Hash and feature selection-based software and hardware equipment fingerprint fine-grained generation method

Abstract

The invention discloses a Hash and feature selection-based software and hardware equipment fingerprint fine-grained generation method, and belongs to the technical field of power network software and hardware equipment fingerprint biomarker generation. Based on the information entropy and the self-correlation coefficient, the features used for generating the device fingerprints are screened out from the numerous features, representative and stable features are extracted from the device fingerprints, and selected feature parameters are converted into hash codes used for the device fingerprint generation and matching step; different devices are identified and distinguished hierarchically and accurately through a fine-grained device fingerprint generation step based on feature selection, a hash code and a Bayesian network.

Description

Technical Field

[0001] The invention belongs to the technical field of power network software and hardware device fingerprint generation, and relates to a software and hardware device fingerprint fine-grained generation method based on hash and feature selection. Background Art

[0002] The power Internet is a key network infrastructure that meets the development needs of intelligent software and hardware for power grids, with the characteristics of low latency, high reliability and wide coverage. It is an emerging business model and application model formed by the deep integration of the new generation of information and communication technology and the power industry. By building technical platforms such as the Internet of Things, cloud computing, and big data, the power Internet realizes the interconnection of power grid equipment, data sharing, intelligent operation and maintenance, and smart electricity use, providing all-round information support and intelligent services for the power industry network. The power Internet identification system is an important part of its network system and a spiritual hub supporting the interconnection of the power Internet. The core of its identification system lies in the generation of fingerprints of hardware and software devices in the power environment. Device fingerprints refer to identity symbols that can uniquely identify physical assets such as machines and products and virtual resources such as systems and software. Traditional identification methods are prone to duplicate identification codes, and the meaning of identification codes is publicly available. It is easy for people to guess that the coding method is counterfeited and forged, which may cause serious system intrusion and even huge losses. Summary of the invention

[0003] The present invention aims at solving the problems in the prior art and provides a method for generating fine-grained fingerprints of hardware and software devices based on hash and feature selection.

[0004] A fine-grained generation method for software and hardware device fingerprints based on hash and feature selection includes the following steps: based on information entropy and autocorrelation coefficient, the features used to generate device fingerprints are screened out from a large number of features, representative and stable features are extracted from the device fingerprints, and the selected feature parameters are converted into hash codes for device fingerprint generation and matching steps. Through the fine-grained device fingerprint generation steps based on feature selection, hash codes and Bayesian networks, different devices can be accurately identified and distinguished in a hierarchical manner.

[0005] A method for generating fine-grained fingerprints of hardware and software devices based on hashing and feature selection also includes the following steps:

[0006] Step S1: When the system receives a fingerprint generation request, it calls the data interface or the backend database according to the request type.

[0007] Step S2: Extract different feature data for different modules, and perform algorithm analysis on the extracted feature data to filter out redundant and invalid attribute features and segment coarse-grained attributes.

[0008] Step S3: Perform hash operation on the extracted attribute features to generate a fine-grained fingerprint of the device.

[0009] Step S4: The fingerprint is transmitted to the Bayesian network constructed by historical data for classification and the result is fed back to the server.

[0010] Step S5: The fingerprints of different modules of the device are combined in the order of chip, installed module and operating system to generate a device fingerprint and add a corresponding verification code to it.

[0011] The advantages of the present invention are that it combines feature selection and hashing steps to achieve more efficient and accurate device identification and classification.

[0012] Improved accuracy: This method improves the accuracy of device identification by filtering out the most representative and discriminative features from device fingerprints through a feature selection step.

[0013] Reduced computational cost: Using the hashing step, the selected features can be converted into compact hash codes, which reduces storage and computational costs. Compared with traditional device fingerprint generation methods, this method is more efficient in maintaining device databases and performing real-time identification.

[0014] Enhanced security: With fine-grained device fingerprinting, fraud and unauthorized access can be better prevented. More accurate device identification helps improve the system's ability to identify trusted devices, thereby enhancing security.

