Beidou geological disaster displacement and image video monitoring fusion method and system based on high-reliability communication link

Abstract

The invention provides a Beidou geological disaster displacement and image video monitoring fusion method and system based on a high-reliability communication link, and the method comprises the steps: collecting displacement, inclination angle and five-way high-definition video in real time, accessing microclimate and wire clamp temperature data, and carrying out the unified packaging through a data gateway; aPN special line transmission is adopted, and centimeter-level RTK differential service is provided for the unmanned aerial vehicle; based on a Beidou VRS virtual reference station and a long baseline monitoring resolving technology, a Beidou ground-based enhanced station network is multiplexed, and a virtual observation value is generated for real-time dynamic RTK resolving; inverting a three-dimensional water vapor field based on oblique path troposphere delay PWV, realizing tower-level rainfall prediction, and generating a risk assessment report by fusing an A I algorithm; superposing a displacement thermodynamic diagram, a real-time video and an early warning label in a BIM + GI digital twinning scene, and supporting 360-degree panoramic stitching and alarm trigger snapshot; and carrying out intelligent early warning decision. According to the invention, through high-precision Beidou monitoring, multi-modal data fusion and intelligent early warning, the real-time performance, precision and automation level of geological disaster monitoring are significantly improved.

Description

Technical Field

[0001] The present application relates to the field of geological disaster monitoring technology, and specifically to a disaster warning method that combines Beidou high-precision positioning, multi-mode communication links, and visual analysis, which is particularly suitable for real-time monitoring in complex environments such as mountainous areas and landslides. Background Art

[0002] Geological disasters (such as landslides, collapses, and ground subsidence) are characterized by suddenness and destructiveness, seriously threatening the safety of people's lives and property and the stable operation of major infrastructure. Traditional monitoring methods (such as total stations and inclinometers) have the following technical bottlenecks:

[0003] Insufficient monitoring accuracy: Conventional GNSS single-point positioning accuracy is only at the meter level, which is difficult to meet the needs of millimeter-level deformation monitoring; manual inspections are inefficient and cannot achieve all-weather continuous monitoring. Difficulty in fusing multi-source data: Displacement, video, meteorological and other data are scattered and independent, lacking a unified spatiotemporal benchmark, making collaborative analysis difficult; existing systems mostly use a single sensor and cannot fully capture the multi-dimensional precursor characteristics of geological disasters. Insufficient real-time and reliability: Remote areas have poor communication conditions and high data return delays, which affect the timeliness of early warnings; traditional RTK requires dense deployment of physical base stations, which has high construction and maintenance costs. Weak intelligent decision-making capabilities: Relying on manual experience to judge risks, the false alarm and missed alarm rate is high; early warning and emergency response are disconnected, and there is a lack of automated handling mechanisms.

[0004] Therefore, there is an urgent need for a high-precision, low-cost monitoring method that can achieve real-time early warning and automated response to geological disasters through Beidou high-precision positioning, multimodal data fusion and intelligent decision-making. Summary of the Invention

[0005] In view of this, it is necessary to provide a Beidou geological disaster displacement and image video monitoring fusion method based on a highly reliable communication link. Through high-precision Beidou monitoring, multimodal data fusion and intelligent early warning, the real-time, accuracy and automation level of geological disaster monitoring can be significantly improved. It is suitable for high-risk scenarios such as transmission towers and slopes.

[0006] The present application provides a method and system for integrating Beidou geological disaster displacement and image video monitoring based on a highly reliable communication link, including: multi-source data acquisition: real-time acquisition of millimeter-level displacement, tilt and five-way high-definition video through Beidou monitoring devices, access to micro-meteorological and wire clamp temperature data, and unified packaging through a data gateway; highly reliable data transmission: using APN dedicated line transmission, marking Beidou timing timestamps to ensure timing alignment, and providing centimeter-level RTK differential services for drones; virtual base station construction: based on Beidou VRS virtual reference station and long baseline monitoring solution technology, reuse Beidou The ground-based augmented station network generates virtual observation values ​​for real-time dynamic RTK solution; cloud-based intelligent analysis: based on the oblique path tropospheric delayed PWV inversion of the three-dimensional water vapor field, it realizes tower-level rainfall prediction and integrates AI algorithms to generate risk assessment reports; multimodal monitoring: superimposes displacement heat maps, real-time video and warning annotations in the BIM+GIS digital twin scene, supports 360° panoramic stitching and alarm trigger snapshots; intelligent early warning decision-making: through big data and AI real-time analysis, automatically pushes alarms when exceeding thresholds or identifying collapse precursors, and initiates drone emergency inspections. This invention significantly improves the real-time, accuracy and automation level of geological disaster monitoring through high-precision Beidou monitoring, multimodal data fusion and intelligent early warning, and is suitable for high-risk scenarios such as transmission towers and slopes.

[0007] In a first aspect, an embodiment of the present application provides a Beidou geological disaster displacement and image and video monitoring fusion method based on a highly reliable communication link, the method comprising:

[0008] Step 1: Multi-source data collection: real-time acquisition of millimeter-level displacement, tilt angle, and five-directional high-definition video through Beidou monitoring devices; access to micro-meteorological and wire clamp temperature data, and unified packaging through the data gateway;

[0009] Step 2: High-reliability data transmission: Using APN dedicated lines for transmission, all data is marked with Beidou timing timestamps to ensure timing alignment; through the temporary base station mode, centimeter-level RTK differential services are provided for drones;

[0010] Step 3: Based on the BeiDou VRS virtual reference station and long baseline monitoring solution technology, the completed BeiDou ground-based augmentation station network is reused to build a virtual reference station to replace the traditional physical base station. Using the observation data of multiple reference stations, an error correction model is established to generate virtual observation values ​​for real-time dynamic RTK solution.

[0011] Step 4: Cloud-based intelligent analysis uses the slant-path tropospheric delayed PWV to invert the three-dimensional water vapor field and achieve tower-level rainfall prediction. An AI algorithm integrates displacement rate, inclination mutation, and video features to generate a risk assessment report.

[0012] Step 5: Multimodal monitoring: Using a 3D visualization platform, displacement heat maps, real-time video, and warning area annotations are superimposed on the BIM+GIS digital twin scene. Through 360° panoramic stitching and downward-looking wide-angle monitoring, alarms trigger automatic video capture.

[0013] Step 6: Intelligent early warning and decision-making: Through big data analysis and artificial intelligence algorithms, the monitoring data is solved and trended in real time to generate early warning information and automatically trigger the alarm mechanism. When the displacement rate exceeds the threshold or AI identifies the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is simultaneously launched.

[0014] Optionally, in an implementation of the first aspect of the present invention, step 1: multi-source data acquisition, real-time acquisition of millimeter-level displacement, tilt angle, and five-directional high-definition video through a Beidou monitoring device; access to micro-meteorological and wire clamp temperature data, and unified packaging via a data gateway, includes:

[0015] Beidou geological disaster monitoring integrated device is deployed on towers and slopes to collect millimeter-level high-precision displacement settlement, inclination, image and video data in real time. The Beidou geological disaster monitoring integrated device is based on GNSS high-precision receiver positioning technology, low-power high-precision GNSS chip and five channel cameras in front, back, left, right and bottom. It can monitor and stitch dual-channel 180° panoramic high-definition images, and integrate MEMS sensor technology.

[0016] The integration of narrowband communication technology and embedded system sleep and wake-up technology can provide millimeter-level high-precision displacement monitoring information and 360° panoramic image video monitoring around the clock;

[0017] The Beidou geological disaster monitoring integrated device also has a data gateway function, which can access data from various sensors such as micro-meteorology and wire clamp temperature measurement, and can be selected and applied according to the monitoring scenario;

[0018] The data is transmitted back to the unified IoT platform through the IoT APN dedicated line. Based on the positioning and mapping capabilities of the Beidou platform and the China Southern Power Grid Zhikan platform, multi-source data fusion applications are carried out through the system during operation to support digital production applications.

