Wind power plant face recognition safety management system and method

Abstract

The invention discloses a wind power plant face recognition safety management system and method, and the system comprises a data obtaining module which is used for obtaining wind power plant environment data and personnel biological characteristic data; the face recognition module is used for performing face recognition and dynamically adjusting the user permission based on the wind power plant environment data and the personnel biological characteristic data; the face recognition module comprises a self-adaptive multi-mode biological feature fusion system, an ultra-high-speed face matching engine based on quantum computing, a situation awareness dynamic authority management system and a homomorphic encryption face recognition module; personnel biological characteristic data are fused through a self-adaptive multi-mode biological characteristic fusion system, a super-high-speed face matching engine based on quantum computing is used for recognizing a face, and meanwhile, based on wind power plant environment data, a dynamic authority management system based on context awareness is used for dynamically adjusting the authority level; the homomorphic encryption face recognition module is adopted to encrypt the face recognition process, and face recognition can be achieved while personal privacy is protected.

Description

Technical Field

[0001] The invention belongs to the technical field of wind power generation, and relates to a wind farm face recognition safety management system and method. Background Art

[0002] As the global demand for renewable energy continues to grow, the wind power industry has ushered in unprecedented development opportunities. As the core facility of wind power generation, the safety management of wind farms is directly related to power generation efficiency and operational stability, and is an important part of ensuring the safety of power supply. However, as the scale of wind farms continues to expand, their safety management faces increasingly complex challenges.

[0003] Traditional wind farm safety management methods mainly rely on manual inspections and fixed camera monitoring. Although manual inspections can directly discover potential safety hazards, they are limited by the number of personnel and inspection frequency, making it difficult to achieve comprehensive and real-time monitoring. Although fixed camera monitoring can cover a certain area, it has obvious deficiencies in personnel identification, environmental adaptability, and real-time response. Especially under complex and changeable environmental conditions, such as strong winds, dust, and extreme temperatures, the performance and reliability of traditional monitoring systems are often greatly reduced.

[0004] In recent years, with the rapid development of artificial intelligence technology, face recognition technology has begun to be applied in the field of wind farm safety management. By introducing advanced face recognition algorithms and hardware equipment, the system can quickly and accurately identify the identities of wind farm personnel, providing a new technical means for safety management. However, in actual applications, the wind farm face recognition safety management system still faces many technical difficulties and challenges.

[0005] First, wind farms are usually located in remote areas with complex and changeable environmental conditions, which places extremely high demands on the performance and reliability of face recognition systems. In harsh environments such as strong winds and dust, traditional face recognition algorithms often have difficulty accurately identifying facial features, resulting in increased false alarm and missed alarm rates. At the same time, environmental factors such as extreme temperatures may also have an adverse effect on the stability and service life of face recognition equipment.

[0006] Secondly, the wind farm is vast and sparsely populated, which places high demands on the system's coverage and response speed. Traditional fixed camera monitoring systems often have difficulty achieving full coverage, while increasing the number and density of cameras will bring high costs and maintenance pressure. In addition, even if the system can capture facial images, the system's response speed and processing efficiency will also face severe tests due to the sparse distribution of personnel.

[0007] Furthermore, existing face recognition systems often encounter problems of low computational efficiency when processing large-scale face data. As the number of personnel in wind farms continues to increase, the amount of face data that the system needs to process has also exploded. Traditional computing methods and hardware equipment are difficult to meet the needs of real-time recognition, resulting in a slow response speed of the system or even failure to work properly.

[0008] In addition, in wind farm safety management, although face recognition technology can improve management efficiency, it may also lead to the risk of personal privacy leakage. Therefore, how to ensure system security while taking into account personal privacy protection has become an important factor in the design and application of wind farm face recognition safety management system.

[0009] Therefore, with the increasing awareness of data security and privacy protection, how to achieve effective face recognition while protecting personal privacy has become a pressing issue to be addressed. Summary of the invention

[0010] The purpose of the present invention is to solve the technical problems in the prior art that the wind farm environment is complex, the calculation efficiency of face recognition technology is low, and it is easy to cause privacy leakage, and to provide a wind farm face recognition security management system and method.

