present invention includes the following steps:

[0053] Step S1, as follows Figure 2 As shown, the system includes a server 1, multiple cultural relic terminal devices 2 connected to the server 1, and a risk control system 3. Each cultural relic terminal device 2 is located in a fixed display case containing museum artifacts. The terminal device 2 can access information about the artifacts within its corresponding display case and the environmental information of the artifact's surrounding environment. The artifact information includes artifact identification information (artifact code, artifact material, major damage, collecting institution, etc.), artifact preservation conditions information (preservation location, display status, display conditions, etc.), artifact management and usage information (maintenance management, inspection management, personnel, system management, guidance services, etc.), and visitor information. The environmental information includes information on artifact protection facilities (environmental monitoring equipment, microenvironment control equipment, etc.) and artifact preservation environment information (humidity and heat indicators, light intensity, pollutants, biological hazards, climate indicators, etc.). The cultural relic terminal devices 2 can be mobile phones, tablets, desktop computers, laptops, etc., and are equipped with QR code-based applications.

[0054] Step S2: The information on the cultural relics and the environment obtained by the cultural relic terminal device 2 is input into the risk control system 3 through the server 1. The environmental variables and environmental risk indicators of the cultural relics in the display case corresponding to each cultural relic terminal device 2 are analyzed to determine the key environmental variables of different types of cultural relics.

[0055] This invention employs mutual information analysis to extract key environmental variables. Mutual information between two variables is a measure of their interdependence, gauging the correlation between two sets of events. It can also be understood as measuring the information shared between two variables, i.e., the degree to which understanding one variable reduces the uncertainty about the other. The steps for extracting key environmental variables include:

[0056] Step S21, construct the model for the set of environmental variables X and the model for the environmental risk indicator y as follows:

[0057] X = [x1, x2, ..., x i ,…,x m ]∈R n×m (1)

[0058] y∈R n (2)

[0059] Where m is the total number of environmental variables, n is the number of samples obtained by sampling the preservation environment of the museum's collection, and R is the set of real numbers. i Let i represent the i-th environmental variable in the set of environmental variables, where i is the ordinal number of all environmental variables, i = 1, 2, 3, ..., m.

[0060] Because different variables have different measures (units), they will have different magnitudes. If the raw data is used, larger-scale data will occupy a larger proportion in the model, which is obviously unreasonable. Therefore, it is necessary to standardize each variable in the above model to ensure that different variables contribute equally to the overall model.

[0061] Therefore, in step S22, for each environment variable x i And the environmental risk indicator y is standardized. This invention uses the standard deviation standardization method, and each environmental variable x is standardized. i The sampled value x it The standardized calculation formula is as follows:

[0062]

[0063]

[0064] Where n is the number of samples obtained by sampling the preservation environment of the museum's collection, representing... Represents the standardized i-th environment variable x i The t-th sample value, Let s represent the mean of all sampled values ​​of the i-th environmental variable, and let s represent the standard deviation of all sampled values ​​of the i-th environmental variable. i Let x represent the i-th environment variable in the set of environment variables. tRepresents the i-th environment variable x i The t-th sample value.

[0065] The standardized i-th environment variable x i All sampled values ​​constitute the standardized environmental variables.

[0066] The standardized calculation formula for the environmental risk indicator y is the same as that for environmental variables, and will not be repeated here.

[0067] Step S23: Match each standardized environmental variable with the standardized environmental risk indicator. Perform mutual information calculations, sort the results from largest to smallest, and select the top six environmental variables as key environmental variables.

[0068] The i-th standardized environmental variable With standardized environmental risk indicators The formula for calculating mutual information is as follows:

[0069]

[0070] in, Represents the marginal entropy. This represents conditional entropy. express and The joint probability density, express marginal probability density, express The marginal probability density.

[0071] By calculating mutual information, the correlation between environmental variables and environmental risk indicators can be visually observed, providing a theoretical basis for the selection of key environmental variables. It should be noted that when selecting key environmental variables, multiple variables can be chosen in order of their numerical values, or the environmental variables whose calculated results exceed a certain threshold can be selected. In this embodiment, the final selected key environmental variables are the six ranked variables (i.e., the top six), including ambient temperature, ambient humidity, light intensity, sulfur dioxide concentration, carbon dioxide concentration, and volatile organic compounds.

[0072] Step S3: Based on the environmental risk indicators and key environmental variables of different types of cultural relics, construct an environmental risk status classification model for different types of cultural relics, and obtain the preset classification control limits for the environmental risks of different types of cultural relics.

[0073] Step S3 specifically includes:

[0074] Step S31: Construct a training model based on a temporally correlated neural network (BERT) as an environmental risk state classification model. Since the state represented by a single sampling point differs between different preceding and following sampling points, a single sampling point cannot fully reflect the current state. That is, throughout the process, a single sampling point does not exist independently but has a temporal relationship with several preceding and following sampling points; therefore, temporally correlated features need to be introduced into the neural network.

