Slope monitoring wireless measurement system and method based on unmanned aerial vehicle relay

Abstract

The application discloses a kind of based on unmanned plane throwing relay's side slope monitoring wireless measurement system and method, it is related to side slope safety monitoring technical field.The system includes impact-resistant throwing type sensor subsystem, unmanned plane throwing subsystem, unmanned plane relay subsystem and data processing center subsystem.Unmanned plane throwing subsystem will impact-resistant sensor node quickly, uniformly laid on the surface of side slope;Sensor node is woken up and vibration, acceleration and pressure data are collected after sensing impact;Unmanned plane relay subsystem flies to the airspace of laying area, automatically collects sensor node data and returns to data processing center;Data processing center carries out cleaning, feature extraction and intelligent early warning to data.The application solves the problem that sensor deployment is difficult in complex dangerous side slope area, and data transmission is unreliable, realizes the automation and intelligentization of side slope monitoring.

Description

Technical Field

[0001] This invention relates to the field of geotechnical engineering safety monitoring and geological disaster early warning technology, specifically to a wireless measurement system and method for slope monitoring based on UAV spraying relay. Background Technology

[0002] Slope stability is a critical issue for the construction and operation safety of projects such as highways, railways, mines, and water conservancy. Slope disasters such as landslides and collapses are characterized by their suddenness and destructive power, making real-time and effective monitoring an important means of early warning and disaster mitigation.

[0003] Currently, traditional slope monitoring technologies have the following limitations:

[0004] Manual inspections and the deployment of fixed monitoring stations present challenges: For slopes with complex terrain, steep inclines, or difficult-to-reach terrain, manual inspections are not only inefficient and costly, but also pose significant safety risks. Traditional fixed monitoring stations (such as total stations and fixed inclinometers) require pre-construction of civil engineering at specific locations, making deployment inflexible and limiting their monitoring range, thus failing to achieve comprehensive coverage of large-scale slopes.

[0005] Although wired transmission methods provide stable data transmission, they are difficult to lay in complex slope terrain, and the lines are easily damaged by disasters such as rockfalls and landslides, resulting in poor system reliability and maintainability.

[0006] While wireless sensor technology has solved the cabling problem to some extent, its deployment still relies heavily on manual labor, making it difficult to implement in hazardous areas. Furthermore, conventional sensor nodes often lack the ability to withstand the massive impact of being dropped from a height, hindering rapid deployment through automation. Additionally, the communication range of a single sensor is limited, and signals are easily blocked in complex terrain, making it difficult to directly transmit data over long distances to a data processing center.

[0007] The methods described above are all insufficient to quickly and flexibly build a high-density, comprehensive monitoring network in the target area before or after a disaster, resulting in untimely and incomplete data acquisition, which fails to provide sufficient data support for accurate disaster prediction and assessment. Summary of the Invention

[0008] This invention provides a wireless measurement system for slope monitoring based on drone-based spraying relay, which solves the technical problem of quickly, flexibly, and cost-effectively deploying a high-density wireless monitoring network in slope areas with complex terrain and difficult access by personnel, and achieving reliable data transmission and accurate risk warning.

[0009] This invention is achieved through the following technical solution:

[0010] In a first aspect, this application provides a wireless measurement system for slope monitoring based on UAV spraying relay, characterized in that it includes an impact-resistant spraying sensor subsystem, a UAV spraying subsystem, a UAV relay subsystem, and a data processing center subsystem.

[0011] The impact-resistant spray sensor subsystem includes multiple sensor nodes, which are configured to wake up after sensing an impact, collect and transmit vibration, acceleration and pressure data of the slope.

[0012] The drone spraying subsystem is configured to fly according to a preset path, control the spraying mechanism to release multiple sensor nodes, and deploy the sensor nodes on the slope surface.

[0013] The UAV relay subsystem includes a relay UAV and a relay communication gateway mounted thereon. The relay communication gateway includes a first communication module and a second communication module, and is configured to fly over the sensor deployment area according to a preset cruise path, collect data from the sensor nodes, and send the aggregated data to the data processing center subsystem through the second communication module.

[0014] The data processing center subsystem is configured to receive and process the aggregated data, and to perform risk assessment and graded early warning.

[0015] A further optimization scheme is that the UAV spraying subsystem is configured to fly along a bow-shaped flight path, and trigger a spraying action once every preset flight distance based on GPS positioning information;

[0016] The outer shell of the sensor node is spherical or streamlined, made of engineering plastics or composite materials, and covered with a flexible buffer layer.