[0019] Optionally, in an implementation of the first aspect of the present invention, step 3: based on the Beidou VRS virtual reference station and long baseline monitoring solution technology, reusing the completed Beidou ground-based augmentation station network, building a virtual reference station to replace the traditional physical base station, and establishing an error correction model through observation data from multiple reference stations to generate virtual observation values ​​for real-time dynamic RTK solution, including:

[0020] Step 3.1: Data collection is performed through a physical reference station network. Existing BeiDou ground-based augmentation stations must be reused and meet the following requirements: station spacing ≤ 30 km; dual-frequency or multi-frequency BeiDou receivers must be used; data sampling rate ≥ 1 Hz; raw observation values ​​must be transmitted back in real time; and inter-station time synchronization must be achieved using the PTP protocol.

[0021] Step 3.2: Obtain the error model through network RTK error modeling, where the carrier phase

[0022] The formula for the measurement is:

[0023]

[0024] in, is the carrier phase observation value, R is the distance from the receiver to the satellite, I is the ionospheric delay, T is the tropospheric delay, c is the speed of light, δt r and δt s are the clock errors of the receiver and the satellite respectively, O is the satellite orbit error, M is the multipath effect error, λ is the carrier phase wavelength, and N is the integer ambiguity;

[0025] Step 3.3: Through Geometry-Free combined extraction, regional linear interpolation or low-order surface fitting is used to correct the error of the observation value of the reference station. The formula for the corrected carrier phase observation value is:

[0026]

[0027] Among them, I cor and T cor are the correction values ​​for ionospheric and tropospheric delays, δt r,cor and δt s,cor are the correction values ​​of receiver and satellite clock errors respectively;

[0028] Step 3.4: Based on the observations of the main reference station, combined with the geometric correlation terms and atmospheric delay correction, the carrier phase observations of the virtual reference station are generated. The formula is as follows:

[0029]

[0030] in, is the carrier phase observation, is the observation value of the subjective station, is the geometric distance error gradient, is the atmospheric delay error gradient.

[0031] Optionally, in an implementation of the first aspect of the present invention, it is characterized in that the error model is corrected using a LAMBDA algorithm, including:

[0032] Step 4.1: Error modeling:

[0033] Step 4.1.1: Ionospheric delay: extract the ionospheric delay through dual-frequency geometry-free combination and establish a low-order polynomial regional spatial model. The ionospheric delay formula is:

[0034]

[0035] Where I is the ionospheric delay value, f1 and f2 are the frequencies of the two frequency points respectively, and are the carrier phase observation values ​​in weeks, λ1 and λ2 are the carrier phase wavelengths corresponding to the carrier phase observation values;

[0036] Step 4.1.2: Tropospheric delay: The dry component is corrected by the Saastamoinen model, and the wet component is estimated by a random walk process;

[0037] Step 4.1.3: Orbit / clock error: Eliminate using precise ephemeris products or inter-station differentials;

[0038] Step 4.2: LAMBDA ambiguity is fixed;

[0039] Step 4.2.1: Q N Perform integer Gaussian transform to reduce the correlation between ambiguities. The transformation formula is:

[0040] Z T Q N Z=D,

[0041] Among them, Q N is the initial covariance matrix, D is a diagonal matrix, Z is the descending correlation matrix, |Z|=1;

[0042] Blur after transformation:

[0043] Step 4.2.2: Search for the optimal integer solution in the transformed space by integer least squares search Use the shrinking search strategy Bootstrapping to accelerate calculation;

[0044] Step 4.2.3: Inverse transformation and verification, Inverse transform back to the original space: The fixed reliability was verified by the ratio test RatioTes;

[0045] Step 4.3: Fix the solution error correction and fix the ambiguity Substitute into the observation equation and inversely calculate the residual term;

[0046] Step 4.4: Use the residuals to correct the ionosphere / troposphere model parameters and update the covariance matrix QN , enter the next epoch solution and perform iterative optimization of the error model.

[0047] Optionally, in an implementation of the first aspect of the present invention, step 4: cloud-based intelligent analysis, inverting the three-dimensional water vapor field based on the oblique path tropospheric delay PWV to achieve tower-level rainfall prediction; and generating a risk assessment report based on the AI ​​algorithm fusing displacement rate, inclination mutation, and video features, includes:

[0048] Cloud-based intelligent analysis, based on the BeiDou-3 PPP-B2b precise point positioning service, combines meteorology and GNSS technology to generate tower-level rainfall forecasts and risk assessment reports, including:

[0049] Using the BeiDou-3 PPP-B2b service, the atmospheric precipitable water volume (PWV) is inverted from observation data. Combined with a three-dimensional water vapor tomography model, the three-dimensional water vapor field in the monitoring area is reconstructed to achieve high-precision water vapor forecasts and make refined predictions of surrounding rainfall.

[0050] The BeiDou comprehensive geological disaster monitoring device can be quickly deployed as a temporary base station, transmitting observation data to the BeiDou platform in real time. This complements the base station network, improves the availability of differential services, and provides centimeter-level real-time dynamic positioning services for drone inspections.

[0051] Based on AI algorithms, a more accurate risk assessment report is generated by fusing displacement rate, tilt mutation, and multi-dimensional data of video features.

[0052] Optionally, in an implementation of the first aspect of the present invention, step 5: multimodal monitoring: superimposing displacement heat maps, real-time videos, and warning area annotations in the BIM+GIS digital twin scene through a three-dimensional visualization platform, and automatically capturing alarm-triggered videos through 360° panoramic stitching and downward-looking wide-angle monitoring, including:

[0053] S5.1. Build a multimodal data fusion architecture, including data layer integration. Data types in the data layer integration include Beidou displacement data, 3D geological models, geospatial basemaps, real-time video streams, and meteorological data. Fusion rules include spatial alignment and temporal synchronization. Spatial alignment involves achieving millimeter-level alignment through a seven-parameter conversion between the Beidou CGCS2000 coordinate system and the BIM local coordinate system. Temporal synchronization involves using a hybrid NTP+PTP protocol to ensure that the time stamp error between video frames and displacement data is less than 10ms.

[0054] S5.2, Displacement Trend Prediction and Target Recognition: A multimodal fusion model is constructed. Displacement trend prediction uses an LSTM neural network to analyze time series data and corrects model errors in combination with historical InSAR data. Video target recognition uses the YOLOv7 algorithm to detect crack extension and rock slip, and matches and locates them with a 3D oblique photography model. Risk level assessment generates a risk assessment matrix based on multiple indicators such as displacement rate, rainfall, and crack width, and outputs a four-level warning system: red / orange / yellow / blue. The multimodal fusion model evaluates the multimodal data and outputs the assessment results.

[0055] S5.3, Generate Displacement Heat Map: Generate displacement heat map through key technologies of 3D visualization platform, use WebGL shader to achieve GPU accelerated rendering, and support real-time update of multiple monitoring points;

[0056] S5.4, Building a BIM+GIS Digital Twin Scene: Engine selection includes Cesium for global terrain and satellite imagery, Three.js for loading BIM models and achieving local high-precision rendering, and layered display control through layer visibility logic, enabling seamless integration of 3D scenes and videos.

[0057] S5.5, Panoramic Monitoring Integration: Using OpenCV's SIFT feature matching and GPU-accelerated stitching technology, the PnP algorithm solves camera extrinsics, maps video pixel coordinates to 3D scene coordinates, and performs panoramic video stitching and the fusion of real-time monitoring video and 3D models.

[0058] S5.6, set multi-level warning trigger conditions: When the Beidou displacement exceeds the threshold and the warning level reaches yellow or above, the system will call the PTZ camera preset position to capture key frames and store them. At the same time, it will use AI to analyze cracks or collapses and generate a report that will be pushed to the BIM annotation, realizing the linkage and interaction of multimodal data;

[0059] S5.7, a crack / collapse detection model based on a YOLOv8+Transformer hybrid architecture: This model uses the Transformer architecture to process complex spatiotemporal relationships, enhance AI analysis, and improve detection accuracy.

