Auxiliary video monitoring system for operation and maintenance work of booster station

Abstract

The invention relates to the technical field of booster station monitoring systems, and discloses a booster station operation and maintenance work video auxiliary monitoring system comprising a three-dimensional perception interaction layer used for collecting equipment state data and providing a man-machine interaction interface and connected with an edge intelligent fusion layer through a wired or wireless transmission channel; and the edge intelligent fusion layer is used for locally analyzing and caching the state data. Through standardized interface interconnection of the three-dimensional perception interaction layer, the edge intelligent fusion layer and the cloud collaborative management layer, a collaborative working mode of data acquisition, local analysis and cloud optimization is formed, the architecture design breaks through the dependence of a traditional scheme on cloud centralized processing, data preprocessing sinks to the edge layer, the network transmission pressure is effectively reduced, and the network performance is improved. The local intelligent analysis capability can still be maintained when the network fluctuates, and compared with the delay problem caused by full data uploading in the prior art, the system response real-time performance and the availability under the network disconnection scene are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of booster station monitoring systems, and in particular to a video-assisted monitoring system for booster station operation and maintenance work. Background Art

[0002] As the core hub for achieving voltage level increases in power systems, booster stations' technical architecture and operation and maintenance models are directly related to the stability and economic efficiency of power grid transmission. These facilities are typically equipped with multiple power transformers, which use electromagnetic induction to gradually boost low-voltage electricity (35kV or below) from the generator to higher voltages such as 110kV, 220kV, or even 500kV. This significantly reduces current in long-distance transmission lines, effectively minimizing energy losses caused by line impedance according to Joule's law, and expanding the power transmission radius to hundreds of kilometers.

[0003] Electrical equipment within the station can be divided into primary and secondary equipment based on their functions. Primary equipment includes SF6 circuit breakers with arc extinguishing capacities of thousands of amperes, transformers for current and voltage conversion, and lightning arresters to protect electrical equipment from lightning strikes. The SF6 pressure gauge for a 500kV circuit breaker is typically installed 2.5 meters above the ground, while the pressure gauge for a 220kV current transformer is over 4 meters. Secondary equipment includes relay protection devices, measuring instruments, and automated control systems, which together form a complex power conversion network. The operation and maintenance of these high-altitude equipment present unique challenges. Traditional manual inspections require operators to climb to the height of the equipment with tools and perform close-range work within the narrow structure. Donning safety gear and performing pre-ascendancy checks alone can take 20-30 minutes, and the work presents multiple risks, such as falls and electric shock.

[0004] In the existing operation and maintenance technology system, although telescopic observation can avoid some high-altitude operations, it has significant limitations due to optical principles: under direct strong light, the reflectivity of the pressure gauge glass surface is high, which also leads to an increase in the error rate of the pointer reading; the air turbulence in the outdoor environment in winter will reduce the image clarity, and the error of the pointer reading using a telescope will be significantly increased. The more prominent problem is that the entire process of a single inspection from equipment positioning, optical focusing to data recording is time-consuming, and the corresponding labor costs will also increase. The efficiency bottleneck of the traditional method has become a key factor restricting the lean operation and maintenance of the substation. With the intelligent transformation of the power grid, how to break through the safety and efficiency bottlenecks of high-altitude equipment inspections through technological innovation has become an important research direction in the field of power operation and maintenance. Therefore, those skilled in the art have proposed a video-assisted monitoring system for the operation and maintenance of substations to solve the above-mentioned technical problems. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a video-assisted monitoring system for the operation and maintenance of a substation, which solves the problem that the existing observation method is easily affected by the external environment.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a video-assisted monitoring system for booster station operation and maintenance work, comprising The 3D perception interaction layer is used to collect device status data and provide a human-computer interaction interface. This layer is connected to the edge intelligent fusion layer through wired or wireless transmission channels. The edge intelligent fusion layer is used to analyze and cache state data locally and establish a two-way communication link with the cloud collaborative management layer through the network communication module; Cloud-based collaborative management layer for distributed model optimization and remote visualization of device status; Among them, the three-dimensional perception interaction layer, edge intelligent fusion layer and cloud collaborative management layer realize data interaction through standardized interface protocols, and the interface protocols include MQTT and RESTful API.

