Wind power plant intelligent video monitoring and abnormal behavior recognition system, method and equipment and medium

Abstract

The invention discloses a wind power plant intelligent video monitoring and abnormal behavior recognition system, method and device and a medium, and belongs to the technical field of wind power generation, and the system comprises a multi-mode collection module which is used for collecting various data of a wind power plant; the edge calculation module is used for receiving and primarily processing data of the wind power plant; the communication network module is used for transmitting the processed data of the wind power plant to the central server cluster module; the central server cluster module is used for receiving the data transmitted by the communication network module and executing deep learning model training and large-scale data analysis; and the unmanned aerial vehicle subsystem is used for executing large-range inspection and close-range inspection and transmitting the acquired data to the central server cluster module through the communication network module. The problems that a traditional wind power plant monitoring system is difficult in data integration, weak in analysis capacity and poor in real-time performance are solved, the operation efficiency, safety and reliability of a wind power plant are remarkably improved, and technical support is provided for sustainable development of the wind energy industry.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind power generation, and in particular relates to a wind farm intelligent video monitoring and abnormal behavior recognition system, method, equipment and medium. Background Art

[0002] Wind farm monitoring systems are key technologies to ensure safe and efficient operation of wind power generation equipment. Such systems usually include various sensors, data acquisition equipment, communication networks and central control units to monitor the operating status, environmental conditions and possible abnormal conditions of wind turbines in real time. An effective monitoring system can detect potential problems in a timely manner, prevent equipment failures, and optimize operating parameters, thereby improving the overall efficiency and reliability of wind farms.

[0003] Currently, most wind farm monitoring systems rely primarily on a single type of sensor data, such as operating parameters collected by supervisory control and data acquisition (SCADA) systems or image data collected by video surveillance systems. These systems perform well in specific areas, but often lack comprehensiveness and in-depth analysis capabilities. In addition, with the continuous expansion of wind farm scale and the increase in distributed deployment, traditional centralized monitoring methods face huge challenges in data transmission, storage, and real-time processing. Especially under severe weather conditions, communication interruptions may cause the monitoring system to fail, affecting the safe operation of wind farms.

[0004] Therefore, a major technical problem facing wind farm monitoring systems is how to effectively integrate and analyze information from multiple heterogeneous data sources. This includes multimodal data such as video streams, sound, vibration, temperature, wind speed, etc. These data types are different, the data volume is huge, and there is temporal and spatial correlation. Traditional data processing methods are difficult to fully mine the valuable information contained therein. At the same time, how to reduce the pressure of data transmission and storage and improve the real-time response capability of the system while ensuring the comprehensiveness of the data and the depth of analysis is also an urgent problem to be solved. Summary of the invention

[0005] In order to overcome the shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide a wind farm intelligent video monitoring and abnormal behavior recognition system to solve the technical problems of single sensor limitations, low data processing efficiency, insufficient communication reliability and low anomaly detection accuracy existing in traditional wind farm monitoring systems.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: The present invention discloses a wind farm intelligent video monitoring and abnormal behavior recognition system, comprising: Multimodal acquisition module, used to collect data from wind farms; An edge computing module, used for receiving and preliminarily processing the wind farm data collected by the multi-modal acquisition module; A communication network module, used to transmit the wind farm data processed by the edge computing module to the central server cluster module; The central server cluster module is used to receive data transmitted by the communication network module and perform deep learning model training and large-scale data analysis; The drone subsystem is used to perform large-scale inspections and close-range inspections, and transmit the collected data to the central server cluster module through the communication network module; The data of the wind farm includes: video image data, acoustic data, vibration data, stress distribution data, meteorological data, magnetic field data and gravity data.