[0060] S5.7, a crack / collapse detection model based on a YOLOv8+Transformer hybrid architecture: This model uses the Transformer architecture to process complex spatiotemporal relationships, enhance AI analysis, and improve detection accuracy.

[0061] S5.8, Monitoring point density is adapted to harsh environments: When the monitoring point density exceeds 1000 points / km 2When the system is in low light, it automatically switches to cluster display mode; in harsh environments, it performs low-light enhancement and switches to STARVIS sensor mode combined with Retinex algorithm to improve image quality; in the event of communication interruption, the edge node caches the data of the last 15 minutes and resumes the transmission after the network is restored;

[0062] S5.9, Hololens2 superimposes real-time displacement data, superimposes real-time displacement data onto the on-site field of view through Hololens2, and uploads key data of warning events to the chain to meet regulatory audit requirements.

[0063] Optionally, in an implementation of the first aspect of the present invention, step 6: intelligent early warning and decision-making: through big data analysis and artificial intelligence algorithms, real-time solution and trend analysis of monitoring data are performed to generate early warning information and automatically trigger an alarm mechanism. When the displacement rate exceeds a threshold or AI identifies a precursory feature of collapse, an emergency response mechanism is triggered, an alarm is automatically pushed to the emergency management platform, and a drone inspection plan is simultaneously initiated, including:

[0064] The Beidou geological disaster monitoring scene monitoring panel provides a custom monitoring threshold function, distinguishing between red, orange, yellow and blue levels to achieve precise control of tower risks;

[0065] Provides scene visualization. Through the Nanfang Grid's intelligent map marking and sensor 3D model installation functions, it visualizes the location of towers equipped with Beidou geological disaster monitoring terminals, risk levels, real-time alarms, and real-time monitoring data curve information, enabling remote real-time monitoring of tower operating conditions.

[0066] It provides a monitoring report export function, which generates special monitoring reports for key controlled line towers to improve the risk management of geological disasters on transmission lines.

[0067] In a second aspect, an embodiment of the present application provides a Beidou geological disaster displacement and image and video monitoring fusion system based on a highly reliable communication link, which is applied to the Beidou geological disaster displacement and image and video monitoring fusion method based on a highly reliable communication link as described in the first aspect, including:

[0068] Perception layer: Beidou monitoring devices are used to collect millimeter-level displacement, tilt, and five-directional high-definition video in real time; micro-meteorological and wire clamp temperature data are accessed and uniformly packaged through the data gateway;

[0069] Transport layer: Highly reliable data transmission: APN dedicated line transmission is used, and all data is marked with Beidou timing timestamps to ensure timing alignment; through the temporary base station mode, centimeter-level RTK differential services are provided for drones;

[0070] Processing layer: Based on BeiDou VRS virtual reference station and long baseline monitoring solution technology, the completed BeiDou ground-based augmentation station network is reused to build virtual reference stations to replace traditional physical base stations. Using observation data from multiple reference stations, an error correction model is established to generate virtual observation values ​​for real-time dynamic RTK solution.

[0071] Analysis layer: Cloud-based intelligent analysis uses oblique path tropospheric delayed PWV to invert the three-dimensional water vapor field, enabling tower-level rainfall prediction. AI algorithms fuse displacement rate, inclination mutations, and video features to generate risk assessment reports.

[0072] Platform layer: Multimodal monitoring: Through the 3D visualization platform, displacement heat maps, real-time videos, and warning area annotations are superimposed on the BIM+GIS digital twin scene. Through 360° panoramic stitching and downward-looking wide-angle monitoring, alarm triggers automatic video capture;

[0073] Application layer: Intelligent early warning and decision-making: Through big data analysis and artificial intelligence algorithms, the monitoring data is solved and trended in real time, early warning information is generated and the alarm mechanism is automatically triggered. When the displacement rate exceeds the threshold or AI identifies the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is simultaneously launched.

[0074] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0075] processor;

[0076] a memory for storing processor-executable instructions;

[0077] Among them, the processor is configured to implement the Beidou geological disaster displacement and image and video monitoring fusion method based on a high-reliability communication link as described in the first aspect when executing the instructions.

[0078] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a program, and the program instructs a device to execute the Beidou geological disaster displacement and image and video monitoring fusion method based on a high-reliability communication link as described in the first aspect.

[0079] The application discloses a Beidou disaster displacement and image video monitoring fusion method based on a high-reliability communication link, and comprises the following steps: multi-source data acquisition: real-time acquisition of millimeter-level displacement, inclination and five-direction high-definition video through a Beidou monitoring device, and access to micro-meteorological and wire clip temperature data, and unified encapsulation through a data gateway; high-reliability data transmission: APN special line transmission is adopted, a Beidou time stamp is marked to ensure time sequence alignment, and centimeter-level RTK differential service is provided for a UAV; virtual reference station construction: based on a Beidou VRS virtual reference station and a long baseline monitoring solution technology, a Beidou ground-based enhancement station network is multiplexed to generate virtual observation values for real-time dynamic RTK solution; cloud-side intelligent analysis: based on an oblique path tropospheric delay PWV, a three-dimensional water vapor field is inversed to realize tower-level rainfall prediction, and an AI algorithm is fused to generate a risk assessment report; multi-modal monitoring: in a BIM+GIS digital twin scene, displacement heat maps, real-time videos and early warning labels are superimposed to support 360-degree panoramic splicing and alarm triggering snapshot; intelligent early warning decision: through real-time analysis of big data and AI, an alarm is automatically pushed when a threshold value is exceeded or a collapse precursor is identified, and a UAV emergency inspection is started.

[0080] Beneficial effects:

[0081] (1) High-reliability communication architecture: APN special line + Beidou time synchronization, ensuring low-delay and high-trust data transmission.

[0082] (2) Virtual reference station optimization: multiplexing an existing Beidou ground-based enhancement station network reduces construction cost and improves long baseline solution accuracy.

[0083] (3) Multi-modal fusion: BIM+GIS three-dimensional visualization platform integrates displacement heat maps, panoramic videos and AI risk analysis to realize digital twin dynamic monitoring.

[0084] (4) Intelligent early warning linkage: based on multi-dimensional AI models such as LSTM displacement prediction and YOLOv8 crack detection, a hierarchical emergency response mechanism is triggered.

[0085] (5) UAV cooperative inspection: through a temporary reference station mode, a centimeter-level positioning is provided for a UAV to form a “monitoring-warning-disposal” closed loop. BRIEF DESCRIPTION OF DRAWINGS

[0086] Figure 1 The Beidou disaster displacement and image video monitoring fusion method based on a high-reliability communication link provided by the embodiment of the application is shown in the flowchart.

[0087] Figure 2The virtual observation value generation process schematic diagram is provided based on the Beidou VRS virtual reference station and the long baseline monitoring solution technology.

[0088] Figure 3 The model schematic diagram of the traction force of the unmanned plane and the repulsion force generated by the obstacle to the unmanned plane is provided for an embodiment of the application.

[0089] Figure 4 The error model correction process schematic diagram using the LAMBDA algorithm is provided for an embodiment.

[0090] Figure 5 The electronic device schematic diagram is provided for an embodiment of the application. DETAILED DESCRIPTION

[0091] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are some of the embodiments of the application, but not all the embodiments of the application.

[0092] It should be noted that "at least one" in the embodiments of the application means one or more, and more means two or two or more. Unless otherwise defined, all technical and scientific terms used in the present application have the same meanings as those commonly understood by those skilled in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing the specific embodiments of the application, and are not intended to limit the application.