[0017] A buffer mechanism is provided between the outer shell and the internal module. The buffer mechanism is at least one of a suspension structure, a silicone potting layer, or a spring damping system.

[0018] A further optimized solution is that the scattering mechanism includes a storage bin, a controllable discharge gate driven by a servo motor, and a dispensing mechanism driven by a stepper motor.

[0019] The sorting mechanism is configured to detach the sensor node from the storage bin and drop it by rotating it at a preset angle;

[0020] The UAV spraying subsystem is configured to deploy sensors to a preset target area in a single flight mission.

[0021] A further optimization is that the relay drone is configured to fly along a low-altitude path covering the area where the sensors are deployed;

[0022] The first communication module is a LoRa collection module with a receiving sensitivity higher than a preset sensitivity threshold. It is configured to continuously broadcast data collection beacons and listen to the channel, and receive data sent by ground sensor nodes in response to the beacons.

[0023] The relay communication gateway also includes an embedded computer configured to temporarily store and convert the collected data.

[0024] A further optimization scheme is that the second communication module is a 4G DTU module or a satellite short message module, which is configured to adaptively switch between cellular network and satellite communication based on real-time signal strength;

[0025] When communication between the relay drone and the data processing center is interrupted for more than a preset time, the relay drone will temporarily store the aggregated data in the local storage module and resume transmission from the point of interruption after communication is restored.

[0026] A further optimization is that the sensor node is equipped with an accelerometer or a gyroscope and is configured to generate a wake-up signal when the sensed acceleration value or attitude angle change exceeds a preset threshold.

[0027] The sensor node is configured to perform normal data acquisition at a first sampling frequency. When the acceleration value of the vibration event detected exceeds a preset threshold, it automatically switches to a second sampling frequency for high-speed data acquisition. After high-speed data acquisition continues for a preset duration, it returns to the normal data acquisition mode.

[0028] A further optimization is that the data processing center subsystem is configured as follows:

[0029] Feature values ​​characterizing slope stability are extracted from vibration signals, acceleration signals, and pressure signals. These feature values ​​include vibration frequency spectrum, peak acceleration, pressure change gradient, and root mean square value.

[0030] By comparing the extracted feature values ​​with the preset safety threshold, when it is determined that more than a preset proportion of sensors in the same area continuously report abnormalities, a corresponding level of warning information is generated.

[0031] A further optimization scheme is that the data processing center subsystem is configured to use machine learning models to train and analyze historical and real-time data to identify abnormal patterns that characterize slope instability. The machine learning models include Long Short-Term Memory Network (LSTM), Support Vector Machine (SVM), or Random Forest.

[0032] Input the acceleration time series data of each sensor for a preset period of time into the pre-trained LSTM model to predict the slope instability trend. When the high-risk probability value output by the model exceeds the critical warning probability and shows spatial clustering anomalies, a high-level warning is generated.

[0033] The patrol frequency of relay drones is dynamically adjusted according to the warning level, and encrypted patrols are initiated when the warning level reaches a preset high level.

[0034] A further optimization is that the system also includes an early warning closed-loop feedback mechanism:

[0035] The entire chain of multimodal data for each early warning event is packaged into a digital case library. The case library data is used to retrain the machine learning early warning model to optimize the early warning threshold and recognition accuracy.

[0036] The thresholds and logic of the early warning rules are dynamically adjusted based on the false alarms and missed alarms.

[0037] Secondly, this application provides a wireless measurement method for slope monitoring based on UAV spraying relay, including the following steps:

[0038] The drone spraying subsystem flies along a preset path and controls the spraying mechanism to release multiple sensor nodes, which are then deployed on the slope surface.

[0039] The drone relay subsystem flies over the sensor deployment area according to a preset cruise path, collects data from the sensor nodes, and sends the aggregated data to the data processing center through the second communication module.

[0040] The data processing center receives and processes the aggregated data to conduct risk assessment and tiered early warning.

[0041] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0042] By relying on the systematic collaboration of impact-resistant drop sensors, drone-based drop, drone relay, and a data processing center, a complete closed-loop architecture has been constructed, encompassing drop deployment, sensing and data acquisition, aerial relay, and intelligent analysis. This architecture not only solves the problems of high difficulty and risk associated with manually deploying sensors on complex and dangerous slopes, but also overcomes the limitations of complex terrain blocking wireless signals by utilizing the patrol and collection mode of relay drones over the deployment area, achieving efficient and reliable aggregation of data from distributed sensor nodes. Simultaneously, relying on the centralized analysis capabilities of the data processing center, multi-source monitoring data is transformed into graded early warning results for slope stability, providing comprehensive technical support for the early identification and emergency response of slope disasters. Compared to traditional fixed monitoring or single wireless sensor networks, this architecture improves the deployment flexibility, data acquisition comprehensiveness, and early warning timeliness of slope monitoring. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0044] Figure 1 A schematic diagram of the slope monitoring system architecture based on UAV spraying and relay transmission provided in the embodiments of this application;

[0045] Figure 2 A flowchart illustrating the slope monitoring method based on UAV spraying and relay transmission provided in this application embodiment. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0047] First, some of the technical terms used in this application will be explained to help those skilled in the art understand this application.