[0015] Strong adaptability: Due to the use of feature selection and hashing technology, this method has strong adaptability in different environments and scenarios. Whether on mobile devices, in network communications or in the Internet of Things environment, it can be effectively applied and deployed.

[0016] Reduce the false recognition rate: Through the carefully designed feature selection and hashing technology, the false recognition rate can be effectively reduced, and the stability and reliability of the system can be improved. This means that the system can identify devices more reliably and reduce the negative impact of false positives.

[0017] The proposed feature screening fingerprint generation scheme can effectively extract features of different categories of devices. The features of each type of device are representative. In addition, the diversity of hash functions makes the generated fingerprints more random, making it difficult to find patterns and thus not easy to be imitated.

[0018] The present invention can integrate the inherent attributes and behavioral characteristics of a device to generate an identification of the device, and can facilitate the management and update of the device fingerprint and the access authentication of the device.

[0019] The device fingerprint generation scheme uses a hash function to avoid duplicate device fingerprints in a system and ensure the uniqueness of the device identification.

[0020] Record the attribute elements of the generated fingerprint to facilitate subsequent monitoring and analysis of device behavior so as to update the device fingerprint in a timely manner.

[0021] By inferring the relationship between attributes based on the changes in the original fingerprint data and building a Bayesian network, it is possible to classify devices after the device fingerprint is generated and distinguish whether the device is connected to the network for the first time.

[0022] The present invention proposes a fine-grained generation method for hardware and software device fingerprints based on hashing and feature selection, which realizes efficient and accurate device identification through feature selection and hashing technology.

[0023] First, the feature selection step uses information entropy and autocorrelation coefficient to select the most representative and distinguishing features from the device fingerprint. This process can effectively filter out redundant and invalid attributes, ensure that the selected features can accurately reflect the uniqueness of the device, and improve the accuracy of device identification. In this way, a small number of features that are most useful for device identification can be selected from a large number of features, reducing the complexity and computational complexity of data processing and improving the overall performance of the system.

[0024] Secondly, the selected features are converted into compact hash codes, which significantly reduces storage and computing costs. Compared with traditional device fingerprint generation methods, this method is more efficient in maintaining device databases and performing real-time identification. The use of hash codes not only makes the data more compact, effectively reduces the storage space occupied, but also speeds up data processing and transmission. At the same time, the generated fine-grained device fingerprints can better prevent fraud and unauthorized access, and enhance the security of the system. Through fine-grained device fingerprints, every detail of the device can be accurately identified, preventing any disguise and counterfeiting, and ensuring the reliability and security of device identification.

[0025] Finally, by classifying device fingerprints through the Bayesian network, the category of the device and the uniqueness of the fingerprint can be accurately determined, further improving the intelligence and automation level of device management. The Bayesian network uses historical data and empirical knowledge to build a powerful classification model that can quickly and accurately classify devices after the device fingerprint is generated, and determine whether the device is a first-time network device. In addition, the attribute elements of the generated fingerprint are recorded to facilitate the subsequent monitoring and analysis of device behavior, so as to update the device fingerprint in a timely manner and ensure the accuracy and timeliness of device identification. This mechanism of continuous monitoring and dynamic updating enables the system to respond to changes quickly and improves the flexibility and adaptability of device management.

[0026] This fine-grained generation method of hardware and software device fingerprints based on hashing and feature selection not only solves the problems of duplicate identification codes and easy counterfeiting in traditional device fingerprint generation methods, but also significantly improves the accuracy and security of device identification. Its superior performance in adaptability, efficiency and security makes this method have broad application prospects and can be effectively applied to power network hardware and software scenarios to provide comprehensive and reliable equipment identification and management solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. As shown in the figure:

[0028] Figure 1 The present invention is a flowchart of the method for generating fingerprints of power grid hardware and software devices.