[0075] The time-related neural network used in this invention is as follows: Figure 2 As shown, it includes a multi-head attention mechanism module, a residual connection module, a feedforward neural network layer, a max-pooling layer, a linear layer, and an output layer. For better understanding, the following is a brief description of this temporally correlated neural network.

[0076] 1) The multi-head attention mechanism module is configured to perform the following operations: For an input vector, firstly, construct the query vector (Q), key vector (K), and value vector (V) of the vector through a linear layer; where Q∈R N×m , K∈R N×m V∈R N ×m Secondly, to train the network to pay attention to and learn different types of information, and to avoid the network learning only one type of information from affecting the results, a multi-head attention mechanism is adopted. This involves linearly projecting the query vector (Q), key vector (K), and value vector (V) onto dimensions different from the input dimensions before computation, as shown below:

[0077] head i =Attention(QW i Q ,KW i K VW i V (6)

[0078] Among them, head i It is a multi-head vector segment, W i Q ∈R m×d W i K ∈R m×d With W i V ∈R m×d Here are the parameters obtained from network training, and d = m / i. The calculation method for Attention() is as follows:

[0079]

[0080] Attention_output is the output value after the attention operation, and Softmax is the normalization function.

[0081] After learning different information, all vector fragments are concatenated to ensure that the overall information content remains unchanged, as shown below:

[0082] MultiHead(Q,K,V)=Concat(head1...head h W o (8)

[0083] Among them, W o ∈R m×m These are weights obtained from network training.

[0084] 2) The residual connection module is configured to perform the following operations: add the input result and output result of each layer and then perform a normalization operation to solve the network degradation problem to a certain extent. Equation (9) below represents adding the input to the input after passing through the neural network, and Equation (10) below represents performing layer normalization on the result. Here, layer normalization is used, that is, normalization calculation is performed in each layer to prevent the input data from falling into the saturation region of the activation function.

[0085]

[0086]

[0087] in, This represents the unnormalized values ​​of the features in the l-th layer. This represents the normalized numerical value of the features at layer l. This represents the input of the current layer of the network. This represents the output of the current layer of the network, where ε is a small constant; and

[0088]

[0089] Where H represents the number of neurons in the l-th layer.

[0090] Furthermore, to prevent information corruption during LayerNormalization, the result needs to be processed by an activation function. Therefore, the final result after LayerNormalization is shown below:

[0091]

[0092] Where f(·) represents the activation function, and here the ReLU activation function is used.

[0093] 3) The feedforward neural network layer consists of two fully connected layers. The first layer uses the SELU activation function, while the second layer does not use an activation function. The corresponding formula for this layer is as follows:

[0094]

[0095] Where λ and α are fixed parameters, λ≈1.0507 and α≈1.6733. W1, W2, b1, and b2 are parameters obtained during network training. X l This is the output of the residual network in the last layer.

[0096] 4) The max pooling layer is set to increase the receptive field while reducing the number of parameters, and select the most relevant input from it.

[0097] Step S32: Construct a training set X′ for key environmental variables and a training set y′ for environmental risk indicators. For each key environmental variable x′ in the training set... j Standardize the environmental risk indicator y′.

[0098] The training set X′ for key environmental variables and the training set y′ for environmental risk indicators are obtained by sampling the information on cultural relics and the environment acquired by the cultural relic terminal device 2, and then statistically analyzing them.

[0099] The training sets X′ for key environmental variables and y′ for environmental risk indicators are shown below:

[0100] X′=[x′1,x′2,…,x′ j ,…x′ n ]∈R t×n (13)

[0101] y'∈R t (14)

[0102] Where n is the number of key environmental variables; t is the time window length, i.e., the number of samples selected. In this embodiment, the time window length is 10, meaning 10 samples are input for network training each time; R is the set of real numbers; x j This represents any environment variable in the training set that is a key environment variable.

[0103] For each key environment variable x′ j The method for standardizing the environmental risk indicator y′ is the same as in step S12, and will not be repeated here. It should be noted that for each key environmental variable x′... j The mean and variance values ​​used when standardizing the environmental risk indicator y′ are the same as those calculated by formula (4).

[0104] Step S33, standardize the key environmental variable x′ j The environmental risk indicator y′ is input into the training model constructed in step S31 for training, and the key environmental variable x′ is continuously used. j By fitting the risk index y′, the neural network learns the key environmental variable x′. j The relationship with the risk indicator y′ is used to extract the key environmental variable x′. j The data feature h.