[0048] LoRa: Long Range Radio, a low-power wide-area network communication technology.

[0049] LSTM: Long Short-Term Memory, a type of temporal recurrent neural network;

[0050] DBSCAN: Density-Based Spatial Clustering of Applications with Noise;

[0051] AGC: Automatic Gain Control, a circuit module used to adjust the receiving sensitivity of a communication module;

[0052] eMMC: Embedded Multi Media Card, a type of non-volatile memory chip.

[0053] In one embodiment, such as Figure 1 As shown, the slope monitoring wireless measurement system based on UAV spraying relay includes an impact-resistant spraying sensor subsystem, a UAV spraying subsystem, a UAV relay subsystem, and a data processing center subsystem.

[0054] Specifically, the impact-resistant spray sensor subsystem includes multiple sensor nodes, which are configured to wake up from sleep mode after sensing an impact, collect vibration, acceleration, and pressure data of the slope, and transmit them through a wireless communication module.

[0055] Furthermore, the UAV spraying subsystem is configured to fly along a preset path, controlling the spraying mechanism to release multiple sensor nodes and deploy them on the slope surface. The UAV spraying subsystem uses a six-rotor flight platform with a maximum load capacity of 10 kg, and analyzes the coordinates of the preset path through the flight control system.

[0056] Furthermore, the UAV relay subsystem includes a relay UAV and a relay communication gateway mounted on it. The relay communication gateway is equipped with a first communication module and a second communication module, and is configured to fly over the sensor deployment area according to a preset cruise path. It collects data from the sensor nodes through the first communication module and sends the aggregated data to the data processing center subsystem through the second communication module. The relay communication gateway has a built-in embedded computer running an embedded Linux system, which is used to buffer, sort, and convert the received fragmented data packets during flight.

[0057] Furthermore, the data processing center subsystem is deployed on a server and configured to receive and process aggregated data, perform risk assessment, and issue tiered early warnings. The data processing center subsystem listens for data pushes from the second communication module via a standard Socket network interface and stores the parsed structured data into a time-series database.

[0058] The data processing center subsystem supports both mobile and fixed deployments to adapt to different application scenarios.

[0059] This embodiment constructs a complete closed loop of "deployment of impact-resistant spray sensors, drone spraying, drone relay, and data processing center" through systematic collaboration of impact-resistant spray sensors, drone spraying, drone relay, and data processing center. It not only solves the pain points of high difficulty and high risk of manually deploying sensors on complex and dangerous slopes, but also overcomes the limitations of complex terrain on wireless signals by utilizing the high-altitude advantage of relay drones. It achieves efficient and reliable aggregation of data from distributed sensor nodes, providing complete technical support for the early identification and emergency response of slope disasters.

[0060] In one embodiment, the UAV spraying subsystem is configured to fly along a bow-shaped flight path, triggering a spraying action every preset flight distance based on GPS positioning information. The outer shell of the sensor node is spherical or streamlined, made of impact-resistant polycarbonate material, and covered with a flexible buffer layer; a buffer mechanism, which is a silicone potting layer, is provided between the outer shell and the internal module.

[0061] Specifically, the bow-shaped flight path is automatically generated by the ground control station after marking the target slope area on the electronic map, ensuring coverage without blind spots. The preset distance is calculated based on the ratio of the diagonal length of the target area to the required number of sensor nodes to ensure uniform deployment density.

[0062] Furthermore, the buffer mechanism can also adopt a suspension structure, suspending the internal module inside the outer shell by elastic ropes, or adopt a spring damping system, setting a damping spring between the internal module and the outer shell to further disperse the impact energy upon landing.

[0063] Furthermore, the sorting mechanism can also employ a low-pressure spring ejector to eject the sensor node from the storage bin using spring force, or a pneumatic push rod array to push the sensor node out of the storage bin using pneumatic pressure, adapting to the deployment requirements of sensor nodes of different weights and sizes.