[0029] Figure 2 The figure is a flow chart of the feature selection algorithm in the power grid hardware and software device fingerprint generation method of the present invention.

[0030] Figure 3 The present invention is a flowchart of the power grid hardware and software equipment fingerprint Bayesian network application.

[0031] Figure 4 It is a parameter relationship diagram of the power grid software and hardware device fingerprint Bayesian network application of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] Example 1: Figure 1 As shown, a fine-grained method for generating fingerprints of software and hardware devices based on hashing and feature selection can generate a fingerprint identification of power equipment with coding uniqueness, and at the same time can integrate the characteristics of different levels of modules of power grid software and hardware devices to generate a device fingerprint with coded meaning encryption, thereby improving the security of the power grid software and hardware equipment system.

[0034] The following steps are involved:

[0035] Step S1: When the system receives a fingerprint generation request, it calls the data interface or the backend database according to the request type.

[0036] Step S2: Extract different feature data for different modules, and perform algorithm analysis on the extracted feature data to filter out redundant and invalid attribute features and segment coarse-grained attributes.

[0037] Step S3: Perform hash operation on the extracted attribute features to generate a fine-grained fingerprint of the device.

[0038] Step S4: The fingerprint is transmitted to the Bayesian network constructed by historical data for classification and the result is fed back to the server.

[0039] Step S5: The fingerprints of different modules of the device are combined in the order of chip, installed module and operating system to generate a device fingerprint and add a corresponding verification code to it.

[0040] like Figure 2 As shown, the data preprocessing steps include: providing the original data set with the Onehot encoding categorical data step and the missing value filling step, and collecting feature screening data set step.

[0041] The feature screening steps include: feature discrimination and feature stability steps, feature importance scoring steps, and differential screening steps.

[0042] The device fingerprint generation steps include: post-screening feature step, device fingerprint generation + verification code generation step.

[0043] like Figure 3 As shown, in the initial step, the power equipment module features are extracted and sent to the server, a fine-grained device fingerprint is generated for the module, the generated fingerprint is compared with the existing fingerprint library, a feature vector is generated, and input into the Bayesian network classification to distinguish the equipment category and new and old equipment, and the server feedbacks the results.

[0044] like Figure 4 As shown, the chip is connected to the module, size, supplier, function and power consumption, the module connection address, communication password, port number and communication baud rate, the operating system is connected to the system version, network traffic and kernel parameters, and the above connections are judged and then output.

[0045] Example 2: Figure 1 , Figure 2 , Figure 3 and Figure 4 As shown, a method for generating fingerprints of hardware and software devices based on hashing and feature selection includes the following steps:

[0046] Step S1: When the system receives a fingerprint generation request, it calls the data interface or the backend database according to the request type.

[0047] Step S2: Extract different feature data for different modules, and perform algorithm analysis on the extracted feature data to filter out redundant and invalid attribute features and segment coarse-grained attributes.

[0048] Step S3: Perform hash operation on the extracted attribute features to generate a fine-grained fingerprint of the device.

[0049] Step S4: The fingerprint is transmitted to the Bayesian network constructed by historical data for classification and the result is fed back to the server.

[0050] Step S5: The fingerprints of different modules of the device are combined in the order of chip, installed module and operating system to generate a device fingerprint and add a corresponding verification code to it.

[0051] In step S1, when a coding request is received, data is retrieved from a database or an API interface through a corresponding data coding interface.

[0052] In step S2, the coarse-grained attributes of modules at different levels of the device are segmented, and the complex attribute features are pruned and segmented to achieve feature fine-graining.

[0053] In step S3, the data in different modules of the device are hashed to generate a fine-grained fingerprint of the device.

[0054] In step S2, the method for feature screening to select the features used to generate device fingerprints is: first, preprocess the acquired data, supplement the missing values ​​of some features, and encode the categorical data; then, calculate the discrimination and stability of the features respectively and perform weighted sorting to determine the features used to generate device fingerprints.