[0007] Preferably, the video image data is obtained by collecting all-round video images of the wind turbine through multiple sets of high-definition cameras, as well as thermal imaging images for nighttime monitoring; The acoustic data collects noise levels and performs abnormal sound recognition through an acoustic sensor array; The vibration data is monitored and recorded by a vibration sensor; The stress distribution data is obtained by monitoring the stress distribution data of the tower and the blades through an optical fiber sensing network; The meteorological data includes wind speed data, temperature data, humidity data and air pressure data collected by meteorological sensors; The magnetic field data is detected by a quantum sensor network to detect tiny magnetic field change data of generators and transformers; The gravity data is data of foundation settlement and tower inclination monitored by a quantum gravimeter.

[0008] Preferably, the edge computing module includes: High-performance edge AI processors to perform real-time data processing and preliminary analysis; Tensor processing units, used to accelerate deep learning tasks; The edge computing module runs an adaptive compression algorithm to dynamically adjust the video quality according to the real-time status of the communication network module to optimize data transmission efficiency; The communication network module comprises: 5G private network for high-speed and low-latency data transmission; Satellite communication backup to ensure communication reliability in extreme weather conditions; Quantum communication networks, which are used to provide ultra-high security levels for data transmission to prevent information leakage and hacker attacks; The central server cluster module includes: High-performance CPU and GPU servers for performing complex data analysis and deep learning model training; Quantum annealers, used to solve complex optimization problems; Large-capacity storage arrays are used to store and manage massive amounts of historical and real-time data.

[0009] Preferably, the system further comprises a data visualization platform for receiving and displaying the analysis results of the central server cluster; The system also includes a multimodal deep learning module, which is deployed on the central server cluster module and is used for a modal fusion architecture based on a Transformer architecture to achieve effective fusion of video data, audio data, and sensor data from the multimodal acquisition module; The system also includes a spatiotemporal attention mechanism module, which is deployed on a central server cluster and includes: 3D convolutional neural networks to capture the spatiotemporal features of video data; Long short-term memory network variants, used to enhance the modeling capabilities of long time series data; Graph attention network, used to model the spatial relationship and mutual influence between wind turbines.

[0010] Preferably, the system also includes a federated learning module, which realizes data privacy protection and distributed learning by implementing a secure aggregation protocol, an asynchronous federated learning algorithm and a model compression technology between the edge computing module and the central server cluster module, so that the system can improve model performance by utilizing data from multiple wind farms without sharing the original data.

[0011] Preferably, the system also includes an adaptive threshold algorithm module, which is deployed on the edge computing module and the central server cluster module, and adopts a dynamic time warping algorithm, a Gaussian mixture model and an online anomaly detection algorithm to achieve real-time adjustment of the anomaly detection threshold and improve the system's ability to identify various abnormal situations.

[0012] Preferably, the system further comprises a biologically inspired algorithm module, which is deployed on the central server cluster module and comprises: Ant colony optimization algorithm, used to optimize the camera network layout in the multimodal acquisition module and the inspection path of the drone subsystem; Artificial bee colony algorithm, used to perform multi-objective optimization in system operation and balance monitoring coverage and energy consumption; Firefly algorithm for rapid identification of anomalous patterns in multimodal data.

[0013] The present invention also discloses a wind farm intelligent video monitoring and abnormal behavior recognition method, comprising: Collect data from wind farms; Receive and preliminarily process the collected wind farm data; Transmit the processed data of the wind farm; Receive transmitted data and perform deep learning model training and large-scale data analysis; Perform wide-range inspections and close-range inspections, and transmit collected data; The data of the wind farm includes: video image data, acoustic data, vibration data, stress distribution data, meteorological data, magnetic field data and gravity data.

[0014] The present invention also discloses an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned wind farm intelligent video monitoring and abnormal behavior identification method are implemented.