[0105] Specifically, first define the loss function MSE:

[0106]

[0107] Among them, y j These are the given key environment variables. This is the model's output value. The network is trained by minimizing this loss function, MSE. After training, the output of the layer preceding the classification layer of the neural network is used as the key environmental variable x′. j The data feature h.

[0108] Step S34: Based on the data characteristics h of the key environmental variable X′, derive the preset graded control limit T for environmental risk. 2 Preset graded control limit T 2 The calculation formula is as follows:

[0109]

[0110]

[0111] In the formula, N represents the number of samples, h i This represents the data features obtained for each sample point during the offline process. This represents the average value of the data characteristics.

[0112] In the formula, N represents the number of samples, which is equal to t in formulas (13) and (14).

[0113] Since a kernel density estimate is required to select the kernel to use for each sampling point, a kernel density estimate is needed. 2 At what confidence level should the value be determined? Specifically, first determine the confidence level of the kernel density estimation, and then select the corresponding T value based on the confidence level. 2 Value. For example, when the confidence level is 0.9, the corresponding preset control limit is the medium-risk preset control limit. When the confidence level is 0.99, the corresponding preset control limit is the high-risk preset control limit.

[0114] Step S4: Obtain several real-time environmental variables of the cultural relics in the collection through the cultural relic terminal device 2. Based on the several real-time environmental variables and the environmental risk status classification model constructed in step S3, obtain the classification control limit at the current moment. Step S4 specifically includes:

[0115] Step S41: Select key real-time environmental variables from the acquired real-time environmental variables, that is, select the key environmental variables (i.e., key real-time environmental variables) determined in step S1 from the acquired real-time environmental variables, and use them as data for subsequent calculations.

[0116] Step S42: Standardize the real-time key environmental variables. The standardization method here is the same as that in step S12, and will not be repeated here. It should be noted that the mean and variance values ​​used in this standardization are the same as those calculated by formula (4) to ensure the consistency between real-time data and modeling data.

[0117] Step S43: Input the standardized real-time key environmental variables into the environmental risk status classification model constructed in step S3 to obtain the classification control limits at the current moment.

[0118]

[0119]

[0120] Where h(X) online ) represents the features of each sample point during the online process. h i Same as formulas 15 and 16.

[0121] Step S5, set the current hierarchical control limit. With preset graded control limits By comparing data, the risk status classification of the current preservation environment of the museum's collection is determined, and a corresponding risk control QR code is generated.

[0122] Specifically, when When the risk level is low, a green QR code is generated, indicating that the preservation environment of the museum's collection is normal; when At that time, the risk level is medium risk, and a yellow QR code is generated, indicating that there are potential risks to the preservation environment of the museum's collection; when At that time, the risk level is high, and a red QR code is generated, indicating that the environment in which the museum's collection is preserved is dangerous.

[0123] Step S6: The generated QR code is input into the cultural relic terminal device 2 via server 1, and the QR code is displayed on the display interface of the cultural relic terminal device 2. In this way, by quantifying, classifying, and characterizing the risks of the preservation environment of museum collections, managers can promptly and fully grasp the risk status of the preservation environment, which is conducive to promoting timely risk management and improving the risk prevention and protection capabilities of museum collections.

[0124] The present invention also provides a storage medium for characterizing the environmental risks of museum artifacts, the storage medium storing computer-executable instructions. These computer-executable instructions, when executed, perform the steps S1-S6 described above.

[0125] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of the invention. Various variations can be made to the above embodiments of the present invention. That is, all simple and equivalent changes and modifications made based on the claims and description of this invention fall within the protection scope of the claims of this patent. All aspects not described in detail in this invention are conventional technical content.