[0007] Preferably, the three-dimensional perception interaction layer includes: The detachable multispectral sensor module integrates a visible light camera, an infrared thermal imaging unit, and a laser ranging sensor. It is physically connected to the edge intelligent fusion layer through a quick-connect interface at the top of the insulating operating rod to obtain image, temperature, and spatial distance data of the device. Wearable interactive terminals communicate with the edge intelligent fusion layer via WiFi or Bluetooth protocols to display data in real time and receive control instructions.

[0008] Preferably, the detachable multi-spectral probe module includes an environment adaptive monitoring unit, which includes: Light and dust sensors, used to detect ambient light intensity and dust concentration; An electromechanical polarized light filter switching mechanism, electrically connected to the light sensor, is used to automatically adjust the filter angle according to light intensity to eliminate reflections; The micro air pump is electrically connected to the dust sensor and is used to trigger the lens cleaning action according to the dust concentration.

[0009] Preferably, the wearable interactive terminal integrates a nine-axis inertial sensor for recognizing gestures and mapping them into control instructions for adjusting the focal length of the visible light camera or switching spectral modes.

[0010] Preferably, the edge intelligent fusion layer includes: A low-power edge computing module with a built-in lightweight image recognition engine for local identification of meter pointer positions and appearance defects. The adaptive power supply unit uses a lithium battery and solar panel parallel power supply structure to support continuous operation during network outages.

[0011] Preferably, the low-power edge computing module sets a multi-level power consumption mode. When the abnormality detection confidence of the image recognition engine exceeds a preset threshold, it automatically switches to the enhanced computing power mode and activates the infrared thermal imaging unit for secondary data collection.

[0012] Preferably, the edge intelligent fusion layer includes a data cache unit, which stores status data in timestamp order and packages the data into blockchain blocks when the network is interrupted. The blocks include a timestamp, a previous block hash, and a data hash value.

[0013] Preferably, the cloud collaborative management layer includes: The federated learning scheduling module uses a distributed training framework, where each edge node uploads anonymized defect features and aggregates them to generate a global model. A lightweight digital twin engine that builds a 3D model of the device based on WebGL technology and maps status data in real time for visualization.

[0014] Preferably, the video-assisted monitoring system for the operation and maintenance of the substation is characterized in that it also includes a multi-device self-organizing network communication module, which builds a Mesh network based on the LoRaWAN protocol. When the network is interrupted, the probe of the three-dimensional perception interaction layer and the wearable terminal form a multi-hop communication link through the edge node, and cache offline data through the blockchain mechanism.

[0015] Preferably, the multi-device ad hoc network communication module sets up a master-slave node authentication mechanism, verifies the device identity through a pre-shared key, and uploads offline data to the cloud collaborative management layer in block order when the network is restored.

[0016] The present invention provides a video-assisted monitoring system for booster station operation and maintenance, which has the following beneficial effects: 1. The present invention constructs a three-layer heterogeneous architecture system, which forms a collaborative working mode of data collection, local analysis and cloud optimization through the standardized interface interconnection of the three-dimensional perception interaction layer, the edge intelligent fusion layer and the cloud collaborative management layer. This architectural design breaks through the traditional solution's reliance on centralized cloud processing, sinks data preprocessing to the edge layer, effectively reduces network transmission pressure, and can still maintain local intelligent analysis capabilities during network fluctuations. Compared with the delay problem caused by full data upload in the existing technology, it significantly improves the system's real-time response and availability in network disconnection scenarios.

[0017] 2. The detachable multi-spectral probe module in the present invention integrates visible light, infrared thermal imaging and laser ranging units, and cooperates with the environmental adaptive adjustment mechanism to build a multi-dimensional equipment status acquisition system. This processing method enables operation and maintenance personnel to remotely obtain equipment images, temperature and spatial distance data through an insulated operating rod on the ground. The influence of strong light reflection is eliminated with the help of a polarized light filter, and the lens self-cleaning is achieved using a micro air pump. Compared with the optical limitations of traditional telescope inspections and the safety hazards of high-altitude operations, this not only improves the accuracy of meter recognition to a new level, but also fundamentally avoids the risks of high-altitude operations.

[0018] 3. The cloud-based collaborative management layer of this invention incorporates a federated learning framework and a lightweight digital twin engine to build a distributed intelligent optimization and 3D visualization system. This system aggregates models by uploading anonymized defect features from edge nodes, avoiding the data privacy risks of centralized training. It also uses WebGL technology to compress 3D models, reducing hardware deployment costs. This processing combination differs from the high barriers to entry of existing heavy-duty twin platforms. While ensuring model generalization capabilities, it enables remote operations and maintenance (O&M) to achieve immersive visual monitoring of device status through standard mobile terminals.