[0064] This embodiment achieves uniform deployment of sensor nodes on the slope surface by using a bow-shaped flight path combined with GPS positioning to trigger the deployment mechanism. The spherical shell, flexible buffer layer and composite buffer structure can effectively absorb the impact energy when the sensor nodes land, which greatly improves the survival rate and working stability of the sensors after deployment, and lays a physical foundation for the reliable acquisition of subsequent monitoring data.

[0065] In one embodiment, the dispensing mechanism includes a storage bin, a controllable discharge gate driven by a servo motor, and a dispensing mechanism driven by a stepper motor. The dispensing mechanism is configured to detach and drop sensor nodes from the storage bin by rotating them at a preset angle. The UAV dispensing subsystem is configured to deploy sensors to a preset target area in a single flight mission.

[0066] Specifically, the storage bin is cylindrical and axially mounted under the belly of the UAV, capable of accommodating approximately 20 sensors. Its volume is determined based on the endurance and deployment density requirements of a single flight mission. The controllable discharge door adopts a double-door structure. After receiving flight control commands, the servo motor drives the linkage mechanism, causing the two doors to open and close synchronously.

[0067] Furthermore, the sorting mechanism employs a rotating dial, with an arc-shaped groove on its outer circumference that matches the shape of the sensor node. A stepper motor drives the dial to rotate 60°, and the groove carries a single sensor node to directly below the discharge port, where the node falls under the influence of gravity.

[0068] This embodiment achieves the orderly and controllable release of sensor nodes through the cooperation of storage bins, controllable discharge gates and rotary wheel-type distribution mechanisms. A single flight mission can complete the full coverage deployment of the preset target area, avoiding the high risk and low efficiency of manual deployment. It is suitable for rapid monitoring deployment scenarios of large-area slopes.

[0069] In one embodiment, the relay drone is configured to fly along a low-altitude path covering the area where sensors are deployed. The first communication module, a LoRa data collection module with a receiving sensitivity higher than a preset sensitivity threshold, is configured to continuously broadcast data collection beacons and listen to the channel, receiving data transmitted by ground sensor nodes in response to the beacons. The relay communication gateway also includes an embedded computer configured to temporarily store and convert the collected data according to protocols.

[0070] Specifically, the flight altitude for the low-altitude path is dynamically adjusted based on the slope's undulations and elevation differences to ensure line-of-sight. The preset sensitivity threshold is set according to the receiver sensitivity curve of LoRa modulation technology to ensure the demodulation of weak far-end signals.

[0071] Furthermore, the data collection beacon contains the unique identifier of the relay UAV and the current timestamp. After receiving the beacon, the ground sensor node sends the data after a random delay according to the preset backoff algorithm to avoid signal collision.

[0072] Furthermore, the embedded computer converts the proprietary protocol data uploaded by different sensor nodes into standard data frames in JSON format.

[0073] This embodiment utilizes the high-altitude advantage of the relay drone cruising at approximately 50m altitude, combined with the broadcast beacon mechanism of the high-sensitivity LoRa collection module, to overcome terrain obstruction and expand the communication coverage of a single flight to an area with a radius of approximately 300m. Combined with the protocol conversion capabilities of the embedded computer, it improves the success rate of sensor data collection and transmission stability in complex terrain.

[0074] In one embodiment, the second communication module is a 4G DTU module or a satellite short message module, configured to adaptively switch between cellular network and satellite communication based on real-time signal strength. If communication between the relay drone and the data processing center is interrupted for more than a preset duration, the relay drone will temporarily store the aggregated data in a local storage module, and resume transmission from the point of interruption once communication is restored.

[0075] Specifically, real-time signal strength is obtained by reading the AGC register value at the bottom layer of the 4G DTU or satellite communication module. The adaptive switching logic is set with a signal strength threshold. When the 4G signal strength is lower than this threshold and continues to be lower than this threshold for more than several heartbeat cycles, the system automatically cuts off the 4G link power and wakes up the satellite short message module to establish a connection.

[0076] Furthermore, the local storage module uses eMMC flash memory chips, and the breakpoint resume mechanism is implemented by recording the sequence number of the sent data. When the communication link is restored, the system prioritizes sending the data in the buffer queue to ensure the continuity of the data stream.

[0077] This embodiment adapts to the complex and ever-changing communication environment of the slope area through the adaptive switching mechanism between the 4G DTU and the satellite short message module. Combined with the breakpoint resume logic based on sequence number records, it effectively avoids the problem of data loss caused by communication interruption, ensures that the data processing center can obtain the complete monitoring dataset, and greatly improves the robustness of the system in harsh communication environments.