[0011] In order to achieve the above object, the present invention adopts the following technical solutions: A first aspect of the present invention provides a wind farm face recognition safety management system, comprising: A data acquisition module, used to acquire wind farm environmental data and personnel biometric data; A face recognition module, which is used to perform face recognition and dynamically adjust user permissions based on wind farm environmental data and personnel biometric data; Among them, the face recognition module includes an adaptive multimodal biometric fusion system, an ultra-high-speed face matching engine based on quantum computing, a context-aware dynamic authority management system and a homomorphic encryption face recognition module; the biometric data of the personnel is fused by the adaptive multimodal biometric fusion system, and the face is recognized by the ultra-high-speed face matching engine based on quantum computing. At the same time, based on the wind farm environmental data, the context-aware dynamic authority management system is used to dynamically adjust the authority level; the homomorphic encryption face recognition module is used to encrypt the face recognition process.

[0012] Furthermore, the wind farm face recognition safety management system of the present invention further includes an edge computing layer, and the edge computing layer includes: A lightweight deep learning framework for implementing preliminary face recognition; Bio-inspired self-organizing repair networks for implementing dynamic routing protocols.

[0013] Furthermore, the personnel biometric data includes face, gait and voiceprint features.

[0014] Furthermore, the adaptive multimodal biometric fusion system adopts a feature fusion network based on the Transformer architecture to fuse face, gait and voiceprint features.

[0015] Furthermore, the ultra-high-speed face matching engine based on quantum computing adopts an improved Grover quantum search algorithm.

[0016] Furthermore, the context-aware dynamic permission management system is implemented based on a graph neural network, uses a graph database to store the relationship between users, permissions and resources, and adopts a deep reinforcement learning algorithm to achieve dynamic adjustment of permission levels.

[0017] Furthermore, the homomorphic encryption face recognition module adopts BGV fully homomorphic encryption.

[0018] Furthermore, the face recognition module also includes a federated learning model update mechanism, which adopts a federated averaging algorithm for model aggregation and uses differential privacy technology to protect data privacy.

[0019] A second aspect of the present invention provides a wind farm face recognition safety management method, comprising the following steps: Obtain wind farm environmental data and personnel biometric data; Performing preliminary detection on the biometric data of the person to obtain first biometric data of the person; Performing feature fusion on the biometric data of the first person, and performing face matching on the fused features; Based on the wind farm environment data and the matching facial data, user permissions are adjusted.

[0020] Furthermore, BGV fully homomorphic encryption is used for data encryption during face matching.

[0021] Compared with the prior art, the present invention has the following beneficial effects: The present invention discloses a wind farm face recognition security management system. The data acquisition module can capture the environmental data of the wind farm and the biometric data of personnel in real time and accurately, providing a solid data foundation for subsequent face recognition; the adaptive multimodal biometric fusion system can intelligently fuse multiple biometric data, improve the accuracy and robustness of recognition, and effectively cope with the complex and changeable wind farm environment. The ultra-high-speed face matching engine based on quantum computing uses the powerful parallel processing capability of quantum computing to achieve high-speed matching of facial features, significantly improving the speed and accuracy of face recognition. The context-aware dynamic permission management system can intelligently adjust the user permission level according to the wind farm environmental data and user behavior characteristics, ensuring the rationality and security of user permissions in a complex and changeable environment. The homomorphic encryption face recognition module adopts advanced homomorphic encryption technology in the face recognition process, ensuring that data is matched and recognized in an encrypted state, effectively preventing data leakage and privacy infringement.

[0022] Furthermore, the federated learning model update mechanism allows the system to perform model training and optimization without centralizing data, improving the scalability and flexibility of the system; by continuously learning and updating the model, the system can adapt to changes in the wind farm environment and personnel characteristics, maintain continuous optimization of its recognition performance, and provide users with more accurate and reliable services.