[0015] The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned wind farm intelligent video monitoring and abnormal behavior identification method are implemented.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention realizes all-round intelligent monitoring of wind farms and automatic identification of abnormal behaviors by working together through the edge computing module and the central server cluster module, integrating multimodal sensor data, combining edge computing with cloud computing, and applying quantum technology and advanced deep learning algorithms, significantly improving the comprehensiveness, real-time and intelligent level of the wind farm monitoring system. The system can perceive the state of the wind farm in all directions, realize rapid and accurate identification of abnormal behaviors, and greatly improve the fault prediction and prevention capabilities. The introduction of edge computing reduces the burden of data transmission and improves the system response speed, while the application of quantum technology enhances data security and computing efficiency. The present invention solves the technical problems of difficult data integration, weak analysis ability and poor real-time performance of traditional wind farm monitoring systems, limitations of single sensors, low data processing efficiency, insufficient communication reliability, and low accuracy of abnormal detection, significantly improving the operational efficiency, safety and reliability of wind farms, and providing technical support for the sustainable development of the wind energy industry.

[0017] Furthermore, the processed results are transmitted to the data visualization platform and presented to the operator in an intuitive and easy-to-understand manner. The central server cluster module is responsible for performing deep learning model training and large-scale data analysis. It includes high-performance CPU and GPU servers, quantum annealers, and large-capacity storage arrays. The quantum annealer is used to solve complex optimization problems, such as wind turbine layout optimization, which is of great significance to improving the overall efficiency of wind farms. The communication network module ensures efficient and secure data transmission between the components of the system, including 5G private network, satellite communication backup, and quantum communication network. The 5G private network realizes high-speed and low-latency data transmission, the satellite communication backup ensures communication reliability under extreme weather conditions, and the quantum communication network is used for data transmission with ultra-high security level, effectively preventing information leakage and hacker attacks.

[0018] Furthermore, the spatiotemporal attention mechanism module enhances the system's ability to process time series data, enabling the system to effectively capture the spatiotemporal characteristics of video data, enhance the modeling capabilities of long time series data, and accurately model the spatial relationships and mutual influences between fans.

[0019] Furthermore, the federated learning module effectively protects the data privacy of participants through secure aggregation protocols, asynchronous federated learning algorithms and model compression technology, while adapting to the network conditions and computing power of different wind farms. This distributed learning method enables the system to use data from multiple wind farms to improve model performance without sharing the original data.

[0020] Furthermore, the adaptive threshold algorithm module uses a dynamic time warping algorithm, a Gaussian mixture model, and an online anomaly detection algorithm to achieve modeling of complex multimodal distributions and real-time adjustment of anomaly detection thresholds. This greatly improves the system's ability to identify various abnormal situations, including equipment failures, environmental anomalies, and security threats.

[0021] Furthermore, the bio-inspired algorithm and adaptive learning method used enable the system to have self-optimization capabilities, which can continuously improve the monitoring accuracy and efficiency. These improvements work together to significantly improve the operational efficiency, safety and reliability of wind farms, providing strong technical support for the sustainable development of the wind energy industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of the hardware architecture of the wind farm intelligent video monitoring and abnormal behavior recognition system of the present invention; Figure 2 A schematic diagram of the software architecture of the wind farm intelligent video monitoring and abnormal behavior recognition system of the present invention; Figure 3 The figure is a data flow diagram of the wind farm intelligent video monitoring and abnormal behavior recognition system of the present invention. DETAILED DESCRIPTION

[0023] The present invention will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0024] The following detailed description is an exemplary description, which is intended to provide further detailed description of the present invention. Unless otherwise specified, all technical terms used in the present invention have the same meaning as those generally understood by those skilled in the art to which the present application belongs. The terms used in the present invention are only for describing specific embodiments, and are not intended to limit exemplary embodiments according to the present invention.