Claims

1. A method of characterizing the risk of the preservation environment of a collection of cultural heritage objects, characterized in that, The application comprises the following steps: Step S1, providing a server, a plurality of cultural relic terminal devices connected in communication with the server, and a risk control system, obtaining collection cultural relic information and environment information by using the cultural relic terminal devices; Step S2, inputting the collection cultural relic information and the environment information to the risk control system through the server, analyzing the environmental variables and environmental risk indicators in the storage environment of the collection cultural relics in the display cabinet corresponding to each cultural relic terminal device, and determining the key environmental variables of different types of cultural relics; The application comprises the following steps: Step S21: constructing a model of the environmental variable set X and a model of the environmental risk indicator y as follows: ， ， wherein, is the number of all environmental variables, is the number of samples obtained by sampling the preservation environment of the cultural relics in the collection, is the ordinal number of all environmental variables, denotes the i-th environmental variable in the set of environmental variables, , is the set of real numbers; Step S22: standardizing each environmental variable in the environmental variable set and the environmental risk index y, the sampling value of the standardized environmental variable satisfies the following relationship: ， ， wherein n is the number of samples obtained by sampling the preservation environment of the collection cultural relics, denotes the t-th sample value of the i-th environmental variable after standardization, denotes the mean of all sample values of the i-th environmental variable , denotes the standard deviation of all sample values of the i-th environmental variable , denotes the i-th environmental variable in the set of environmental variables, denotes the t-th sample value of the i-th environmental variable . All the sampled values of the i-th environment variable after standardization constitute the standardized environment variable ; Step S23: comparing each standardized environmental variable with the standardized environmental risk indicator The mutual information is calculated, the calculation results are sorted in descending order of numerical value, and the environmental variables ranked in the top six are taken as the key environmental variables; the standardized environmental variable of the i th The mutual information of the standardized environmental risk indicator satisfies the following relationship: I( )=H -H = , ) , where H represents the marginal entropy of X represents the conditional entropy, , ) represents the joint probability density of X and Y represents the marginal probability density of X represents the marginal probability density of Y represents the marginal probability density of X represents the marginal probability density of Y Step S3, constructing an environmental risk state classification model of different types of cultural relics according to the environmental risk indicators and key environmental variables of different types of cultural relics, and obtaining a preset classification control limit of the environmental risk of different types of cultural relics; Step S3 further comprises: Step S31: constructing a training model based on a time series correlation neural network as an environmental risk state classification model; The time series correlation neural network comprises a multi-head attention mechanism module, a residual connection module, a feedforward neural network layer, a max pooling layer, a linear layer, and an output layer; The multi-head attention mechanism module is configured to construct a query vector, a key vector, and a value vector of the input vector, perform linear projection on the query vector, the key vector, and the value vector to different dimensions from the input dimension, calculate a multi-head vector segment, and splice all the multi-head vector segments; The residual connection module is configured to add the input result and the output result of each layer and perform a normalization operation; the feedforward neural network layer is composed of two fully connected layers, the first layer uses an activation function SELU, and the second layer does not use an activation function; the max pooling layer is configured to increase the receptive field while reducing the number of parameters and select the most relevant input; Step S32: Constructing the training set X of key environmental variables with the training set y of environmental risk indicators , for each key environmental variable in the training set and the environmental risk indicator y is normalized; Step S33: standardizing the key environmental variables and the environmental risk index y The training model constructed in the input step S31 is trained, and the key environmental variables and the risk index y are constantly utilized to make the neural network learn the relationship between the key environmental variables and the risk index y , so as to extract the data features h of the key environmental variables ; the training of the network is completed by minimizing the loss function; and the output result of the previous layer of the output layer of the neural network is used as the data features h of the key environmental variables after the training is completed. Step S34, deriving preset classification control limits of environmental risk according to data features h of key environmental variables ;​ The preset hierarchical control limit satisfies the following relationship: ， ， N represents the number of samples, represents the features obtained for each sample point in the offline process, represents the average value of the features; when the confidence level is 0.9, the corresponding preset classification control limit is the medium-risk preset classification control limit ; when the confidence level is 0.99, the corresponding preset classification control limit is the high-risk preset classification control limit ; Step S4, obtaining a plurality of real-time environmental variables of the collection cultural relics by using the cultural relic terminal devices, and obtaining a classification control limit at the current time according to the plurality of real-time environmental variables and the environmental risk state classification model constructed in step S3; Step S5, compare the current time with the preset hierarchical control limit , , , when , the risk level is low risk, a green two-dimensional code is generated, indicating that the collection cultural relics preservation environment is normal; when , the risk level is medium risk, a yellow two-dimensional code is generated, indicating that the collection cultural relics preservation environment has hidden dangers; when , the risk level is high risk, a red two-dimensional code is generated, indicating that the collection cultural relics preservation environment is dangerous. hierarchy control limit at the current time satisfies the following relation: ， ， wherein, represents a feature representing each sample point in the online process; Step S6, inputting the generated risk control two-dimensional code to the cultural relic terminal device through the server, and displaying the risk control two-dimensional code on the display interface of the cultural relic terminal device.

2. The method of claim 1, wherein, The collection cultural relic information comprises cultural relic identity information, cultural relic storage condition information, cultural relic management and use information, and visitor information.

3. The method of claim 1, wherein, The environmental information comprises cultural relic protection facility information and cultural relic storage environment information.

4. The method of claim 1, wherein, The cultural relic terminal device is a mobile phone, a tablet computer, a desktop computer, or a portable notebook computer.

5. The method of claim 1, wherein, Step S4 further comprises: Step S41: selecting real-time key environmental variables from the obtained plurality of real-time environmental variables; Step S42: standardizing the real-time key environmental variables; Step S43: input the normalized real-time key environmental variables into the environmental risk state classification model constructed in step S3 to obtain the classification control limit at the current time .

Images