[0023] Furthermore, the present invention significantly improves the safety management level of wind farms through intelligent face recognition and authority management, strict data encryption and privacy protection, and the ability to update and continuously optimize models; it not only improves the operating efficiency of wind farms and the convenience of personnel management, but also effectively prevents illegal intrusions and safety accidents, providing strong technical support for the sustainable development of wind farms.

[0024] Furthermore, the present invention also discloses a wind farm face recognition security management method, which first obtains the environmental data of the wind farm and the biometric data of the personnel, and through the introduction of the environmental data, provides a more comprehensive and accurate basis for the subsequent authority adjustment; performs preliminary detection on the biometric data of the personnel, effectively excludes invalid or low-quality feature data, and lays a solid foundation for subsequent feature fusion and face matching; the feature fusion process can effectively integrate a variety of biometric data, improves the robustness and accuracy of recognition, and provides a more reliable means of personnel identity authentication for the wind farm; uses BGV fully homomorphic encryption to encrypt data in the face matching process, ensuring the security of data during transmission and storage; at the same time, can adjust the user's authority according to the wind farm environmental data and the matching face data, which not only improves the safety management level of the wind farm, but also makes the authority management more flexible and intelligent, and can make timely and accurate responses according to different environments and personnel situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0026] Figure 1 This is a block diagram of the wind farm face recognition safety management system of the present invention; Figure 2 This is a block diagram of the wind farm face recognition safety management method of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention described and marked in the drawings here can be arranged and designed in various different configurations.

[0028] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0029] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0030] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. indicate an orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed when in use, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0031] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0032] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0033] The present invention is further described in detail below in conjunction with the accompanying drawings: See also Figure 1 ,The present invention provides a wind farm face recognition safety management system, including a data acquisition module, an edge computing layer, a face recognition module and an application layer.

[0034] The data acquisition module is used to acquire wind farm environment data and personnel biometric data; the wind farm environment data and personnel biometric data are acquired through the perception layer.

[0035] The perception layer is the foundation of the entire system and consists of intelligent sensor devices distributed throughout the wind farm, including but not limited to high-definition cameras, millimeter-wave radars, thermal imaging cameras, environmental sensors, and sound collectors. These devices are responsible for collecting environmental data and personnel biometric data from all directions and angles.

[0036] The perception layer is used to collect wind farm environmental data and personnel biometric data; it is implemented using a distributed intelligent sensor network. The distributed intelligent sensor network includes high-definition cameras, millimeter-wave radars, thermal imaging cameras, environmental sensors, and sound collectors. The high-definition camera in this embodiment uses a 1 / 1.8-inch CMOS sensor with a resolution of 3840*2160 (4K), a frame rate of 120fps, and is equipped with a 6mm wide-angle and 50mm telephoto dual lens. This configuration can capture clear facial images at different distances and angles. The operating frequency of the millimeter-wave radar is 77GHz, the detection range is 100 meters, and the angular resolution is 0.1°; the millimeter-wave radar can provide reliable personnel detection capabilities under severe weather and lighting conditions. The thermal imaging camera uses a non-cooled vanadium oxide detector with a temperature detection range of to , thermal sensitivity ; Thermal imaging camera provides the system with recognition capability at night and in low visibility conditions. Environmental sensor integrated temperature (accuracy ), humidity (accuracy ±2%RH), wind speed (range 0-60m / s), and light intensity (range 0-200,000lux) sensors. These sensors provide rich environmental information and help the system adjust its recognition strategy. The sound collector uses a MEMS (Micro-Electro-Mechanical Systems) microphone array, 360° omnidirectional sound pickup, a signal-to-noise ratio of >70dB, and a noise suppression capability of >30dB. The sound collector provides audio features for the system and enhances the ability of multimodal recognition.

[0037] The edge computing layer is used to pre-process and preliminarily analyze the collected data; it is composed of edge computing units distributed throughout the wind farm. In this embodiment, the hardware configuration of each edge computing unit is that the processor uses a 64-core ARM architecture CPU with a main frequency of 3.0GHz; an integrated embedded GPU that supports 384 CUDA cores; 32GB LPDDR5 high-speed memory; 1TB NVMe SSD for storage, supporting high-speed data reading and writing; integrated 5G module and Wi-Fi 6, supporting high-speed wireless communication.