[0025] On the one hand, the present invention provides a wind farm intelligent video monitoring and abnormal behavior recognition system, the system comprising: Multimodal acquisition module, used to collect various data of wind farms; The edge computing module is installed on the wind turbine tower and is used to receive and preliminarily process the data from the multimodal acquisition module; The communication network module is used to transmit the data processed by the edge computing module to the central server cluster module; The central server cluster module is used to receive data transmitted by the communication network module and perform deep learning model training and large-scale data analysis; The drone subsystem is used to perform large-scale inspections and close-range inspections, and transmit the collected data to the central server cluster module through the communication network module; Data visualization platform, used to receive and display the analysis results of the central server cluster module; The system works together through the edge computing module and the central server cluster module, integrates multimodal data, applies deep learning algorithms and quantum technology, and realizes all-round intelligent monitoring of wind farms and automatic identification of abnormal behaviors.

[0026] Furthermore, the multimodal acquisition module includes: A high-definition camera network, which includes a 360-degree panoramic camera installed on the top of each wind turbine, PTZ cameras installed around each wind turbine, and infrared thermal imaging cameras for nighttime monitoring; Acoustic sensor arrays, using MEMS microphones, for noise monitoring and abnormal sound recognition; Vibration sensors, based on piezoelectric materials, are installed at key locations on the wind turbine; Fiber-optic sensing network, using fiber Bragg grating technology, to monitor tower and blade stress distribution; Meteorological sensors to monitor wind speed, temperature, humidity and air pressure; A network of quantum sensors, including SQUID magnetic sensors, which detect tiny magnetic field changes in generators and transformers, and quantum gravimeters, which monitor foundation settlement and tower tilt.

[0027] Furthermore, the edge computing module includes: High-performance edge AI processors to perform real-time data processing and preliminary analysis; Tensor processing units, used to accelerate deep learning tasks; Among them, the edge computing module runs an adaptive compression algorithm to dynamically adjust the video quality according to the real-time status of the communication network to optimize data transmission efficiency.

[0028] Furthermore, the communication network module includes: 5G private network for high-speed, low-latency data transmission; Satellite communication backup to ensure communication reliability in extreme weather conditions; Quantum communication networks are used to provide ultra-high security levels for data transmission, preventing information leakage and hacker attacks.

[0029] Furthermore, the central server cluster module includes: High-performance CPU and GPU servers for performing complex data analysis and deep learning model training; Quantum annealers, used to solve complex optimization problems such as wind turbine layout optimization; Large-capacity storage arrays are used to store and manage massive amounts of historical and real-time data.

[0030] Furthermore, the system also includes a multimodal deep learning module, which is deployed on the central server cluster module and adopts a modal fusion architecture based on the Transformer architecture to achieve effective fusion of video, audio and sensor data from the multimodal sensor network, thereby improving the accuracy of anomaly detection.

[0031] Furthermore, the system also includes a spatiotemporal attention mechanism module, which is deployed on the central server cluster module and includes: 3D convolutional neural networks to capture the spatiotemporal features of video data; Long short-term memory network variants, used to enhance the modeling capabilities of long time series data; Graph attention network, used to model the spatial relationship and mutual influence between wind turbines.

[0032] Furthermore, the system also includes a federated learning module, which realizes data privacy protection and distributed learning by implementing a secure aggregation protocol, an asynchronous federated learning algorithm and a model compression technology between the edge computing module and the central server cluster module, so that the system can improve model performance by utilizing data from multiple wind farms without sharing the original data.

[0033] Furthermore, the system also includes an adaptive threshold algorithm module, which is deployed on the edge computing module and the central server cluster module, and adopts a dynamic time warping algorithm, a Gaussian mixture model and an online anomaly detection algorithm to achieve real-time adjustment of the anomaly detection threshold and improve the system's ability to identify various abnormal situations.

[0034] Furthermore, the system also includes a biologically inspired algorithm module, which is deployed on the central server cluster module and includes: Ant colony optimization algorithm, used to optimize the camera network layout in multimodal sensor networks and the inspection path of drone subsystems; Artificial bee colony algorithm, used to perform multi-objective optimization in system operation and balance monitoring coverage and energy consumption; Firefly algorithm for rapid identification of anomalous patterns in multimodal data.