[0093] It should be noted that in the embodiments of the application, the terms "first", "second", etc. are only used for the purpose of distinguishing description, and cannot be understood as indicating or implying relative importance, nor can it be understood as indicating or implying sequence. The features limited by "first", "second" can explicitly or implicitly include one or more of the features. In the description of the embodiments of the application, the terms "exemplary" or "for example" are used to mean as an example, illustration or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of the application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the terms "exemplary" or "for example" are used to present the relevant concept in a specific manner.

[0094] Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0095] Embodiment one

[0096] The embodiment discloses a Beidou disaster displacement and image video monitoring fusion method based on a high-reliability communication link, and comprises the following steps: multi-source data acquisition: real-time acquisition of millimeter-level displacement, inclination and five-direction high-definition video by a Beidou monitoring device, and access to micro-meteorological and clamp temperature data, and unified encapsulation through a data gateway; high-reliability data transmission: APN special line transmission is adopted, a Beidou time stamp is marked to ensure time sequence alignment, and centimeter-level RTK differential service is provided for a UAV; virtual reference station construction: based on a Beidou VRS virtual reference station and a long baseline monitoring solution technology, a Beidou ground-based enhancement station network is multiplexed to generate virtual observation values for real-time dynamic RTK solution; cloud-side intelligent analysis: based on a slant-path tropospheric delay PWV, a three-dimensional water vapor field is inversed to realize tower-level rainfall prediction, and an AI algorithm is fused to generate a risk assessment report; multi-modal monitoring: in a BIM+GIS digital twin scene, a displacement heat map, real-time video and early warning labels are superimposed to support 360-degree panoramic splicing and alarm triggering snapshot; intelligent early warning decision: through real-time analysis of big data and AI, an alarm is automatically pushed when a threshold value is exceeded or a collapse precursor is identified, and a UAV emergency inspection is started.

[0097] Figure 1 A Beidou disaster displacement and image video monitoring fusion method based on a high-reliability communication link is provided for an embodiment of the application. Figure 1 As shown in the figure, a Beidou disaster displacement and image video monitoring fusion method based on a high-reliability communication link comprises the following steps:

[0098] Step 1: multi-source data acquisition, real-time acquisition of millimeter-level displacement, inclination and five-direction high-definition video by a Beidou monitoring device; access to micro-meteorological and clamp temperature data, and unified encapsulation through a data gateway.

[0099] Specifically, the step 1: multi-source data acquisition, real-time acquisition of millimeter-level displacement, inclination and five-direction high-definition video by a Beidou monitoring device; access to micro-meteorological and clamp temperature data, and unified encapsulation through a data gateway, comprises the following steps:

[0100] A Beidou disaster monitoring comprehensive device is deployed on a tower and a slope to real-time acquire millimeter-level high-precision displacement settlement, inclination, image video data, wherein the Beidou disaster monitoring comprehensive device is based on GNSS high-precision receiver positioning technology, low-power high-precision GNSS chips and five-channel front, rear, left and right downward cameras, double-path 180-degree panoramic high-definition image monitoring splicing, MEMS sensor technology, narrowband communication technology and embedded system hibernation wake-up technology are fused together to provide millimeter-level high-precision displacement monitoring information and 360-degree panoramic image video monitoring all day round;

[0101] The Beidou geological disaster monitoring comprehensive device has a data gateway function, can access micro-meteorological, wire clamp temperature measurement and various sensor data, and selects and applies according to the monitoring scene;

[0102] The data is returned to the unified Internet of Things platform through the Internet of Things APN special line, and the positioning and map capabilities based on the Beidou platform and the South Network Sky platform are used to carry out multi-source data fusion application through the system, and support digital production application.

[0103] It can be understood that in the present embodiment, millimeter-level displacement, inclination and five-direction high-definition video data are collected in real time by the Beidou monitoring device, micro-meteorological and wire clamp temperature data are accessed, and are returned to the Internet of Things platform after being uniformly packaged by the data gateway. This multi-source data acquisition method can realize comprehensive perception of geological disasters, including key indicators such as deformation and crack expansion. The Beidou monitoring equipment can realize millimeter-level precision displacement monitoring and support integration of various sensor data. In addition, the Beidou system also has a short message communication function, which can complete data transmission in a network-free environment.

[0104] In the high-risk areas of geological disasters such as towers, slopes, etc., the Beidou geological disaster monitoring comprehensive device (integrating GNSS receiver, multi-channel camera and environmental sensor) is deployed to build an all-weather, multi-dimensional three-dimensional monitoring network. The GNSS high-precision module includes a dual-frequency / multi-frequency low-power GNSS chip (such as and Xintong UB482), which supports BDS-3 full-frequency signal. Combined with MEMS inertial sensor compensation for high-frequency vibration noise (such as accelerometer + gyroscope), the millimeter-level displacement precision in dynamic environment (horizontal ±2mm +1ppmRMS) is improved. At the same time, five cameras (200 million pixel wide-angle lens) are configured in front, back, left, right and down view, supporting: dual 180° panoramic stitching (such as fisheye lens + OpenCV image fusion), wide-angle monitoring (covering the slope foot crack development area), video coding standard H.265 / H.264 + AI edge computing (such as crack identification algorithm).

[0105] At the same time, according to the application scene, multi-source sensors such as micro-meteorological, wire clamp temperature, rain gauge, crack meter, soil moisture meter, inclinometer, etc. can be accessed, and selected and applied according to the monitoring scene.

[0106] Step 2: High-reliability data transmission: APN special line transmission is adopted, all data is marked with Beidou time stamp, and time sequence alignment is ensured; through the temporary reference station mode, centimeter-level RTK differential service is provided for unmanned aerial vehicles.

[0107] Specifically, all data is transmitted via dedicated APN lines and marked with Beidou timing timestamps to ensure time alignment. Simultaneously, a temporary base station mode is used to provide centimeter-level RTK differential services to drones. This process highlights the high reliability and real-time performance of the Beidou system, which can meet the monitoring needs in complex geological environments.

[0108] It can be understood that in this embodiment, a customized APN dedicated line (logically isolated from the public network) is used to build a dedicated network channel, and the end-to-end IPSec encrypted tunnel ensures the security of data transmission. The temporary base station needs to be equipped with a dual-frequency BDS-3 receiver (such as Huace Navigation i70). Support 1Hz raw observation value output (including carrier phase, pseudorange, Doppler shift). Quick deployment: Deploy the temporary base station 30 minutes before the drone takes off (the site selection requires a wide field of view and a stable foundation). Differential data generation: The base station solves the satellite orbit / clock error in real time. Broadcast the RTCM3.3 format differential signal through the L-band radio / UHF digital radio. Finally, the drone-side solution is performed.

[0109] Step 3: Based on the Beidou VRS virtual reference station and long baseline monitoring and solution technology, the completed Beidou ground-based augmentation station network is reused to build a virtual reference station to replace the traditional physical base station. Through the observation data of multiple reference stations, an error correction model is established to generate virtual observation values ​​for real-time dynamic RTK solution.

[0110] As will be appreciated, this embodiment utilizes BeiDou VRS virtual reference stations and long-baseline monitoring and solution technology, reusing the already established BeiDou ground-based augmentation station network to construct virtual reference stations in place of traditional physical base stations. An error correction model is established using observation data from multiple reference stations to generate virtual observations for real-time kinematic RTK solutions. This process utilizes GNSS high-precision positioning technology, combined with regional linear interpolation or low-order surface fitting methods, to correct errors in the observations, thereby improving solution accuracy.

[0111] Virtual Reference Station (VRS) technology generates carrier phase observations at a virtual reference station using observation data from multiple base stations. Its core steps include error modeling, ambiguity resolution, and synthesis of carrier phase observations.

[0112] Figure 2 A schematic diagram of a process for generating virtual observation values ​​based on BeiDou VRS virtual reference station and long baseline monitoring solution technology provided in one embodiment of the present application. Figure 3 Schematic diagram of a model showing the traction force of a drone and the repulsive force exerted on the drone by obstacles provided in one embodiment of the present application.