[0038] The edge computing layer also includes: A lightweight deep learning framework for local face detection and preliminary recognition; the lightweight deep learning framework in this embodiment is optimized based on TensorFlow Lite, supports INT8 quantization, and can efficiently run complex neural network models on edge devices. This lightweight deep learning framework implements preliminary face detection and feature extraction functions, using MobileNetV3 as the basic network structure.

[0039] Bio-inspired self-organizing repair network for implementing dynamic routing protocol; dynamic routing protocol is implemented based on improved ant colony algorithm (ACO). The routing selection probability is calculated as follows:

[0040] in, It is at the node i Select the next hop node j The probability of Is the path The pheromone concentration on is the heuristic information, and are the parameters of pheromone importance and heuristic information importance respectively. The algorithm can automatically adjust the routing strategy according to the network status and improve the robustness of the network.

[0041] Distributed data preprocessing module, used for data cleaning, normalization and noise reduction; realizes preprocessing functions such as data cleaning, normalization, noise reduction, etc. For image data, the adaptive histogram equalization (CLAHE) algorithm is used for enhancement; for audio data, wavelet transform is used for noise reduction.

[0042] The edge security module is used to implement data encryption, identity authentication and access control; it uses the AES-256 algorithm to encrypt sensitive data stored locally, and uses the digital signature algorithm (ECDSA) based on elliptic curve cryptography (ECC) for identity authentication.

[0043] Resource scheduling optimizer, used to dynamically allocate computing resources; based on reinforcement learning algorithms (such as DQN), dynamic allocation of computing resources is achieved, including state space and action space; the state space includes current CPU utilization, memory usage, and task queue length, etc. The action space includes adjusting CPU frequency, allocating GPU resources, and adjusting task priorities, etc.

[0044] The face recognition module is responsible for processing data from all edge computing layers for in-depth analysis and decision-making. The face recognition module is deployed in the central control room of the wind farm, using a high-performance server cluster. The specific configuration used in this embodiment is: using the latest generation of Xeon processors, each server is equipped with a dual-channel processor, each with 64 cores and a main frequency of 3.5GHz; each server is equipped with 1TB DDR5 ECC memory to support high-speed data processing and large-scale concurrent operations; using a distributed storage architecture with a total capacity of 1PB, using high-speed SSD as the cache layer and large-capacity HDD as the storage layer; using 100Gbps InfiniBand network interconnection to ensure high-speed data transmission within the cluster.

[0045] The face recognition module includes a high-performance server cluster with an adaptive multimodal biometric fusion system, an ultra-high-speed face matching engine based on quantum computing, a context-aware dynamic permission management system, a homomorphic encrypted face recognition module, and a federated learning model update mechanism.

[0046] The adaptive multimodal biometric fusion system is implemented using the deep learning framework PyTorch. The core is a feature fusion network based on the Transformer architecture. Its input is feature vectors from different sensors, such as a 256-dimensional face feature vector, a 128-dimensional gait feature vector, and a 64-dimensional voiceprint feature vector. The fusion process can be expressed as:

[0047] in, is the fused feature vector, , , are respectively the query, key and value matrices obtained by linear transformation of each modal feature, is the output projection matrix. The network dynamically adjusts the weights of each feature through the attention mechanism to adapt to different environmental conditions.

[0048] The ultra-high-speed face matching engine based on quantum computing works in conjunction with a dedicated quantum processing unit (QPU). The QPU uses superconducting quantum bit technology and maintains an operating temperature of around 10mK. The ultra-high-speed face matching engine based on quantum computing uses an improved Grover quantum search algorithm, with a search process iteration number of R The calculation is as follows:

[0049] in, N is the size of the search space, M is the number of target states. The ultra-high-speed face matching engine based on quantum computing can complete the search of millions of face libraries within milliseconds.