[0035] The present invention provides a wind farm intelligent video monitoring and abnormal behavior recognition system, which realizes all-round intelligent monitoring of wind farms and automatic recognition of abnormal behaviors by integrating multimodal data, applying deep learning algorithms and quantum technology. This system aims to solve the problems of single sensor limitations, low data processing efficiency, insufficient communication reliability, low anomaly detection accuracy, etc. in traditional wind farm monitoring systems, and significantly improve the operational efficiency and safety of wind farms.

[0036] The core of this system lies in its comprehensiveness and innovation. Through hardware facilities such as multimodal acquisition modules, edge computing modules, central server cluster modules, advanced communication network modules, drone subsystems, etc., combined with multimodal deep learning, spatiotemporal attention mechanism, federated learning, adaptive threshold algorithm, biologically inspired algorithm, quantum perception and quantum machine learning and other software modules, a comprehensive, efficient and reliable wind farm monitoring system is constructed.

[0037] Figure 1 The hardware architecture diagram of the wind farm intelligent video monitoring and abnormal behavior recognition system of the present invention is as follows: As the data source of the system, the multimodal acquisition module continuously collects various types of data from the wind farm and transmits them to the edge computing module. The edge computing module is located on each wind turbine tower and receives and preliminarily processes these raw data. The edge computing module runs the video acquisition and preprocessing module to compress and reduce noise on the raw video data, and performs preliminary intelligent analysis such as target detection and simple anomaly recognition through the edge AI analysis module.

[0038] After processing the data locally, the edge computing module transmits the processed data to the central server cluster module through the communication network module. The communication network module plays the role of a data transmission bridge to ensure efficient and reliable connection between edge devices and the central system. As a mobile data acquisition platform, the drone subsystem also transmits the collected video and image data to the central server cluster module in real time through the communication network module and receives control instructions from the central system.

[0039] The central server cluster module receives data from all edge computing modules and drone subsystems for deeper analysis and processing. The central AI analysis module uses this aggregated data for deep learning and advanced analysis, while the quantum computing module handles particularly complex computing tasks such as large-scale optimization problems or complex pattern recognition. The processed results are transmitted to the data visualization platform and presented to the operator in an intuitive and easy-to-understand way.

[0040] Figure 2 FIG. 1 is a schematic diagram of the software architecture of the wind farm intelligent video monitoring and abnormal behavior recognition system of the present invention; Figure 2 As shown, the operating system and middleware provide basic software support for the edge computing module and the central server cluster module, ensuring the effective management of hardware resources and the smooth operation of each software module.

[0041] Figure 3 Schematic diagram of the data flow of the wind farm intelligent video monitoring and abnormal behavior recognition system of the present invention; Figure 3 As shown in the figure, these components form a complete closed-loop system of data collection, transmission, processing and analysis through close interaction. From data generation to final decision support, each component plays a key role in it, and together they realize intelligent monitoring and abnormal behavior identification of wind farms.

[0042] The multimodal acquisition module is the basis of this system, which includes a high-definition camera network, an acoustic sensor array, a vibration sensor, a fiber optic sensor network, a meteorological sensor, and a quantum sensor network. The high-definition camera network consists of a 360-degree panoramic camera (with a resolution of no less than 8K) installed on the top of each wind turbine, 3-4 PTZ cameras with 30x optical zoom around each wind turbine, and an infrared thermal imaging camera for night monitoring and equipment thermal anomaly detection. The acoustic sensor array uses MEMS microphones for noise monitoring and abnormal sound recognition. The vibration sensor is based on piezoelectric materials and is installed in key parts of the wind turbine. The fiber optic sensor network uses fiber Bragg grating (FBG) technology to monitor the stress distribution of the tower and blades. Meteorological sensors include anemometers, thermometers, hygrometers, barometers, etc. The quantum sensor network includes SQUID magnetic sensors and quantum gravimeters, which are used to detect small magnetic field changes in generators and transformers, and to monitor foundation settlement and tower tilt.