[0019] 4. The multi-device self-organizing network communication module constructed based on the LoRaWAN protocol of the present invention is combined with the blockchain caching mechanism to form a distributed communication system that adapts to complex scenarios. This processing method automatically switches to Mesh relay mode when the network is interrupted, maintains data transmission through multi-hop communication, and uses blockchain technology to ensure the integrity and traceability of offline data. Compared with traditional transmission solutions that rely on fixed networks, it effectively solves the network coverage problem of remote substations. In network disconnection scenarios such as emergency repairs, it can still ensure the continuous recording and subsequent synchronization of operation and maintenance data, thereby improving the system's environmental adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a schematic diagram of the overall system framework of the present invention; Figure 2 This is a schematic diagram of the framework of the three-dimensional perception interaction layer of the present invention; Figure 3 This is a schematic diagram of the edge intelligent fusion layer framework architecture of the present invention; Figure 4 This is a schematic diagram of the cloud collaborative management layer framework of the present invention; Figure 5 This is a schematic diagram of the framework architecture of the multi-device ad hoc network communication module of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0022] Please see the attached Figure 1 -Attached Figure 5 The embodiment of the present invention provides a video-assisted monitoring system for booster station operation and maintenance, including The 3D perception interaction layer is used to collect device status data and provide a human-computer interaction interface. This layer is connected to the edge intelligent fusion layer through wired or wireless transmission channels. The edge intelligent fusion layer is used to analyze and cache state data locally and establish a two-way communication link with the cloud collaborative management layer through the network communication module; Cloud-based collaborative management layer for distributed model optimization and remote visualization of device status; Among them, the three-dimensional perception interaction layer, edge intelligent fusion layer and cloud collaborative management layer realize data interaction through standardized interface protocols, including MQTT and RESTful API.

[0023] The video-assisted monitoring system for substation operations and maintenance, provided by this invention, utilizes an innovative three-layer heterogeneous architecture. Through the organic collaboration of a three-dimensional perception interaction layer, an edge intelligence fusion layer, and a cloud-based collaborative management layer, it establishes a closed-loop operations and maintenance system encompassing "on-site data collection - real-time edge analysis - cloud-based intelligent optimization." The three-dimensional perception interaction layer is centered around a detachable multispectral probe module and a wearable interactive terminal. The former integrates visible light imaging, infrared thermal imaging, and laser ranging units, enabling contactless data collection from high-altitude equipment via an insulated operating lever. The latter utilizes a lightweight wristband design and converts gesture commands into device control signals via a nine-axis inertial sensor, mitigating the operational risks of traditional manual ascents.

[0024] The edge intelligent fusion layer deploys low-power edge computing modules and adaptive power supply units. Based on the ARM+DSP heterogeneous processor architecture, it implements local computing tasks such as meter identification and defect detection, avoiding transmission delays when uploading full data to the cloud. At the same time, a hybrid power supply solution of lithium batteries and solar panels is used to ensure the continuous operation of remote booster stations.

[0025] The system's architectural design transcends the traditional heavy-duty approach of industrial monitoring equipment, integrating lightweight technologies from the consumer electronics industry with the demands of industrial monitoring. The multispectral probe module in the 3D sensing interaction layer utilizes a lightweight aluminum alloy housing, keeping the overall weight within a comfortable handheld range. Combined with an electromechanical polarization filter switching mechanism and a micro-air pump cleaning unit, it automatically adapts to complex environments such as strong light and dust, significantly improving data acquisition stability compared to traditional telescopic inspections.

[0026] The edge intelligence fusion layer incorporates a lightweight deep learning engine, leveraging lightweight network architectures like MobileNet to enable local identification of equipment defects. This avoids the high bandwidth requirements and processing latency of traditional cloud-based servers, making it particularly suitable for outdoor booster stations with poor network coverage. The cloud-based collaborative management layer combines a federated learning framework with a lightweight digital twin engine to build a distributed intelligent optimization system. Each edge node only uploads anonymized defect features for model aggregation, protecting data privacy while improving model generalization capabilities. This also enables 3D visualization and remote monitoring of equipment status.