[0050] The context-aware dynamic permission management system is implemented based on graph neural networks (GNN); in the context-aware dynamic permission management system, the complex relationships between users, permissions, and resources are stored in the Neo4j graph database. In order to effectively process these graph-structured data, the system uses the GraphSAGE algorithm for graph embedding, so that the potential connections and features between nodes can be captured. Then, the system uses the DDPG (Deep Deterministic Policy Gradient) algorithm, an advanced deep reinforcement learning algorithm, to achieve dynamic output of continuous permission levels. In this way, the system can flexibly adjust the permission level according to the current situation and user needs. The homomorphic encryption face recognition module uses the BGV fully homomorphic encryption scheme, with a key length of 3072 bits and uses NTL (Number Theory Library) to implement the underlying cryptographic operations. In the encryption domain, an approximate activation function (such as Instead of ReLU), a simplified CNN structure is constructed to realize encrypted domain feature extraction.

[0051] The federated learning model update mechanism is implemented based on the TensorFlow Federated framework, and the FedAvg algorithm is used for model aggregation. Gaussian noise is added to the client gradient to achieve differential privacy protection, and the privacy budget ε is set to 0.1.

[0052] The application layer provides users with a friendly interactive interface and rich functions, such as real-time monitoring, alarm management, authority control, etc. The application layer includes: The visual interface is a web application developed using the React framework, which provides functions such as real-time video stream display, personnel location map, and alarm information display.

[0053] The alarm management system, based on rule engine and machine learning algorithm, realizes intelligent alarm triggering and automatic classification of alarm levels.

[0054] The permission management interface is connected to the dynamic permission management system of the cloud computing layer to provide user permission configuration and management functions.

[0055] Data analysis tools integrate data visualization tools (such as ECharts) to provide analysis reports such as personnel flow trends and security incident statistics.

[0056] Mobile application that supports remote monitoring and emergency response functions.

[0057] In order to improve the reliability of the system in harsh environments, the wind farm face recognition safety management system of the present invention also includes an intelligent nanomaterial meteorological protection system, including: Self-cleaning nanocoating: Composed of titanium dioxide (TiO2) nanoparticles and fluorinated polymers, it is prepared by a sol-gel method and applied to the device surface using plasma-assisted chemical vapor deposition (PACVD) technology.

[0058] Temperature-responsive phase change material: n-octadecane microcapsules (phase change temperature: 28°C) were selected and prepared by interfacial polymerization and blended with the device housing material to form a composite material.

[0059] Transparent conductive antistatic coating: It is composed of indium tin oxide (ITO) nanowires and graphene composite materials, prepared by spin coating and then laser sintering. Performance indicators: transmittance>90%, square resistance<100Ω / sq.

[0060] Through the close collaboration between the face recognition module and the edge computing layer, the present invention realizes the organic combination of distributed intelligence and central decision-making. The edge computing layer performs preprocessing and preliminary analysis at the source of the data, which greatly reduces the burden of the face recognition module and improves the real-time performance and reliability of the system. The face recognition module, with its powerful computing power, realizes complex data fusion, large-scale search and intelligent decision-making functions.

[0061] See also Figure 2 An embodiment of the present invention provides a wind farm face recognition safety management method, comprising the following steps: S1, obtain wind farm environmental data and personnel biometric data; obtain wind farm environmental data and personnel biometric data. Environmental data includes the geographical location, climate conditions, equipment status, etc. of the wind farm, which can be collected through the wind farm monitoring system and sensors. Personnel biometric data mainly includes facial images, iris information, etc., which can be collected through biometric identification devices installed at the entrance or key areas of the wind farm.

[0062] S2, perform preliminary detection on the biometric data of the personnel to obtain the first biometric data of the personnel; after obtaining the biometric data of the personnel, perform preliminary detection on it to screen out the first biometric data of the personnel that meets the quality requirements. This step can be achieved by checking the clarity and integrity of the image to ensure that the subsequent feature fusion and face matching can proceed smoothly.