[0043] Edge computing modules are installed on each wind turbine tower, including high-performance edge AI processors and tensor processing units (TPUs). These devices run adaptive compression algorithms to dynamically adjust video quality based on network conditions. The video compression rate R is calculated as follows:

[0044] in, Indicates the original video size, Indicates the size of the compressed video. In this way, the system can effectively reduce the burden of data transmission while ensuring the quality of the video.

[0045] The central server cluster module is responsible for performing deep learning model training and large-scale data analysis. It includes high-performance CPU and GPU servers, quantum annealers, and large-capacity storage arrays. Quantum annealers are used to solve complex optimization problems, such as wind turbine layout optimization, which is of great significance to improving the overall efficiency of wind farms.

[0046] The communication network module ensures efficient and secure data transmission between the components of the system, including 5G private network, satellite communication backup and quantum communication network. The 5G private network realizes high-speed and low-latency data transmission, the satellite communication backup ensures the reliability of communication under extreme weather conditions, and the quantum communication network is used for data transmission with ultra-high security level, effectively preventing information leakage and hacker attacks.

[0047] The drone subsystem is used to perform large-scale inspections and close-range inspections, including fixed-wing long-endurance drones, multi-rotor drones, and drone charging stations. The flight path optimization of the drone is achieved using the ant colony algorithm (ACO). The calculation formula for the path length L is as follows:

[0048] in, Indicates the number of points that need to be visited. represents the distance from point i to point j. By optimizing the flight path, the system can maximize the inspection efficiency while minimizing energy consumption.

[0049] The multimodal deep learning module is the core of the system, which realizes the effective fusion of video, audio and sensor data. This module adopts a modal fusion architecture based on the Transformer architecture, introduces a cross-modal attention mechanism, and uses a self-supervised pre-training method to improve the generalization ability of the model. The calculation formula is as follows:

[0050] in, represents the attention score of the i-th modality, In this way, the system can adaptively adjust the importance of different modal data, thereby improving the accuracy of anomaly detection.

[0051] The spatiotemporal attention mechanism module enhances the system's ability to process time series data, including 3D convolutional neural networks, long short-term memory network (LSTM) variants, and graph attention networks (GAT). The combination of these technologies enables the system to effectively capture the spatiotemporal characteristics of video data, enhance the modeling capabilities of long time series data, and accurately model the spatial relationship and mutual influence between wind turbines.

[0052] The federated learning module effectively protects the data privacy of participants through secure aggregation protocols, asynchronous federated learning algorithms and model compression technology, while adapting to the network conditions and computing power of different wind farms. This distributed learning method enables the system to use data from multiple wind farms to improve model performance without sharing the original data.

[0053] The adaptive threshold algorithm module uses the dynamic time warping (DTW) algorithm, Gaussian mixture model (GMM) and online anomaly detection algorithm to achieve modeling of complex multimodal distribution and real-time adjustment of anomaly detection thresholds. This greatly improves the system's ability to identify various abnormal situations, including equipment failures, environmental anomalies and security threats.

[0054] The bio-inspired algorithm module draws on the wisdom of nature, including the ant colony optimization algorithm (ACO), the artificial bee colony algorithm (ABC), and the firefly algorithm. These algorithms are used to optimize the camera network layout and drone inspection paths, balance monitoring coverage and energy consumption, and quickly identify abnormal patterns. By imitating the intelligent behavior of biological groups, the system achieves efficient and flexible optimization and decision-making processes.

[0055] The quantum perception and quantum machine learning modules use quantum technology to improve the computing power and sensitivity of the system. The quantum sensor network improves detection sensitivity, quantum feature mapping enhances data feature expression, and the quantum-classical hybrid neural network implements nonlinear activation functions in quantum circuits. The application of these quantum technologies gives the system significant advantages in dealing with complex problems and massive data.

