A smart grid-oriented privacy electric meter query method based on face recognition

Abstract

The application discloses a privacy electric meter query method based on face recognition for smart grid, which comprises a registration stage and a query stage. In the registration stage, the user binds the face information with the electric meter information and registers in the smart grid. The smart grid extracts the features of the face information of the user, converts the feature vectors obtained by the feature extraction into binary vectors, and then converts the binary vectors into mask polynomials, and binds the mask polynomials with the power consumption information of the user. In the query stage, the smart grid constructs a final binding polynomial by using the mask polynomial and the power consumption information, sends the polynomial to the user, and the user calculates the power consumption information by using the face information and the polynomial. The face information and the power consumption information are encoded and bound by the polynomial, and the user and the smart grid end do not need to store the key in the process, effectively solving the problems of key leakage and privacy query in the smart grid.

Description

Technical Field

[0001] This invention belongs to the field of information security and designs a privacy meter query method based on facial recognition for smart grids. Background Technology

[0002] A smart grid is a new type of power network system that integrates various advanced technologies based on power generation, transmission, distribution, and consumption equipment. In a smart grid, users can access the network to check the power consumption of smart devices and dynamically adjust the usage time of these devices, thereby saving on electricity costs and achieving higher cost-effectiveness.

[0003] Currently, in smart grids, encryption schemes are used to protect transmitted data when users make inquiries, preventing personal information leaks caused by attackers on insecure channels. However, these schemes suffer from key leakage issues. If the encryption key used for data transmission is compromised, attackers can easily compromise the messages on the channel. Introducing biometric identification can effectively address this problem. However, biometrics themselves contain users' personal information, and the Personal Information Protection Law stipulates the need for secure use of users' personal information. Therefore, how to conduct secure inquiries in smart grids using biometrics without disclosing personal information remains a challenge. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides a face recognition-based privacy meter query method for smart grids, which can be used to protect user data privacy and security during meter querying.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A privacy-preserving electricity meter query method based on facial recognition for smart grids, the method comprising the following two stages:

[0007] Registration Phase: Users bind their facial information to their electricity meter information to register within the smart grid. The registration phase includes the following steps:

[0008] Step 1: Initialize smart grid parameters;

[0009] Step 2: The user sends their facial information and electricity meter information to the smart grid system. The system extracts features from the facial information to obtain a feature vector.

[0010] Step 3: Encode the feature vector into a binary vector;

[0011] Step 4: Project the binary code to obtain the feature set;

[0012] Step 5: Generate the mask polynomial. Generate the mask polynomial based on the feature set;

[0013] Step 6: Bind Information. Bind the mask polynomial and the meter information to complete the registration.

[0014] In step 1 of the registration phase, the following parameters are selected and disclosed: security parameter λ, binary vector length n, matching threshold t, large prime number p, and prime number field. And double shot All calculations in this method are performed over the prime number field.

[0015] In step 2 of the registration phase, the FaceNet deep learning model is used to extract facial information feature vectors, thus obtaining the feature vectors.

[0016] v represents a feature vector of length n;

[0017] Represents the real number field;

[0018] v i This represents the element in v that is in the real number field.

[0019] In step 3 of the registration phase, position-sensitive hashing is used to encode the feature vector v into a binary vector of length n, resulting in a binary vector b = LSH(v, bucket). num bucket size code length ).

[0020] v represents a feature vector of length n;

[0021] b represents a binary vector of length n;

[0022] LSH stands for Locality Sensitive Hash;

[0023] bucket uum This indicates the number of hash buckets in the LSH algorithm;

[0024] bucket size This represents the width of the hash bucket in the LSH algorithm;

[0025] code length This represents the dimension of the vector in the LSH algorithm.

[0026] In step 4 of the registration phase, all indices with a bit value of 1 are extracted to obtain the feature set.

[0027] This represents a finite field containing p elements.

[0028] In step 5 of the registration phase, the elements in P are rearranged to obtain... Generate mask polynomial

[0029] σ denotes a bijection used for rearranging elements.

[0030] In step 6 of the registration phase, the mask polynomial will be... With meter information Bind and get Registration complete.

[0031] Query Phase: The user initiates a query request to the smart grid system to obtain the information of the linked electricity meter. The query phase includes the following steps:

[0032] Step 1: The system uses the meter information as a secret to generate a secret polynomial;

[0033] Step 2: The system constructs and sends the final binding polynomial to the user based on the secret polynomial and the mask polynomial;

[0034] Step 3: The user uses the final binding polynomial to calculate and obtain the bound meter information.

[0035] In step 1 of the query phase, all the latest meter data is retrieved. As a constant term, in Randomly select the coefficients of the remaining polynomials to generate N secret polynomials of degree t-1 {κ1(X), ..., κ...} N (X)}.

[0036] In step 2 of the query phase, the final binding polynomial is obtained based on the mask polynomial. and V i The coefficients are sent to the user.