[0113] Specifically, in this embodiment, Figure 2 、 3As shown, step 3: Based on the BeiDou VRS virtual reference station and long baseline monitoring solution technology, the completed BeiDou ground-based augmentation station network is reused to build a virtual reference station to replace the traditional physical base station. Through the observation data of multiple reference stations, an error correction model is established to generate virtual observation values ​​for real-time dynamic RTK solution, including:

[0114] Step 3.1: Data collection is performed through a physical reference station network. Existing BeiDou ground-based augmentation stations must be reused and meet the following requirements: station spacing ≤ 30 km; dual-frequency or multi-frequency BeiDou receivers must be used; data sampling rate ≥ 1 Hz; raw observation values ​​must be transmitted back in real time; and inter-station time synchronization must be achieved using the PTP protocol.

[0115] Step 3.2: Obtain the error model through network RTK error modeling, where the formula for the carrier phase observation value is:

[0116]

[0117] in, is the carrier phase observation value, R is the distance from the receiver to the satellite, I is the ionospheric delay, T is the tropospheric delay, c is the speed of light, δt r and δt s are the clock errors of the receiver and the satellite respectively, O is the satellite orbit error, M is the multipath effect error, λ is the carrier phase wavelength, and N is the integer ambiguity;

[0118] Step 3.3: Through Geometry-Free combined extraction, regional linear interpolation or low-order surface fitting is used to correct the error of the observation value of the reference station. The formula for the corrected carrier phase observation value is:

[0119]

[0120] Among them, I cor and T cor are the correction values ​​for ionospheric and tropospheric delays, δt r,cor and δt s,cor are the correction values ​​of receiver and satellite clock errors respectively;

[0121] Step 3.4: Based on the observations of the main reference station, combined with the geometric correlation terms and atmospheric delay correction, the carrier phase observations of the virtual reference station are generated. The formula is as follows:

[0122]

[0123] in, is the carrier phase observation, is the observation value of the subjective station, is the geometric distance error gradient, is the atmospheric delay error gradient.

[0124] In Global Navigation Satellite System (GNSS), LAMBDA algorithm is an efficient method for integer ambiguity resolution. The core idea is to reduce the correlation between ambiguities by Z-transformation, thus simplifying the search space of integer ambiguities. Specifically, the error model is modified using LAMBDA algorithm, including:

[0125] Step 4.1: Error modeling:

[0126] Step 4.1.1: Ionospheric delay, extract ionospheric delay by dual-frequency geometry-free combination, and establish a low-order polynomial regional space model, where the ionospheric delay formula is:

[0127]

[0128] where I is the ionospheric delay value, f1 and f2 are the frequencies of the two frequency points respectively, and are the carrier phase observations in cycles, λ1 and λ2 are the carrier phase wavelengths corresponding to the carrier phase observations respectively;

[0129] Step 4.1.2: Tropospheric delay: dry component is corrected by Saastamoinen model, and wet component is estimated by random walk process;

[0130] Step 4.1.3: Orbit / clock bias: use precise ephemeris product or inter-station difference to eliminate;

[0131] Step 4.2: LAMBDA ambiguity fixing;

[0132] Step 4.2.1: Perform integer Gaussian transformation on Q N to reduce the correlation between ambiguities, and the transformation formula is:

[0133] Z T Q N Z=D,

[0134] where Q N is the initial covariance matrix, D is a diagonal matrix, Z is the decorrelation matrix, and |Z| = 1;

[0135] Transformed ambiguities:

[0136] Step 4.2.2: Search for the optimal integer solution in the transformed space by integer least squares Adopt Bootstrapping to accelerate the calculation;

[0137] Step 4.2.3: Inverse transformation and verification, Inverse transform back to the original space: Fixed reliability is verified by ratio test RatioTes;

[0138] Step 4.3: Fixed solution error correction, the fixed ambiguity Substitute into the observation equation, and calculate the residual term;

[0139] Step 4.4: Update the covariance matrix Q using the residual to correct the ionospheric / tropospheric model parameters N , into the next epoch solution, error model iteration optimization.

[0140] Step 4: Cloud intelligent analysis, based on the slant path tropospheric delay PWV inversion three-dimensional water vapor field, realize the tower level rainfall prediction; based on AI algorithm fusion displacement rate, angle mutation, video features, generate risk assessment report.

[0141] Specifically, the step 4: cloud intelligent analysis, based on the slant path tropospheric delay PWV inversion three-dimensional water vapor field, realize the tower level rainfall prediction; based on AI algorithm fusion displacement rate, angle mutation, video features, generate risk assessment report, including:

[0142] Cloud intelligent analysis is based on Beidou III PPP-B2b precise point positioning service, combined with meteorology and GNSS technology, realizes the tower level rainfall prediction and risk assessment report generation, including:

[0143] Using Beidou III PPP-B2b service, the atmospheric precipitable water PWV is inverted through observation data, and combined with three-dimensional water vapor tomography model, the three-dimensional water vapor field of the monitoring area is reconstructed, high-precision water vapor prediction is realized, and the surrounding rainfall is finely predicted;

[0144] The Beidou disaster comprehensive monitoring device can be quickly deployed as a temporary reference station, real-time back observation data to the Beidou platform, become the supplement of the reference station network, improve the availability of differential service, provide centimeter-level real-time dynamic positioning service for unmanned aerial vehicle inspection;

[0145] Based on AI algorithm fusion displacement rate, angle mutation, video features multi-dimensional data, generate more accurate risk assessment report.

[0146] It can be understood that in the embodiment, based on Beidou III PPP-B2b precise point positioning service, combined with meteorology and GNSS technology, realizes the tower level rainfall prediction and risk assessment report generation. For example, by inverting atmospheric precipitable water PWV and combining three-dimensional water vapor tomography model, the three-dimensional water vapor field of the monitoring area is reconstructed, high-precision rainfall prediction is realized. In addition, the AI algorithm fusion displacement rate, angle mutation and video features generate more accurate risk assessment report.

[0147] Step 5: Multimodal monitoring: Through the 3D visualization platform, displacement heat maps, real-time videos, and warning area annotations are superimposed in the BIM+GIS digital twin scene. Through 360° panoramic stitching and downward wide-angle monitoring, alarms trigger automatic video capture.

[0148] Specifically, Step 5: Multimodal Monitoring: Using a 3D visualization platform, displacement heat maps, real-time video, and warning area annotations are superimposed on the BIM+GIS digital twin scene. Through 360° panoramic stitching and downward-looking wide-angle monitoring, alarm-triggered video is automatically captured, including:

[0149] S5.1. Build a multimodal data fusion architecture, including data layer integration. Data types in the data layer integration include Beidou displacement data, 3D geological models, geospatial basemaps, real-time video streams, and meteorological data. Fusion rules include spatial alignment and temporal synchronization. Spatial alignment involves achieving millimeter-level alignment through a seven-parameter conversion between the Beidou CGCS2000 coordinate system and the BIM local coordinate system. Temporal synchronization involves using a hybrid NTP+PTP protocol to ensure that the time stamp error between video frames and displacement data is less than 10ms.

[0150] S5.2, Displacement Trend Prediction and Target Recognition: A multimodal fusion model is constructed. Displacement trend prediction uses an LSTM neural network to analyze time series data and corrects model errors in combination with historical InSAR data. Video target recognition uses the YOLOv7 algorithm to detect crack extension and rock slip, and matches and locates them with a 3D oblique photography model. Risk level assessment generates a risk assessment matrix based on multiple indicators such as displacement rate, rainfall, and crack width, and outputs a four-level warning system: red / orange / yellow / blue. The multimodal fusion model evaluates the multimodal data and outputs the assessment results.