[0055] The method for calculating the distinguishability of features is to evaluate the amount of information of each feature in the data set through information entropy. The larger the information entropy, the more information the feature contains, and the more uncertain and difficult to distinguish the data; conversely, the smaller the information entropy, the more certain and easy to distinguish the data of the feature, and the less distinguishable it is.

[0056] The calculation method is Where X is the characteristics of the hardware and software equipment in the power environment, x i is the feature parameter; the final value is used to identify the degree of discrimination of the feature.

[0057] The method for calculating the stability of features is to evaluate the stability of each feature field in the data set through the autocorrelation coefficient (ACF). If a feature field has high stability, then its value changes less at different times, and it can more accurately reflect the inherent characteristics of the device, thereby improving the uniqueness and robustness of the device fingerprint. When the absolute value of the ACF coefficient is greater than 0.8, it means that the data has a strong correlation and may have a certain volatility; when the absolute value of the ACF coefficient is less than 0.2, it means that the data has almost no correlation and the changes are stable. The calculation method is

[0058]

[0059] The final value is used to indicate the stability of the feature.

[0060] Let X t is the characteristics of hardware and software equipment in the power environment, X t-k The features after translation, The method of feature selection based on the average value of a set of feature parameters and the discrimination and stability of the comprehensive features is: Stability and discrimination are often contradictory in feature selection. Some features with high stability may contribute less to the discrimination between different categories, while some features with high discrimination may not be stable enough and are easily affected by noise and interference. This method is used to balance between stability and discrimination and select the optimal feature subset. The formula is: feature w =weight·H w -(1-weight)·|ACF w |Where weight is the weight of discrimination and stability. Generally speaking, the weight of discrimination is greater than the weight of stability; H w and ACF w are the information entropy and ACF coefficient of the wth feature respectively. The feature with the highest final score will be selected as the fingerprint generation factor.

[0061] Finally, the feature method used to generate device fingerprints is determined as follows: differential screening. Specifically, in order to maintain generality, the attribute features and behavior features are first arranged from large to small. i The calculation formula is: i =feature i+1 -feature i Among them, feature i+1 Represents the feature of the i+1th feature, feature iThe feature of the i-th feature. The maximum score means that the feature score drops sharply, which means the end position of the feature selection.

[0062] In this way, the most important features can be filtered out, while those features that have less impact on the device fingerprint can be ignored.

[0063] In step S3, the hash algorithm for generating the device fingerprint includes the hash algorithm including SHA-1, SHA-256, SHA-384, and SHA-512 algorithms.

[0064] In step S4, the Bayesian network structure is first constructed using historical data and empirical knowledge.

[0065] In step S4, the training data is preprocessed and the parameters of the Bayesian network are learned.

[0066] In step S4, the fingerprint is transmitted to the Bayesian network model trained based on historical data to determine the type of the device and whether the fingerprint of the device is generated for the first time, and the determination result is transmitted to the server.

[0067] In step S5, after the fingerprints of different modules of the device are combined, a check code is added to the suffix, and the check code algorithm is a parity check algorithm, a Hamming check algorithm or a cyclic redundancy check algorithm.

[0068] A check code is generated using a predetermined check code algorithm.

[0069] When selecting the real scenes of electric meters and terminals in the power grid, the present invention simulates and generates the corresponding simulation data of electric meters and terminals respectively. In order to accurately monitor and analyze the operation status of the power grid, according to the real scene, in order to screen the features that need to be extracted by the modules, the following feature attributes are selected as the screened indicators: communication protocol, communication mode, module ID, manufacturer, connection mode, rated current, rated voltage and other attribute elements.

[0070] First, the data in the dataset is preprocessed to supplement the missing values ​​of some features and encode the categorical data.

[0071] In order to reduce the impact of encoding on the algorithm, the encoded data is normalized to the interval [0,1] to facilitate subsequent data stability analysis.