[0063] S3, feature fusion is performed on the biometric data of the first person, and face matching is performed on the fused features; Feature fusion is performed on the biometric data of the first person after preliminary detection. Feature fusion refers to the fusion of information of multiple biometric features (such as face, iris, etc.) to improve the accuracy and robustness of recognition. After feature fusion, we will use the fused features for face matching. This step can be achieved by comparing the fused features with the facial features stored in the database to determine the identity of the user. In the process of face matching, we use BGV fully homomorphic encryption to encrypt the data. BGV fully homomorphic encryption is an encryption method that can calculate encrypted data without decryption. It can realize the function of face matching while protecting user privacy. The application of this technology further improves the security and reliability of face recognition.

[0064] S4, based on the wind farm environment data and the matching face data, adjust the user's permissions.

[0065] The permissions of users will be adjusted based on the wind farm environment data and the matched facial data. Specifically, the user's access rights can be dynamically adjusted according to the actual situation of the wind farm (such as equipment status, climate conditions, etc.) and the user's identity and permission level. For example, in the event of equipment failure or severe weather conditions, we can restrict the access rights of some users to ensure the safe operation of the wind farm.

[0066] By implementing the above method, the present invention successfully improves the safety management level of wind farms; on the one hand, by adopting face recognition technology and feature fusion method, it realizes the rapid and accurate identification of user identity; on the other hand, by adopting BGV fully homomorphic encryption technology, it effectively protects the privacy and data security of users. In addition, by adjusting the user's authority based on the wind farm environmental data and the matching face data, the operation efficiency and safety of the wind farm are further improved.

[0067] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A wind farm face recognition safety management system, characterized in that: include: A data acquisition module, used to acquire wind farm environmental data and personnel biometric data; A face recognition module, which is used to perform face recognition and dynamically adjust user permissions based on wind farm environmental data and personnel biometric data; Among them, the face recognition module includes an adaptive multimodal biometric fusion system, an ultra-high-speed face matching engine based on quantum computing, a context-aware dynamic authority management system and a homomorphic encryption face recognition module; the biometric data of the personnel is fused by the adaptive multimodal biometric fusion system, and the face is recognized by the ultra-high-speed face matching engine based on quantum computing. At the same time, based on the wind farm environmental data, the context-aware dynamic authority management system is used to dynamically adjust the authority level; the homomorphic encryption face recognition module is used to encrypt the face recognition process.

2. The wind farm face recognition safety management system according to claim 1, characterized in that: Also included is an edge computing layer, the edge computing layer comprising: A lightweight deep learning framework for implementing preliminary face recognition; Bio-inspired self-organizing repair networks for implementing dynamic routing protocols.

3. The wind farm face recognition safety management system according to claim 1, characterized in that: The personnel biometric data includes face, gait and voiceprint features.

4. The wind farm face recognition safety management system according to claim 3, characterized in that: The adaptive multimodal biometric fusion system adopts a feature fusion network based on the Transformer architecture to fuse face, gait and voiceprint features.

5. The wind farm face recognition safety management system according to claim 1, characterized in that: The ultra-high-speed face matching engine based on quantum computing adopts an improved Grover quantum search algorithm.

6. The wind farm face recognition safety management system according to claim 1, characterized in that: The context-aware dynamic permission management system is implemented based on a graph neural network, uses a graph database to store the relationship between users, permissions and resources, and adopts a deep reinforcement learning algorithm to achieve dynamic adjustment of permission levels.

7. The wind farm face recognition safety management system according to claim 1, characterized in that: The homomorphic encryption face recognition module adopts BGV fully homomorphic encryption.

8. The wind farm face recognition safety management system according to claim 1, characterized in that: The face recognition module also includes a federated learning model update mechanism, which adopts a federated averaging algorithm to perform model aggregation and uses differential privacy technology to protect data privacy.

9. A wind farm face recognition safety management method, characterized in that: The following steps are involved: Obtain wind farm environmental data and personnel biometric data; Performing preliminary detection on the biometric data of the person to obtain first biometric data of the person; Performing feature fusion on the biometric data of the first person, and performing face matching on the fused features; Based on the wind farm environment data and the matching facial data, user permissions are adjusted.

10. The wind farm face recognition safety management method according to claim 9, characterized in that: BGV fully homomorphic encryption is used for data encryption during face matching.

Images