[0151] S5.3, Generate Displacement Heat Map: Generate displacement heat map through key technologies of 3D visualization platform, use WebGL shader to achieve GPU accelerated rendering, and support real-time update of multiple monitoring points;

[0152] S5.4, Building a BIM+GIS Digital Twin Scene: Engine selection includes Cesium for global terrain and satellite imagery, Three.js for loading BIM models and achieving local high-precision rendering, and layered display control through layer visibility logic, enabling seamless integration of 3D scenes and videos.

[0153] S5.5, Panoramic Monitoring Integration: Using OpenCV's SIFT feature matching and GPU-accelerated stitching technology, the PnP algorithm solves camera extrinsics, maps video pixel coordinates to 3D scene coordinates, and performs panoramic video stitching and the fusion of real-time monitoring video and 3D models.

[0154] S5.6, set multi-level warning trigger conditions: When the Beidou displacement exceeds the threshold and the warning level reaches yellow or above, the system will call the PTZ camera preset position to capture key frames and store them. At the same time, it will use AI to analyze cracks or collapses and generate a report that will be pushed to the BIM annotation, realizing the linkage and interaction of multimodal data;

[0155] S5.7, a crack / collapse detection model based on a YOLOv8+Transformer hybrid architecture: This model uses the Transformer architecture to process complex spatiotemporal relationships, enhance AI analysis, and improve detection accuracy.

[0156] S5.7, a crack / collapse detection model based on a YOLOv8+Transformer hybrid architecture: This model uses the Transformer architecture to process complex spatiotemporal relationships, enhance AI analysis, and improve detection accuracy.

[0157] S5.8, Monitoring point density is adapted to harsh environments: When the monitoring point density exceeds 1000 points / km 2 When the system is in low light, it automatically switches to cluster display mode; in harsh environments, it performs low-light enhancement and switches to STARVIS sensor mode combined with Retinex algorithm to improve image quality; in the event of communication interruption, the edge node caches the data of the last 15 minutes and resumes the transmission after the network is restored;

[0158] S5.9, Hololens2 superimposes real-time displacement data, superimposes real-time displacement data onto the on-site field of view through Hololens2, and uploads key data of warning events to the chain to meet regulatory audit requirements.

[0159] As you can understand, in this embodiment, displacement heat maps, real-time video, and warning zone annotations are overlaid on the BIM+GIS digital twin scene through a 3D visualization platform. For example, WebGL shaders are used to achieve GPU-accelerated rendering and support real-time updates of multiple monitoring points. Furthermore, when an alarm is triggered, video is automatically captured to document the disaster's development.

[0160] Step 6: Intelligent early warning and decision-making: Through big data analysis and artificial intelligence algorithms, the monitoring data is solved and trended in real time to generate early warning information and automatically trigger the alarm mechanism. When the displacement rate exceeds the threshold or AI identifies the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is simultaneously launched.

[0161] Specifically, Step 6: Intelligent Early Warning and Decision-making: Through big data analysis and artificial intelligence algorithms, monitoring data is solved and trended in real time to generate early warning information and automatically trigger the alarm mechanism. When the displacement rate exceeds the threshold or AI identifies the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is simultaneously launched, including:

[0162] The Beidou geological disaster monitoring scene monitoring panel provides a custom monitoring threshold function, distinguishing between red, orange, yellow and blue levels to achieve precise control of tower risks;

[0163] Provides scene visualization. Through the Nanfang Grid's intelligent map marking and sensor 3D model installation functions, it visualizes the location of towers equipped with Beidou geological disaster monitoring terminals, risk levels, real-time alarms, and real-time monitoring data curve information, enabling remote real-time monitoring of tower operating conditions.

[0164] It provides a monitoring report export function, which generates special monitoring reports for key controlled line towers to improve the risk management of geological disasters on transmission lines.

[0165] As you can see, this embodiment fully leverages the high-precision positioning capabilities of the BeiDou system, IoT technology, and AI algorithms to achieve intelligent management of the entire geological disaster monitoring chain. This not only improves the efficiency and accuracy of disaster warnings, but also provides strong support for disaster emergency response, achieving significant results in monitoring geological disasters such as landslides and debris flows.

[0166] Example 2

[0167] like Figure 4 As shown, the present application provides a Beidou geological disaster displacement and image and video monitoring fusion system based on a high-reliability communication link, which is applied to the Beidou geological disaster displacement and image and video monitoring fusion method based on a high-reliability communication link as described in Example 1, including: a perception layer, a transmission layer, a processing layer, an analysis layer, a platform layer, and an application.

[0168] It can be understood that in this embodiment, the perception layer is used to collect millimeter-level displacement, inclination, and five-directional high-definition video in real time through the Beidou monitoring device; access micro-meteorological and wire clamp temperature data, and uniformly package them through the data gateway.

[0169] It can be understood that in this embodiment, the transport layer is used for high-reliability data transmission: APN dedicated line transmission is adopted, and all data are marked with Beidou timing timestamps to ensure timing alignment; through the temporary base station mode, centimeter-level RTK differential services are provided for drones.

[0170] It can be understood that in this embodiment, the processing layer is used to reuse the completed Beidou ground-based augmentation station network based on the Beidou VRS virtual reference station and long baseline monitoring solution technology, build a virtual reference station to replace the traditional physical base station, and establish an error correction model through the observation data of multiple reference stations to generate virtual observation values ​​for real-time dynamic RTK solution.

[0171] It can be understood that in this embodiment, the analysis layer is used for cloud-based intelligent analysis, and the three-dimensional water vapor field is inverted based on the oblique path tropospheric delay PWV to achieve tower-level rainfall prediction; based on the AI ​​algorithm, the displacement rate, inclination mutation, and video features are integrated to generate a risk assessment report.

[0172] It can be understood that in this embodiment, the platform layer is used for multimodal monitoring: through the three-dimensional visualization platform, displacement heat maps, real-time videos and warning area annotations are superimposed in the BIM+GIS digital twin scene, and through 360° panoramic stitching and downward wide-angle monitoring, alarms trigger automatic video capture.

[0173] It can be understood that in this embodiment, the application layer is used for intelligent early warning and decision-making: through big data analysis and artificial intelligence algorithms, the monitoring data is solved and trend analyzed in real time, early warning information is generated and the alarm mechanism is automatically triggered. When the displacement rate exceeds the threshold or AI recognizes the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is started simultaneously.

[0174] Figure 5 This is an electronic device provided by an embodiment of the present application. Figure 5 As shown, the electronic device includes at least the following parts: a processor 101 and a memory 100 , a communication interface 103 , and a bus 102 .

[0175] In the embodiment of the present application, the memory 100 is used to store instructions executable by the processor 101. The processor 101 is configured to execute the instructions to implement the following Figure 5 The device module shown is based on a high-reliability communication link and integrates Beidou geological disaster displacement and image video monitoring.

[0176] In an embodiment of the present application, a computer-readable storage medium includes instructions, and the instructions instruct a device to execute the method of the first aspect. For example, the instructions instruct the device to execute Figure 1 The method is shown in the process steps.

[0177] The program running in the electronic device involved in one embodiment of the present application can be a program that controls a central processing unit (CPU) and the like to realize the functions of the above-mentioned embodiment involved in one embodiment of the present invention (a program that enables a computer to function). Then, the information processed by these devices is temporarily stored in a random access memory (RAM) during its processing, and then stored in various ROMs such as read-only memory (Flash ROM) and hard disk drive (HDD), and is read, modified, and written by the CPU as needed.

[0178] It should be noted that a portion of the electronic device of the above embodiment may also be implemented by a computer. In this case, a program for implementing the control function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer and executed.

[0179] It should be noted that the "computer" mentioned here refers to a computer built into an electronic device, employing hardware including an operating system (OS) and peripheral devices. Furthermore, "computer-readable recording medium" refers to removable media such as floppy disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computers.

[0180] Furthermore, "computer-readable recording media" may include: media that dynamically store programs for a short period of time, such as communication lines when transmitting programs via networks such as the Internet or communication lines such as telephone lines; and media that store programs for a fixed period of time, such as volatile memory within computers acting as servers or clients in this context. Furthermore, the aforementioned program may be a program for implementing a portion of the aforementioned functions, or a program that can achieve the aforementioned functions by combining with a program already stored in a computer.