[0027] The system's core innovation lies in the organic integration of cross-domain technologies to form a new O&M model that balances safety and efficiency. The 3D perception interaction layer solves the challenge of contactless data collection for high-altitude equipment, enabling operators to obtain images, temperature, and spatial location information without having to climb. The edge intelligence fusion layer ensures real-time and complete data processing through local computing and offline caching. The cloud-based collaborative management layer leverages federated learning and digital twin technologies to achieve the transition from single-point device monitoring to full-site intelligent O&M.

[0028] This three-tier architecture design not only breaks through the multiple bottlenecks of traditional solutions, such as high risks of high-altitude operations, strong reliance on data transmission, and low efficiency of intelligent analysis, but also enables flexible expansion of each module through standardized interface protocols, providing a feasible technical path for the digital transformation of substation operations and maintenance.

[0029] The 3D perception interaction layer includes: The detachable multispectral sensor module integrates a visible light camera, an infrared thermal imaging unit, and a laser ranging sensor. It is physically connected to the edge intelligent fusion layer through a quick-connect interface at the top of the insulating operating rod to obtain image, temperature, and spatial distance data of the device. The detachable multi-spectral probe module is installed on the top of the operating rod through an insulated threaded interface. The shell is made of lightweight aluminum alloy, and the overall weight is controlled within a range suitable for handholding. The detachable multi-spectral probe module is internally integrated with: Visible light imaging unit: uses a high-resolution CMOS sensor and an optical zoom lens to clearly capture details of high-altitude equipment; Infrared thermal imaging unit: uses uncooled vanadium oxide detectors to achieve real-time monitoring of the equipment temperature field; Laser ranging module: obtains the spatial distance parameters of the device through the time-of-flight principle, providing data support for subsequent 3D modeling.

[0030] The detachable multispectral probe module contains an environmental adaptive monitoring unit, which includes: Light and dust sensors, used to detect ambient light intensity and dust concentration; Light and dust sensors monitor environmental parameters in real time. When light intensity exceeds a preset threshold, the electromechanical polarized light filter switching mechanism automatically adjusts the filter angle to eliminate reflections on the instrument surface. When the dust concentration exceeds the standard, the micro air pump starts the pulse cleaning function to ensure the cleanliness of the lens surface and maintain the image acquisition quality.

[0031] An electromechanical polarized light filter switching mechanism, electrically connected to the light sensor, is used to automatically adjust the filter angle according to light intensity to eliminate reflections; The micro air pump is electrically connected to the dust sensor and is used to trigger the lens cleaning action according to the dust concentration.

[0032] The wearable interactive terminal integrates a nine-axis inertial sensor to recognize gestures and map them into control instructions for adjusting the focus of the visible light camera or switching spectral modes.

[0033] Wearable interactive terminals communicate with the edge intelligent fusion layer via WiFi or Bluetooth protocols to display data in real time and receive control instructions.

[0034] The wearable interactive terminal features an ergonomic wristband design, integrated with a touchscreen display and a nine-axis inertial sensor. Specifically, the touchscreen displays multispectral data collected by the probe in real time and supports gestures for zooming, focusing, and other operations. The inertial sensor uses an attitude calculation algorithm to map the operator's gestures into control commands, enabling remote adjustment of the probe's focal length and spectral mode. The communication module supports WiFi and Bluetooth dual-mode connections to ensure real-time data transmission with the probe and edge computing module.

[0035] The edge intelligence fusion layer includes: A low-power edge computing module with a built-in lightweight image recognition engine for local identification of meter pointer positions and appearance defects. The low-power edge computing module sets a multi-level power consumption mode. When the anomaly detection confidence of the image recognition engine exceeds the preset threshold, it automatically switches to the enhanced computing power mode and activates the infrared thermal imaging unit for secondary data collection.

[0036] The low-power edge computing module adopts a heterogeneous multi-core processor architecture and integrates an FPGA hardware acceleration unit, specifically including: Lightweight image recognition engine: Based on a deep learning lightweight network architecture, it preprocesses and extracts features from visible light images to achieve meter pointer positioning and equipment defect identification; Multi-level power management mechanism: Dynamically adjusts the processor frequency based on the identification task load, reduces power consumption in normal inspection mode, and increases computing power during abnormal review, ensuring efficient use of computing resources; Multi-source data association processing: When an anomaly is detected in the visible light image, the infrared thermal imaging and laser ranging modules are automatically triggered to form an "image-temperature-distance" association data matrix.

[0037] The adaptive power supply unit uses a lithium battery and solar panel parallel power supply structure to support continuous operation during network outages.