[0056] The self-healing network technology module improves the reliability and fault tolerance of the system through software-defined networking (SDN), blockchain consensus mechanism and reinforcement learning routing algorithm. SDN realizes intelligent network management and fault recovery, blockchain ensures the consistency and reliability of distributed systems, and reinforcement learning routing algorithm dynamically optimizes network traffic and load balancing.

[0057] In summary, the wind farm intelligent video monitoring and abnormal behavior identification system of the present invention has built a comprehensive, intelligent and reliable wind farm monitoring solution by comprehensively applying multiple innovative technologies. The system can not only detect various abnormal situations in real time, but also has predictive maintenance, adaptive learning and self-healing capabilities. Through innovations such as multimodal data fusion, the combination of edge computing and cloud computing, and the application of quantum technology, the system has achieved high-efficiency and high-reliability wind farm monitoring and abnormal behavior identification, significantly improving the operational efficiency and safety of wind farms. The application of this system will provide strong technical support for the sustainable development of the wind energy industry and promote the widespread application of clean energy.

[0058] Another object of the present invention is to provide a wind farm intelligent video monitoring and abnormal behavior recognition method, comprising: Collect data from wind farms; Receive and preliminarily process the collected wind farm data; Transmit the processed data of the wind farm; Receive transmitted data and perform deep learning model training and large-scale data analysis; Perform wide-range inspections and close-range inspections, and transmit collected data; The data of the wind farm includes: video image data, acoustic data, vibration data, stress distribution data, meteorological data, magnetic field data and gravity data.

[0059] The third object of the present invention is to provide an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for intelligent video monitoring and abnormal behavior identification of wind farms when executing the computer program.

[0060] The wind farm intelligent video monitoring and abnormal behavior identification method comprises the following steps: Collect data from wind farms; Receive and preliminarily process the collected wind farm data; Transmit the processed data of the wind farm; Receive transmitted data and perform deep learning model training and large-scale data analysis; Perform wide-range inspections and close-range inspections, and transmit collected data; The data of the wind farm includes: video image data, acoustic data, vibration data, stress distribution data, meteorological data, magnetic field data and gravity data.

[0061] A fourth object of the present invention is to provide a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the wind farm intelligent video monitoring and abnormal behavior identification method are implemented.

[0062] The wind farm intelligent video monitoring and abnormal behavior identification method comprises the following steps: Collect data from wind farms; Receive and preliminarily process the collected wind farm data; Transmit the processed data of the wind farm; Receive transmitted data and perform deep learning model training and large-scale data analysis; Perform wide-range inspections and close-range inspections, and transmit collected data; The data of the wind farm includes: video image data, acoustic data, vibration data, stress distribution data, meteorological data, magnetic field data and gravity data.

[0063] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0064] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0065] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0066] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A wind farm intelligent video monitoring and abnormal behavior recognition system, characterized in that: include: Multimodal acquisition module, used to collect data from wind farms; An edge computing module, used for receiving and preliminarily processing the wind farm data collected by the multi-modal acquisition module; A communication network module, used to transmit the wind farm data processed by the edge computing module to the central server cluster module; The central server cluster module is used to receive data transmitted by the communication network module and perform deep learning model training and large-scale data analysis; The drone subsystem is used to perform large-scale inspections and close-range inspections, and transmit the collected data to the central server cluster module through the communication network module; The data of the wind farm includes: video image data, acoustic data, vibration data, stress distribution data, meteorological data, magnetic field data and gravity data.

2. The wind farm intelligent video monitoring and abnormal behavior recognition system according to claim 1 is characterized in that: The video image data is collected by multiple sets of high-definition cameras to collect all-round video images of the wind turbine, as well as thermal imaging images for nighttime monitoring; The acoustic data collects noise levels and performs abnormal sound recognition through an acoustic sensor array; The vibration data is monitored and recorded by a vibration sensor; The stress distribution data is obtained by monitoring the stress distribution data of the tower and the blades through an optical fiber sensing network; The meteorological data includes wind speed data, temperature data, humidity data and air pressure data collected by meteorological sensors; The magnetic field data is detected by a quantum sensor network to detect tiny magnetic field change data of generators and transformers; The gravity data is data of foundation settlement and tower inclination monitored by a quantum gravimeter.