[0037] In step 3 of the query phase, for the bound polynomial V i Users based on their own set of features Substituting the polynomial V i Calculations yielded Each polynomial value is a value.

[0038] right Given t polynomial values ​​κ, arbitrarily select t values ​​κ[t] and perform Shamir Secret Reconstruction (Shamir_Secret_Recover(t, value[t], meter)) with a threshold of t. If the reconstruction is successful, the secret polynomial κ is obtained. i(X), and then obtain the meter data. If the reconstruction fails, continue to select t values ​​value[t] until all possibilities have been selected.

[0039] t represents the matching threshold;

[0040] value[t] represents any t polynomial values.

[0041] If the meter is not reconstructed after selecting all possible value[t], then the next bound polynomial V is... i+1 Continue the above attempts until the meter is reconstructed or the bound polynomial V is fully traversed.

[0042] This invention encodes and binds facial information and electricity consumption information using polynomials. During this process, neither the user nor the smart grid needs to store keys, effectively solving the problems of key leakage and privacy query in smart grids. Attached Figure Description

[0043] Figure 1 This is a diagram illustrating the registration process for privacy-based electricity meter queries using facial recognition.

[0044] Figure 2 This is a diagram illustrating the interaction between the user and the smart grid system during the query phase of a privacy-based electricity meter query. Detailed Implementation

[0045] The present application will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention, and should not be construed as limiting the scope of protection of the present application.

[0046] In the face-based privacy meter query of this invention, users need to go through a registration phase before they can perform the query operation.

[0047] like Figure 1 As shown, the registration phase of face-based privacy meter query includes the following steps:

[0048] Step 1: Initialize smart grid parameters;

[0049] Step 2: The user sends their facial information and electricity meter information to the smart grid system. The system extracts features from the facial information to obtain a feature vector.

[0050] Step 3: Encode the feature vector into a binary vector;

[0051] Step 4: Project the binary code to obtain the feature set;

[0052] Step 5: Generate the mask polynomial. Generate the mask polynomial based on the feature set;

[0053] Step 6: Bind Information. Bind the mask polynomial and the meter information to complete the registration.

[0054] In step 1 of the registration phase, the following parameters are selected and disclosed: security parameter λ, binary vector length n, matching threshold t, large prime number p, and prime number field. And double shot All calculations in this method are performed over the prime number field.

[0055] In step 2 of the registration phase, the FaceNet deep learning model is used to extract facial information feature vectors, thus obtaining the feature vectors.

[0056] v represents a feature vector of length n;

[0057] Represents the real number field;

[0058] v i This represents the element in v that is in the real number field.

[0059] In step 3 of the registration phase, position-sensitive hashing is used to encode the feature vector v into a binary vector of length n, resulting in a binary vector b = LSH(v, bucket). num bucket size code length ).

[0060] v represents a feature vector of length n;

[0061] b represents a binary vector of length n;

[0062] LSH stands for Locality Sensitive Hash;

[0063] bucket num This indicates the number of hash buckets in the LSH algorithm;

[0064] bucket size This represents the width of the hash bucket in the LSH algorithm;

[0065] code length This represents the dimension of the vector in the LSH algorithm.

[0066] In step 4 of the registration phase, all indices with a bit value of 1 are extracted to obtain the feature set.

[0067] This represents a finite field containing p elements.

[0068] In step 5 of the registration phase, the elements in P are rearranged to obtain... Generate mask polynomial

[0069] σ denotes a bijection used for rearranging elements.

[0070] In step 6 of the registration phase, the mask polynomial will be... With meter information Bind and get Registration complete.

[0071] like Figure 2 As shown, during the query phase, the user interacts with the smart grid system as follows:

[0072] Step 1: The system uses the meter information as a secret to generate a secret polynomial;

[0073] Step 2: The system constructs and sends the final binding polynomial to the user based on the secret polynomial and the mask polynomial;

[0074] Step 3: The user uses the final binding polynomial to calculate and obtain the bound meter information.

[0075] In step 1 of the query phase, all the latest meter data is retrieved. As a constant term, in Randomly select the coefficients of the remaining polynomials to generate N secret polynomials of degree t-1 {κ1(X), ..., κ...} N (X)}.

[0076] In step 2 of the query phase, the final binding polynomial is obtained based on the mask polynomial. and V i The coefficients are sent to the user.

[0077] In step 3 of the query phase, for the bound polynomial V i Users based on their own set of features Substituting the polynomial V i Calculations yielded Each polynomial value is a value.

[0078] right Given t polynomial values ​​κ, arbitrarily select t values ​​κ[t] and perform Shamir Secret Reconstruction (Shamir_Secret_Recover(t, value[t], meter)) with a threshold of t. If the reconstruction is successful, the secret polynomial κ is obtained. i (X), and then obtain the meter data. If the reconstruction fails, continue to select t values ​​value[t] until all possibilities have been selected.