[0038] The edge intelligent fusion layer includes a data cache unit that stores status data in timestamp order. When the network is interrupted, the data is packaged into blockchain blocks, which contain timestamps, previous block hashes, and data hash values.

[0039] The adaptive power supply unit utilizes a lithium battery and flexible solar panels in parallel, with energy management implemented through an MPPT controller to meet the continuous power requirements of field operations. The data cache unit is equipped with a large-capacity storage chip and uses a circular buffering strategy to store data frames. When the network is interrupted, the blockchain cache mechanism is activated, packaging the data into blocks in chronological order. Each block includes a timestamp, the previous block hash, and the data hash value, ensuring the integrity and traceability of offline data.

[0040] The cloud collaborative management layer includes: The federated learning scheduling module uses a distributed training framework, where each edge node uploads anonymized defect features and aggregates them to generate a global model. The federated learning scheduling module builds a distributed model training framework. The specific workflow is as follows: The edge node fine-tunes the initial model based on the local defect dataset and only updates the classification layer parameters; The nodes upload parameter differences to the cloud, which uses a federated averaging algorithm to aggregate and generate a global model while protecting data privacy through differential privacy technology. Parameter sparsification technology is used to reduce the amount of data transmission and achieve efficient iterative optimization of the model.

[0041] A lightweight digital twin engine that builds a 3D model of the device based on WebGL technology and maps status data in real time for visualization.

[0042] The lightweight digital twin engine is developed based on WebGL technology, enabling lightweight construction of equipment 3D models and real-time data mapping: Use geometric compression algorithms to optimize 3D models to ensure smooth rendering on mobile devices; Establish a mapping relationship between physical devices and digital twins, and complete model registration through lidar point cloud data; Receive status data uploaded by edge nodes in real time, and visualize it on the twin in the form of heat maps, numerical annotations, etc., supporting remote interactive viewing.

[0043] The video-assisted monitoring system for the operation and maintenance of the substation also includes a multi-device self-organizing network communication module, which builds a mesh network based on the LoRaWAN protocol. When the network is interrupted, the probes of the three-dimensional perception interaction layer and the wearable terminals form multiple communication links through the edge nodes, and cache offline data through the blockchain mechanism.

[0044] The multi-device ad hoc network communication module sets up a master-slave node authentication mechanism, verifies the device identity through a pre-shared key, and uploads offline data to the cloud collaborative management layer in block order when the network is restored.

[0045] The multi-device ad hoc network communication module builds a star-shaped Mesh hybrid network based on the LoRaWAN protocol, specifically including: Network topology design: The edge computing module serves as the gateway node, and the wearable terminal and probe module serve as the terminal nodes, supporting multi-hop relay communication; Security authentication mechanism: Device identity authentication is achieved through pre-shared keys to prevent unauthorized access; Network disconnection caching strategy: When the network is interrupted, data is transmitted to the edge node through the Mesh network and stored as blockchain blocks. After the network is restored, it is uploaded in sequence to ensure that data is not lost.

[0046] The overall system workflow can be divided into the following stages: Phase 1: Initialization and environment adaptation The operator installs the sensor module onto the insulated operating rod and pairs it with the terminal via NFC. The environmental adaptation unit automatically detects light and dust parameters, triggering polarization filter adjustment and lens cleaning to ensure optimal operating conditions for the data collection device. The edge computing module loads the local recognition model and enters normal inspection power consumption mode.

[0047] The second stage: real-time monitoring and exception handling The probe scans the equipment along a preset trajectory, transmitting multispectral data in real time to the edge computing module. If a meter pointer deviation or a device appearance defect is detected, the system automatically switches to abnormality review mode, activating infrared thermal imaging and laser ranging for secondary data collection. The wearable terminal alerts maintenance personnel with audio and visual notifications, highlighting the abnormal area, and prioritizes caching the abnormal data locally.

[0048] The third stage: collaborative operation and maintenance and model update When multiple people are working, each terminal shares video streams via an ad hoc network communication module, enabling real-time synchronization of abnormal information. A cloud-based federated learning module regularly aggregates training data from edge nodes, updates the global model, and distributes it to each node. The digital twin engine dynamically updates the status characteristics of the 3D equipment model based on the latest inspection data, maintaining consistency between the virtual and physical devices.