[0072] The information entropy algorithm is used to measure the discrimination of data in data preprocessing. Let X be the characteristics of hardware and software equipment in the power environment, x i is the characteristic parameter, and the discrimination measure of the data using the information entropy algorithm is:

[0073]

[0074] Algorithm: Information entropy algorithm. It is used to evaluate the discrimination of data. Based on the frequency distribution of the data, it determines the diversity and entropy of the data by calculating the probability distribution of different values ​​in the data set, thereby calculating the uncertainty and confusion of the attribute data.

[0075] Input: D: A dataset containing n objects, each object has m-dimensional attribute parameters.

[0076] Output: Discrimination evaluation index.

[0077] method:

[0078] By using the information entropy algorithm to analyze the data of n objects in each attribute for each feature in the m dimension, the results are finally sorted in descending order.

[0079] Considering the convenience of expressing this indicator in reality and the rationality of using it in subsequent algorithms, it was finally decided to store the obtained discrimination evaluation indicators in descending order. By sorting the evaluation indicators in descending order, it is easy to identify and extract the data or features with the highest discrimination, which facilitates the subsequent algorithms and analysis.

[0080] The results are shown in Table 1:

[0081] Table 1

[0082]

[0083]

[0084] The stability measure of the data uses the autocorrelation coefficient (ACF) algorithm. Let X t is the characteristics of hardware and software equipment in the power environment, X t-k The features after translation, is the average value of a set of characteristic parameters, and the stability measure of the data using the autocorrelation coefficient (ACF) algorithm is:

[0085]

[0086] Input: D: A dataset containing n objects, each object has m-dimensional attribute parameters, and k is the data translation amount.

[0087] Output: stability assessment indicators.

[0088] method:

[0089] 1) According to the amount of data in the data set, set the corresponding k value.

[0090] 2) By using the information entropy algorithm to analyze the data of n objects in each attribute for each feature in the m dimension, the results are finally sorted in descending order.

[0091] Through comprehensive comparison of simulated data, 3 is selected as the k data offset, and the obtained discrimination evaluation index is stored in ascending order. The index ascending storage can quickly identify the parameters with better stability as a reference.

[0092] The results are shown in Table 2 below:

[0093] Table 2

[0094]

[0095]

[0096] Based on the calculated discrimination and stability, the two evaluation criteria are weighted and integrated. The specific formula is shown as follows:

[0097] feature w =weight·H w -(1-weight)·|ACF w |

[0098] Where weight is the weight, H w and ACF w are the information entropy and ACF coefficient of the w-th feature respectively.

[0099] Through comprehensive comparison of simulated data, the discrimination of data is a more important consideration for generating device fingerprints, and the weight is set to 0.6; at the same time, the calculated comprehensive evaluation scores will be stored in descending order.

[0100] The specific results are shown in Table 3 below:

[0101] Table 3

[0102]

[0103]

[0104] The parameters used to generate fingerprints are finally extracted through the differential algorithm and the selected attributes are stored in the database when the parameters are generated, which is convenient for the subsequent monitoring of the device itself and the device modules.

[0105] By differentiating the fine-grained features screened based on device hierarchies, the topological structure of the Bayesian network is constructed, such as Figure 4 shown.

[0106] Then the parameters of the model are learned according to the training data to determine the conditional probability distribution of each node.

[0107] Considering that the collected training data set is not necessarily complete and it is impossible to obtain a more accurate prior probability, the maximum likelihood estimation method is used for parameter learning.

[0108] The feature vector of the post-fingerprint change and the device fingerprint vector are respectively input into two trained Bayesian network models for testing, and finally the type of device and whether the device is a new device are determined by a pre-set threshold. Finally, the server will feedback the judgment result.