[0181] Furthermore, the electronic device in the above-described embodiments can also be implemented as a collection (device group) consisting of multiple devices. Each device constituting the device group may have a portion or all of the functions or functional blocks of the electronic device in the above-described embodiments. A device group only needs to have all the functions or functional blocks of the electronic device.

[0182] Those skilled in the art should recognize that the above embodiments are merely intended to illustrate the present application and are not intended to limit the present application. As long as they are within the spirit of the present application, appropriate changes and modifications to the above embodiments are within the scope of protection claimed in the present application.

Claims

1. A Beidou geological disaster displacement and image video monitoring fusion method based on a high-reliability communication link, characterized in that: The method comprises: Step 1: Multi-source data collection: real-time acquisition of millimeter-level displacement, tilt angle, and five-directional high-definition video through Beidou monitoring devices; access to micro-meteorological and wire clamp temperature data, and unified packaging through the data gateway; Step 2: High-reliability data transmission: Using APN dedicated lines for transmission, all data is marked with Beidou timing timestamps to ensure timing alignment; through the temporary base station mode, centimeter-level RTK differential services are provided for drones; Step 3: Based on the BeiDou VRS virtual reference station and long baseline monitoring solution technology, the completed BeiDou ground-based augmentation station network is reused to build a virtual reference station to replace the traditional physical base station. Using the observation data of multiple reference stations, an error correction model is established to generate virtual observation values ​​for real-time dynamic RTK solution. Step 4: Cloud-based intelligent analysis uses the slant-path tropospheric delayed PWV to invert the three-dimensional water vapor field and achieve tower-level rainfall prediction. An AI algorithm integrates displacement rate, inclination mutation, and video features to generate a risk assessment report. Step 5: Multimodal monitoring: Using a 3D visualization platform, displacement heat maps, real-time video, and warning area annotations are superimposed on the BIM+GIS digital twin scene. Through 360° panoramic stitching and downward-looking wide-angle monitoring, alarms trigger automatic video capture. Step 6: Intelligent early warning and decision-making: Through big data analysis and artificial intelligence algorithms, the monitoring data is solved and trended in real time to generate early warning information and automatically trigger the alarm mechanism. When the displacement rate exceeds the threshold or AI identifies the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is simultaneously launched.

2. The method for integrating Beidou geological disaster displacement and image and video monitoring based on a high-reliability communication link according to claim 1 is characterized in that: Step 1: multi-source data acquisition, real-time acquisition of millimeter-level displacement, tilt angle, and five-directional high-definition video through Beidou monitoring device; Access micro-meteorological and wire clamp temperature data, which are uniformly packaged through the data gateway, including: Deploy Beidou comprehensive geological disaster monitoring devices on towers and slopes to collect millimeter-level high-precision displacement, settlement, inclination, image and video data in real time. The Beidou comprehensive geological disaster monitoring device is based on GNSS high-precision receiver positioning technology, low-power high-precision GNSS chips, and five-channel cameras for front, back, left, right and bottom. It can monitor and stitch dual-channel 180° panoramic high-definition images, integrating MEMS sensor technology, narrowband communication technology, and embedded system sleep and wake-up technology. It can provide millimeter-level high-precision displacement monitoring information and 360° panoramic image and video monitoring around the clock. The Beidou geological disaster monitoring integrated device also has a data gateway function, which can access data from various sensors such as micro-meteorology and wire clamp temperature measurement, and can be selected and applied according to the monitoring scenario; The data is transmitted back to the unified IoT platform through the IoT APN dedicated line. Based on the positioning and mapping capabilities of the Beidou platform and the China Southern Power Grid Zhikan platform, multi-source data fusion applications are carried out through the system during operation to support digital production applications.

3. The method for integrating Beidou geological disaster displacement and image and video monitoring based on a high-reliability communication link according to claim 1 is characterized in that: Step 3: Based on the BeiDou VRS virtual reference station and long baseline monitoring solution technology, the completed BeiDou ground-based augmentation station network is reused to build a virtual reference station to replace the traditional physical base station. Through the observation data of multiple reference stations, an error correction model is established to generate virtual observation values ​​for real-time dynamic RTK solution, including: Step 3.1: Data collection is performed through a physical reference station network. Existing BeiDou ground-based augmentation stations must be reused and meet the following requirements: station spacing ≤ 30 km; dual-frequency or multi-frequency BeiDou receivers must be used; data sampling rate ≥ 1 Hz; raw observation values ​​must be transmitted back in real time; and inter-station time synchronization must be achieved using the PTP protocol. Step 3.2: Obtain the error model through network RTK error modeling, where the formula for the carrier phase observation value is: in, is the carrier phase observation value, R is the distance from the receiver to the satellite, I is the ionospheric delay, T is the tropospheric delay, c is the speed of light, δt r and δt s are the clock errors of the receiver and the satellite respectively, O is the satellite orbit error, M is the multipath effect error, λ is the carrier phase wavelength, and N is the integer ambiguity; Step 3.3: Through Geometry-Free combined extraction, regional linear interpolation or low-order surface fitting is used to correct the error of the observation value of the reference station. The formula for the corrected carrier phase observation value is: Among them, I cor and T cor are the correction values ​​for ionospheric and tropospheric delays, δt r,cor and δt s,cor are the correction values ​​of receiver and satellite clock errors respectively; Step 3.4: Based on the observations of the main reference station, combined with the geometric correlation terms and atmospheric delay correction, the carrier phase observations of the virtual reference station are generated. The formula is as follows: in, is the carrier phase observation, is the observation value of the subjective station, is the geometric distance error gradient, is the atmospheric delay error gradient.

4. The method for integrating Beidou geological disaster displacement and image and video monitoring based on a high-reliability communication link according to claim 3 is characterized in that Use the LAMBDA algorithm to correct the error model, including: Step 4.1: Error modeling: Step 4.1.1: Ionospheric delay: extract the ionospheric delay through dual-frequency geometry-free combination and establish a low-order polynomial regional spatial model. The ionospheric delay formula is: Where I is the ionospheric delay value, f1 and f2 are the frequencies of the two frequency points respectively, and are the carrier phase observation values ​​in weeks, λ1 and λ2 are the carrier phase wavelengths corresponding to the carrier phase observation values; Step 4.1.2: Tropospheric delay: The dry component is corrected by the Saastamoinen model, and the wet component is estimated by a random walk process; Step 4.1.3: Orbit / clock error: Eliminate using precise ephemeris products or inter-station differentials; Step 4.2: LAMBDA ambiguity is fixed; Step 4.2.1: Q N Perform integer Gaussian transform to reduce the correlation between ambiguities. The transformation formula is: Z T Q N Z=D, Among them, Q N is the initial covariance matrix, D is a diagonal matrix, Z is the descending correlation matrix, |Z|=1; Blur after transformation: Step 4.2.2: Search for the optimal integer solution in the transformed space by integer least squares search Z∈Z n ,The shrinkage search strategy Bootstrapping is used to accelerate the calculation; Step 4.2.3: Inverse transformation and verification, Inverse transform back to the original space: The fixed reliability was verified by the ratio test RatioTes; Step 4.3: Fix the solution error correction and fix the ambiguity Substitute into the observation equation and inversely calculate the residual term; Step 4.4: Use the residuals to correct the ionosphere / troposphere model parameters and update the covariance matrix Q N , enter the next epoch solution and perform iterative optimization of the error model.