3. The wind farm intelligent video monitoring and abnormal behavior recognition system according to claim 1 is characterized in that: The edge computing module includes: High-performance edge AI processors to perform real-time data processing and preliminary analysis; Tensor processing units, used to accelerate deep learning tasks; The edge computing module runs an adaptive compression algorithm to dynamically adjust the video quality according to the real-time status of the communication network module to optimize data transmission efficiency; The communication network module comprises: 5G private network for high-speed and low-latency data transmission; Satellite communication backup to ensure communication reliability in extreme weather conditions; Quantum communication networks, which are used to provide ultra-high security levels for data transmission to prevent information leakage and hacker attacks; The central server cluster module includes: High-performance CPU and GPU servers for performing complex data analysis and deep learning model training; Quantum annealers, used to solve complex optimization problems; Large-capacity storage arrays are used to store and manage massive amounts of historical and real-time data.

4. The wind farm intelligent video monitoring and abnormal behavior recognition system according to claim 1 is characterized in that: The system also includes a data visualization platform for receiving and displaying the analysis results of the central server cluster; The system also includes a multimodal deep learning module, which is deployed on the central server cluster module and is used for a modal fusion architecture based on a Transformer architecture to achieve effective fusion of video data, audio data, and sensor data from the multimodal acquisition module; The system also includes a spatiotemporal attention mechanism module, which is deployed on a central server cluster and includes: 3D convolutional neural networks to capture the spatiotemporal features of video data; Long short-term memory network variants, used to enhance the modeling capabilities of long time series data; Graph attention network, used to model the spatial relationship and mutual influence between wind turbines.

5. The wind farm intelligent video monitoring and abnormal behavior recognition system according to claim 1 is characterized in that: The system also includes a federated learning module, which realizes data privacy protection and distributed learning by implementing a secure aggregation protocol, an asynchronous federated learning algorithm and a model compression technology between the edge computing module and the central server cluster module, so that the system can use data from multiple wind farms to improve model performance without sharing the original data.

6. The wind farm intelligent video monitoring and abnormal behavior recognition system according to claim 1 is characterized in that: The system also includes an adaptive threshold algorithm module, which is deployed on the edge computing module and the central server cluster module, and adopts a dynamic time warping algorithm, a Gaussian mixture model and an online anomaly detection algorithm to achieve real-time adjustment of the anomaly detection threshold and improve the system's ability to identify various abnormal situations.

7. The wind farm intelligent video monitoring and abnormal behavior recognition system according to claim 1 is characterized in that: The system further includes a biologically inspired algorithm module, which is deployed on the central server cluster module and includes: Ant colony optimization algorithm, used to optimize the camera network layout in the multimodal acquisition module and the inspection path of the drone subsystem; Artificial bee colony algorithm, used to perform multi-objective optimization in system operation and balance monitoring coverage and energy consumption; Firefly algorithm for rapid identification of anomalous patterns in multimodal data.

8. A method for intelligent video monitoring and abnormal behavior recognition in a wind farm, characterized in that: include: Collect data from wind farms; Receive and preliminarily process the collected wind farm data; Transmit the processed data of the wind farm; Receive transmitted data and perform deep learning model training and large-scale data analysis; Perform wide-range inspections and close-range inspections, and transmit collected data; The data of the wind farm includes: video image data, acoustic data, vibration data, stress distribution data, meteorological data, magnetic field data and gravity data.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for intelligent video monitoring and abnormal behavior identification of a wind farm as claimed in claim 8 when executing the computer program.

10. A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the method for intelligent video monitoring and abnormal behavior identification of a wind farm according to claim 8.

Images