[0109] After the device category is determined and the fingerprints of each module of the device are generated successfully, the fine-grained fingerprints of each module of the device are merged in the order of chip, installed module and operating system, and the device fingerprint is generated and suffixed with the corresponding verification code. Then it is stored in the database of the corresponding category of devices.

[0110] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating fine-grained fingerprints of hardware and software devices based on hashing and feature selection, characterized in that: Based on information entropy and autocorrelation coefficient, features used to generate device fingerprints are screened out from numerous features, representative and stable features are extracted from the device fingerprints, and the selected feature parameters are converted into hash codes for device fingerprint generation and matching steps. Through the fine-grained device fingerprint generation steps based on feature selection, hash codes and Bayesian networks, different devices can be accurately identified and distinguished hierarchically.

2. According to claim 1, a method for generating fine-grained software and hardware device fingerprints based on hashing and feature selection is characterized in that: It also contains the following steps: Step S1: When the system receives a fingerprint generation request, it calls the data interface or the backend database according to the request type; Step S2: extract different feature data for different modules, and perform algorithm analysis on the extracted feature data to filter out redundant and invalid attribute features and segment coarse-grained attributes; Step S3: Perform hash operation on the extracted attribute features to generate a fine-grained fingerprint of the device; Step S4: transmitting the fingerprint to the Bayesian network constructed by historical data for classification and feeding the result back to the server; Step S5: The fingerprints of different modules of the device are combined in the order of chip, installed module and operating system to generate a device fingerprint and add a corresponding verification code.

3. According to claim 2, a method for generating fine-grained fingerprints of hardware and software devices based on hashing and feature selection is characterized in that: Step S1 also includes the following steps: when receiving a coding request, data is retrieved from the database or API interface through the corresponding data coding interface. Step S2 also includes the following steps: segmenting the coarse-grained attributes of modules at different levels of the device, pruning and segmenting the complex attribute features to achieve fine-grained features, Step S3 also includes the following steps: hashing the data in different modules of the device to generate a fine-grained fingerprint of the device. Step S4 also includes the following steps: first, the Bayesian network structure is constructed using historical data and empirical knowledge.

4. According to claim 2, a method for generating fine-grained fingerprints of hardware and software devices based on hashing and feature selection is characterized in that: Step S2 also includes the following steps: the method for selecting the features for generating device fingerprints by feature screening is: firstly, preprocessing the acquired data, supplementing the missing values ​​of some features, and encoding the categorical data; then respectively calculating the discrimination and stability of the features and weighted sorting them to determine the features for generating device fingerprints. The method for calculating the distinguishability of features is: the information entropy is used to evaluate the amount of information of each feature in the data set. The larger the information entropy, the more information the feature contains, which means the data is more uncertain and difficult to distinguish. On the contrary, the smaller the information entropy, the more certain and easy to distinguish the data of the feature is, and it does not have a high degree of distinguishability. The calculation method is Where X is the characteristics of the hardware and software equipment in the power environment, x i is the feature parameter; the final value is used to identify the degree of discrimination of the feature. The method for calculating the stability of features is: the stability of each feature field in the data set is evaluated by the autocorrelation coefficient (ACF). A feature field has high stability and its value changes less at different times. It can more accurately reflect the inherent characteristics of the device, thereby improving the uniqueness and robustness of the device fingerprint. When the absolute value of the ACF coefficient is greater than 0.8, it means that the data has a strong correlation and may have a certain volatility; when the absolute value of the ACF coefficient is less than 0.2, it means that the data has almost no correlation and the change is stable. The calculation method is The final value is used to identify the stability of the feature. The method of comprehensive feature discrimination and stability for feature screening is: stability and discrimination are often contradictory in feature selection. Some features with higher stability may contribute less to the discrimination between different categories, while some features with higher discrimination may not be stable enough and are easily affected by noise and interference. It is used to balance between stability and discrimination and select the optimal feature subset. The formula is: feature w =weight·H w -(1-weight)·|ACF w |Where weight is the weight of discrimination and stability, it is assumed that the weight of discrimination is greater than the weight of stability; H w and ACF w are the information entropy and ACF coefficient of the wth feature respectively. The feature with the highest final score will be selected as the fingerprint generation factor. Finally, the specific feature method for generating device fingerprints is determined as follows: Differential screening In order to avoid loss of generality, the attribute features and behavior features are first arranged from large to small, and the difference k of the i-th feature is i The calculation formula is: i =feature i+1 -feature i Among them, feature i+1 Represents the feature of the i+1th feature, feature i The feature representing the i-th feature has the largest score, which means that the feature score drops sharply, that is, the end position of the feature is selected. In this way, the most important features are screened out and those features that have little impact on the device fingerprint are ignored.