[0079] t represents the matching threshold;

[0080] value[t] represents any t polynomial values.

[0081] If the meter is not reconstructed after selecting all possible value[t], then the next bound polynomial V is... i+1 Continue the above attempts until the meter is reconstructed or the bound polynomial V is fully traversed.

[0082] The applicant of this invention has provided a detailed description of the embodiments of the invention in conjunction with the accompanying drawings. However, those skilled in the art should understand that the above embodiments are merely preferred embodiments of the invention. The detailed description is only intended to help readers better understand the spirit of the invention and is not intended to limit the scope of protection of the invention. On the contrary, any improvements or modifications made based on the inventive spirit of the invention should fall within the scope of protection of the invention.

Claims

1. A privacy-preserving electricity meter query method based on facial recognition for smart grids, characterized in that: Registration Phase: Users bind their facial information to their electricity meter information and register within the smart grid; the registration phase includes the following steps: Step 1-1: Initialize smart grid parameters; Steps 1-2: The user sends facial information and electricity meter information to the smart grid system. The system extracts features from the facial information to obtain a feature vector. Steps 1-3: Encode the feature vector into a binary vector; Steps 1-4: Project the binary code to obtain the feature set P; Steps 1-5: Generate the mask polynomial; Generate the mask polynomial based on the feature set P; Steps 1-6: Binding Information, binding the mask polynomial and the meter information to complete the registration; Query phase: The user initiates a query request to the smart grid system and obtains the bound meter information; The query phase includes the following steps: Step 2-1: The system uses the meter information as a secret to generate a secret polynomial; Step 2-2: The system constructs and sends the final binding polynomial to the user based on the secret polynomial and the mask polynomial; Steps 2-3: The user uses the final binding polynomial to calculate and obtain the bound meter information; In step 1-5, the elements in P are rearranged to get A mask polynomial is generated σ denotes a bijection for rearranging elements 2. The privacy meter query method based on face recognition for smart grids according to claim 1, characterized in that: In step 1-1 of the registration phase, the following parameters are selected and disclosed: security parameter λ, binary vector length n, matching threshold t, large prime number p, and prime number field. And double shot 3. The privacy meter query method based on face recognition for smart grids according to claim 2, characterized in that: In step 1-2, a FaceNet deep learning model is used to extract a facial information feature vector, obtaining a feature vector v represents a feature vector of length n; denotes the real numbers domain; v i represents an element in v over the field of real numbers.

4. A privacy-preserving electricity meter query method based on face recognition for smart grids according to claim 3, characterized in that: In step 1-3, the feature vector v is binary coded using the location sensitive hashing technique to obtain a binary vector b = LSH(v, bucket num ,bucket size ,code length ) of length n; v represents a feature vector of length n; b represents a binary vector of length n; LSH stands for Locality Sensitive Hash; bucket num denotes the number of hash buckets in the LSH algorithm; bucket size denotes the width of a hash bucket in the LSH algorithm; code length denotes the dimension of the vectors in the LSH algorithm.

5. A privacy-preserving electricity meter query method based on face recognition for smart grids according to claim 3, characterized in that: In step 1-4 of the registration phase, extract all indices with bit equal to 1 to get the feature set denotes a finite field containing p elements.

6. A privacy-preserving electricity meter query method based on face recognition for smart grids according to claim 5, characterized in that: In step 1-6, the mask polynomial is bound with the meter information to get the registration completed.

7. A privacy-preserving electricity meter query method based on face recognition for smart grids according to claim 6, characterized in that: In step 2-1, all the latest meter data are... As a constant term, in Randomly select the coefficients of the remaining polynomials to generate N secret polynomials of degree t-1 {κ1(X),…,κ...} N (X)}.

8. A privacy-preserving electricity meter query method based on face recognition for smart grids according to claim 7, characterized in that: In step 2-2, the final binding polynomial is obtained according to the mask polynomial and the coefficients of V i are sent to the user.

9. A privacy-preserving electricity meter query method based on face recognition for smart grids according to claim 8, characterized in that: In steps 2-3, for the bound polynomial V i Users based on their own set of features Substituting into polynomial V i Calculations yielded Each polynomial value; for Given t polynomial values ​​κ, arbitrarily select t values ​​κ[t] and perform Shamir secret reconstruction (Shamir_Secret_Recover(t, value[t], meter)) with a threshold of t. If the reconstruction is successful, the secret polynomial κ is obtained. i (X), and then obtain its own meter data; If the reconstruction fails, continue to select t values ​​value[t] until all possibilities have been selected; t represents the matching threshold; value[t] represents any t polynomial values; If all possible values of [t] are selected and meter is not reconstructed, the next binding polynomial V i+1 The above attempt is made until meter is reconstructed or all binding polynomials V are traversed.

Images