5. The method for integrating Beidou geological disaster displacement and image and video monitoring based on a high-reliability communication link according to claim 1 is characterized in that: Step 4: Cloud-based intelligent analysis, based on the slant path tropospheric delay PWV inversion of the three-dimensional water vapor field, to achieve tower-level rainfall prediction; Based on the AI ​​algorithm, displacement rate, tilt mutation, and video features are integrated to generate a risk assessment report, including: Cloud-based intelligent analysis, based on the BeiDou-3 PPP-B2b precise point positioning service, combines meteorology and GNSS technology to generate tower-level rainfall forecasts and risk assessment reports, including: Using the BeiDou-3 PPP-B2b service, the atmospheric precipitable water volume (PWV) is inverted from observation data. Combined with a three-dimensional water vapor tomography model, the three-dimensional water vapor field in the monitoring area is reconstructed to achieve high-precision water vapor forecasts and make refined predictions of surrounding rainfall. The BeiDou comprehensive geological disaster monitoring device can be quickly deployed as a temporary base station, transmitting observation data to the BeiDou platform in real time. This complements the base station network, improves the availability of differential services, and provides centimeter-level real-time dynamic positioning services for drone inspections. Based on AI algorithms, a more accurate risk assessment report is generated by fusing displacement rate, tilt mutation, and multi-dimensional data of video features.

6. The method for integrating Beidou geological disaster displacement and image and video monitoring based on a high-reliability communication link according to claim 1 is characterized in that: Step 5: Multimodal monitoring: Using a 3D visualization platform, overlay displacement heat maps, real-time video, and warning area annotations on the BIM+GIS digital twin scene. Through 360° panoramic stitching and downward-looking wide-angle monitoring, alarm-triggered video is automatically captured, including: S5.

1. Build a multimodal data fusion architecture, including data layer integration. Data types in the data layer integration include Beidou displacement data, 3D geological models, geospatial basemaps, real-time video streams, and meteorological data. Fusion rules include spatial alignment and temporal synchronization. Spatial alignment involves achieving millimeter-level alignment through a seven-parameter conversion between the Beidou CGCS2000 coordinate system and the BIM local coordinate system. Temporal synchronization involves using a hybrid NTP+PTP protocol to ensure that the time stamp error between video frames and displacement data is less than 10ms. S5.2, Displacement Trend Prediction and Target Recognition: A multimodal fusion model is constructed. Displacement trend prediction uses an LSTM neural network to analyze time series data and corrects model errors in combination with historical InSAR data. Video target recognition uses the YOLOv7 algorithm to detect crack extension and rock slip, and matches and locates them with a 3D oblique photography model. Risk level assessment generates a risk assessment matrix based on multiple indicators such as displacement rate, rainfall, and crack width, and outputs a four-level warning system: red / orange / yellow / blue. The multimodal fusion model evaluates the multimodal data and outputs the assessment results. S5.3, Generate Displacement Heat Map: Generate displacement heat map through key technologies of 3D visualization platform, use WebGL shader to achieve GPU accelerated rendering, and support real-time update of multiple monitoring points; S5.4, Building a BIM+GIS Digital Twin Scene: Engine selection includes Cesium for global terrain and satellite imagery, Three.js for loading BIM models and achieving local high-precision rendering, and layered display control through layer visibility logic, enabling seamless integration of 3D scenes and videos. S5.5, Panoramic Monitoring Integration: Using OpenCV's SIFT feature matching and GPU-accelerated stitching technology, the PnP algorithm solves camera extrinsics, maps video pixel coordinates to 3D scene coordinates, and performs panoramic video stitching and the fusion of real-time monitoring video and 3D models. S5.6, set multi-level warning trigger conditions: When the Beidou displacement exceeds the threshold and the warning level reaches yellow or above, the system will call the PTZ camera preset position to capture key frames and store them. At the same time, it will use AI to analyze cracks or collapses and generate a report that will be pushed to the BIM annotation, realizing the linkage and interaction of multimodal data; S5.7, a crack / collapse detection model based on a YOLOv8+Transformer hybrid architecture: This model uses the Transformer architecture to process complex spatiotemporal relationships, enhance AI analysis, and improve detection accuracy. S5.7, a crack / collapse detection model based on a YOLOv8+Transformer hybrid architecture: This model uses the Transformer architecture to process complex spatiotemporal relationships, enhance AI analysis, and improve detection accuracy. S5.8, Monitoring point density is adapted to harsh environments: When the monitoring point density exceeds 1000 points / km 2 When the system is in low light, it automatically switches to cluster display mode; in harsh environments, it performs low-light enhancement and switches to STARVIS sensor mode combined with Retinex algorithm to improve image quality; in the event of communication interruption, the edge node caches the data of the last 15 minutes and resumes the transmission after the network is restored; S5.9, Hololens2 superimposes real-time displacement data, superimposes real-time displacement data onto the on-site field of view through Hololens2, and uploads key data of warning events to the chain to meet regulatory audit requirements.

7. The method for integrating Beidou geological disaster displacement and image and video monitoring based on a high-reliability communication link according to claim 1 is characterized in that: Step 6: Intelligent early warning and decision-making: Through big data analysis and artificial intelligence algorithms, monitoring data is solved and trended in real time to generate early warning information and automatically trigger the alarm mechanism. When the displacement rate exceeds the threshold or AI identifies the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is simultaneously launched, including: The Beidou geological disaster monitoring scene monitoring panel provides a custom monitoring threshold function, distinguishing between red, orange, yellow and blue levels to achieve precise control of tower risks; Provides scene visualization. Through the Nanfang Grid's intelligent map marking and sensor 3D model installation functions, it visualizes the location of towers equipped with Beidou geological disaster monitoring terminals, risk levels, real-time alarms, and real-time monitoring data curve information, enabling remote real-time monitoring of tower operating conditions. It provides a monitoring report export function, which generates special monitoring reports for key controlled line towers to improve the risk management of geological disasters on transmission lines.

8. A BeiDou geological disaster displacement and image and video monitoring fusion system based on a high-reliability communication link, applied to the BeiDou geological disaster displacement and image and video monitoring fusion method based on a high-reliability communication link as claimed in any one of claims 1 to 7, characterized in that: include: Perception layer: Beidou monitoring devices are used to collect millimeter-level displacement, tilt angle, and five-directional high-definition video in real time; Access micro-meteorological and wire clamp temperature data and package them uniformly through the data gateway; Transport layer: Highly reliable data transmission: APN dedicated line transmission is used, and all data is marked with Beidou timing timestamp to ensure timing alignment; Provide centimeter-level RTK differential services for drones through temporary base station mode; Processing layer: Based on BeiDou VRS virtual reference station and long baseline monitoring solution technology, the completed BeiDou ground-based augmentation station network is reused to build virtual reference stations to replace traditional physical base stations. Using observation data from multiple reference stations, an error correction model is established to generate virtual observation values ​​for real-time dynamic RTK solution. Analysis layer: Cloud-based intelligent analysis uses oblique path tropospheric delayed PWV to invert the three-dimensional water vapor field, enabling tower-level rainfall prediction. AI algorithms fuse displacement rate, inclination mutations, and video features to generate risk assessment reports. Platform layer: Multimodal monitoring: Through the 3D visualization platform, displacement heat maps, real-time videos, and warning area annotations are superimposed on the BIM+GIS digital twin scene. Through 360° panoramic stitching and downward-looking wide-angle monitoring, alarm triggers automatic video capture; Application layer: Intelligent early warning and decision-making: Through big data analysis and artificial intelligence algorithms, the monitoring data is solved and trended in real time, early warning information is generated and the alarm mechanism is automatically triggered. When the displacement rate exceeds the threshold or AI identifies the precursor characteristics of collapse, the emergency response mechanism is triggered, and the alarm is automatically pushed to the emergency management platform, and the drone inspection plan is simultaneously launched.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the Beidou geological disaster displacement and image and video monitoring fusion method based on a high-reliability communication link as described in any one of claims 1 to 7 when executing the instructions.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, and the program instructs the device to execute the Beidou geological disaster displacement and image and video monitoring fusion method based on a high-reliability communication link as described in any one of claims 1 to 7.

Images