[0049] Phase 4: Hardware Deployment and Compatibility The insulating operating rod is made of epoxy resin fiberglass, meeting the insulation performance requirements in high-voltage environments. The edge computing module uses a high-protection chassis and is installed in the booster station's main control room to ensure electromagnetic compatibility. The cloud server adopts a distributed architecture, supporting simultaneous access and data processing from multiple sites. All hardware components are interconnected through standardized interface protocols, ensuring system scalability and compatibility.

[0050] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. Video-assisted monitoring system for booster station operation and maintenance, characterized by: include The 3D perception interaction layer is used to collect device status data and provide a human-computer interaction interface. This layer is connected to the edge intelligent fusion layer through wired or wireless transmission channels. The edge intelligent fusion layer is used to analyze and cache state data locally and establish a two-way communication link with the cloud collaborative management layer through the network communication module; Cloud-based collaborative management layer for distributed model optimization and remote visualization of device status; Among them, the three-dimensional perception interaction layer, edge intelligent fusion layer and cloud collaborative management layer realize data interaction through standardized interface protocols, and the interface protocols include MQTT and RESTful API.

2. The video-assisted monitoring system for booster station operation and maintenance according to claim 1 is characterized in that: The three-dimensional perception interaction layer includes: The detachable multispectral sensor module integrates a visible light camera, an infrared thermal imaging unit, and a laser ranging sensor. It is physically connected to the edge intelligent fusion layer through a quick-connect interface at the top of the insulating operating rod to obtain image, temperature, and spatial distance data of the device. Wearable interactive terminals communicate with the edge intelligent fusion layer via WiFi or Bluetooth protocols to display data in real time and receive control instructions.

3. The video-assisted monitoring system for booster station operation and maintenance according to claim 2 is characterized in that: The detachable multi-spectral probe module includes an environment adaptive monitoring unit, which includes: Light and dust sensors, used to detect ambient light intensity and dust concentration; An electromechanical polarized light filter switching mechanism, electrically connected to the light sensor, is used to automatically adjust the filter angle according to light intensity to eliminate reflections; The micro air pump is electrically connected to the dust sensor and is used to trigger the lens cleaning action according to the dust concentration.

4. The video-assisted monitoring system for booster station operation and maintenance according to claim 2 is characterized in that: The wearable interactive terminal integrates a nine-axis inertial sensor for recognizing gestures and mapping them into control instructions for adjusting the focus of the visible light camera or switching spectral modes.

5. The video-assisted monitoring system for operation and maintenance of a booster station according to claim 1 is characterized in that: The edge intelligent fusion layer includes: A low-power edge computing module with a built-in lightweight image recognition engine for local identification of meter pointer positions and appearance defects. The adaptive power supply unit uses a lithium battery and solar panel parallel power supply structure to support continuous operation during network outages.

6. The video-assisted monitoring system for booster station operation and maintenance according to claim 5 is characterized in that: The low-power edge computing module sets a multi-level power consumption mode. When the abnormality detection confidence of the image recognition engine exceeds the preset threshold, it automatically switches to the enhanced computing power mode and activates the infrared thermal imaging unit for secondary data collection.

7. The video-assisted monitoring system for booster station operation and maintenance according to claim 5 is characterized in that: The edge intelligent fusion layer includes a data cache unit, which stores status data in timestamp order and packages the data into blockchain blocks when the network is interrupted. The blocks include a timestamp, a previous block hash, and a data hash value.

8. The video-assisted monitoring system for booster station operation and maintenance according to claim 1 is characterized in that: The cloud collaborative management layer includes: The federated learning scheduling module uses a distributed training framework, where each edge node uploads anonymized defect features and aggregates them to generate a global model. A lightweight digital twin engine that builds a 3D model of the device based on WebGL technology and maps status data in real time for visualization.

9. The video-assisted monitoring system for operation and maintenance of a booster station according to any one of claims 1 to 8, characterized in that: It also includes a multi-device self-organizing network communication module, which builds a Mesh network based on the LoRaWAN protocol. When the network is interrupted, the probes of the three-dimensional perception interaction layer and the wearable terminal form multiple communication links through the edge nodes, and cache offline data through the blockchain mechanism.

10. The video-assisted monitoring system for operation and maintenance of a booster station according to claim 9, characterized in that: The multi-device self-organizing network communication module sets up a master-slave node authentication mechanism, verifies the device identity through a pre-shared key, and uploads offline data to the cloud collaborative management layer in block order when the network is restored.

Images