5. According to claim 2, a method for generating fine-grained fingerprints of hardware and software devices based on hashing and feature selection is characterized in that: Step S3 also includes the following steps: the hash algorithm for generating the device fingerprint includes the hash algorithm including SHA-1, SHA-256, SHA-384, and SHA-512 algorithms.

6. According to claim 2, a method for generating fine-grained software and hardware device fingerprints based on hashing and feature selection is characterized in that: Step S4 also includes the following steps: preprocessing the training data and learning the parameters of the Bayesian network.

7. According to claim 2, a method for generating fine-grained software and hardware device fingerprints based on hashing and feature selection is characterized in that: Step S4 also includes the following steps: transmitting the fingerprint to a Bayesian network model trained based on historical data and determining the type of the device and whether the fingerprint of the device is generated for the first time, and transmitting the determination result to the server.

8. According to claim 2, a method for generating fine-grained software and hardware device fingerprints based on hashing and feature selection is characterized in that: Step S5 also includes the following steps: after the fingerprints of different modules of the device are combined, a check code is added to the suffix, and the check code algorithm is a parity check algorithm, a Hamming check algorithm or a cyclic redundancy check algorithm.

9. According to claim 4, a method for generating fine-grained software and hardware device fingerprints based on hashing and feature selection is characterized in that: The information entropy algorithm is used to measure the discrimination of data in data preprocessing. Let X be the characteristics of hardware and software equipment in the power environment, x i is the characteristic parameter, and the discrimination measure of the data using the information entropy algorithm is: Algorithm: The information entropy algorithm is used to evaluate the discrimination of data. Based on the frequency distribution of the data, the diversity and entropy of the data are determined by calculating the probability distribution of different values ​​in the data set, thereby calculating the uncertainty and confusion of the attribute data. Input: D: A dataset containing n objects, each object has m-dimensional attribute parameters, Output: Discrimination evaluation index, method: By using the information entropy algorithm to analyze the data of n objects in each attribute for each feature in the m dimension, the results are finally sorted in descending order. Considering the convenience of expressing this indicator in reality and the rationality of using it in subsequent algorithms, it was finally decided to store the obtained discrimination evaluation indicators in descending order. By sorting the evaluation indicators in descending order, the data or features with the highest discrimination can be identified and extracted.

10. The method for generating fine-grained fingerprints of hardware and software devices based on hash and feature selection according to claim 4, characterized in that: The stability measurement of data in data preprocessing uses the autocorrelation coefficient (ACF) algorithm, let X t is the characteristics of hardware and software equipment in the power environment, X t-k The features after translation, is the average value of a set of characteristic parameters, and the stability measure of the data using the autocorrelation coefficient (ACF) algorithm is: Input: D: a dataset containing n objects, each object has m-dimensional attribute parameters, k is the data translation amount, Output: Stability evaluation index method: 1) According to the amount of data in the data set, set the corresponding k value, 2) By using the information entropy algorithm to analyze the data of n objects in each attribute for each feature in the m dimension, the results are finally sorted in descending order. Through comprehensive comparison of simulated data, 3 is selected as the k data offset, and the obtained discrimination evaluation indicators are stored in ascending order.

Images