[0078] In one embodiment, the sensor node is equipped with a triaxial accelerometer and is configured to generate a wake-up signal when the sensed acceleration value exceeds a preset threshold. The sensor node is configured to perform normal data acquisition at a first sampling frequency, and when the detected acceleration value of a vibration event exceeds the preset threshold, it automatically switches to a second sampling frequency for high-speed data acquisition. After high-speed data acquisition continues for a preset duration, it returns to the normal data acquisition mode.

[0079] Specifically, the preset threshold is set to several times the gravitational acceleration. This threshold is mainly used to identify the instantaneous impact of the sensor being dropped from the drone onto the ground, thereby distinguishing between free-fall impact and environmental vibrations after landing. The wake-up signal triggers the microcontroller's wake-up pin via an external interrupt, enabling the system to quickly switch from deep sleep mode to active mode.

[0080] Specifically, the first sampling frequency is set to 1 time / 10min to meet the low power consumption requirements of long-term monitoring; the second sampling frequency is set to 1 time / s to restore the detailed features of the vibration waveform.

[0081] Specifically, the duration of high-speed data acquisition is set to 30 seconds, based on the vibration duration during a typical landslide's pre-disaster phase, to ensure the capture of complete disaster waveforms.

[0082] This embodiment, through an acceleration threshold-triggered impact wake-up mechanism and a dual sampling frequency switching strategy, ensures that the sensor nodes can capture landing impacts and environmental vibration events in a timely manner. It also significantly reduces device power consumption and extends battery life through the low-frequency conventional acquisition mode, while the high-speed acquisition mode ensures the integrity and accuracy of vibration event waveforms, providing reliable data support for subsequent slope condition analysis.

[0083] In one embodiment, the data processing center subsystem is configured to: extract feature values ​​characterizing slope stability from vibration signals, acceleration signals, and pressure signals, the feature values ​​including vibration frequency spectrum, peak acceleration, pressure change gradient, and root mean square value; compare the extracted feature values ​​with a preset safety threshold, and generate a corresponding level of early warning information when it is determined that more than a preset proportion of sensors in the same area continuously report abnormalities.

[0084] Specifically, the vibration frequency spectrum is obtained by performing frequency domain transformation on the time-domain vibration signal using the Fast Fourier Transform algorithm, and is used to analyze the natural frequency changes of the slope. The root mean square value is obtained by squaring, integrating, averaging, and taking the square root of the acceleration sampling points, reflecting the average energy level of the vibration signal.

[0085] Specifically, a safety threshold for the root mean square (RMS) value of vibration is set as the basic threshold for single-sensor anomaly detection;

[0086] Furthermore, the 24-hour variation trend of air pressure data was calculated to identify the long-term evolution characteristics of the internal stress field of the slope.

[0087] Specifically, the warning level is divided into multiple levels, and the warning information of the corresponding level is generated based on the spatial aggregation ratio of abnormal sensors.

[0088] Preferably, the early warning information is simultaneously notified to on-site management personnel through one or more of the following methods: SMS, platform pop-up window, and audible and visual alarm, clearly indicating the location of the potential disaster, the risk level, and handling suggestions.

[0089] This embodiment effectively avoids false alarms caused by single sensor false alarms by extracting multi-dimensional features such as vibration frequency spectrum, peak acceleration, pressure change gradient, and root mean square value, combined with multi-sensor collaborative threshold judgment logic based on gridded areas. This improves the accuracy and reliability of the warning results and provides a scientific basis for slope safety management.

[0090] In one embodiment, the data processing center subsystem is configured to train and analyze historical and real-time data using machine learning models to identify anomalous patterns characterizing slope instability. These machine learning models include Long Short-Term Memory (LSTM), Support Vector Machine (SVM), or Random Forest. Acceleration time-series data from each sensor over a preset period are input into a pre-trained LSTM model to predict slope instability trends. When the high-risk probability value output by the model exceeds the critical warning probability and exhibits spatial clustering anomalies, a high-level warning is generated. The patrol frequency of the relay drone is dynamically adjusted based on the warning level, and encrypted patrols are initiated when the warning level reaches a preset high level.

[0091] Specifically, the LSTM model is pre-trained in an offline environment using a large amount of historical slope monitoring data to learn the data distribution patterns under normal working conditions.

[0092] Specifically, the initial cruise frequency of the relay drone is set according to operational requirements to maintain a stable data transmission rhythm;

[0093] Furthermore, spatial clustering anomalies are analyzed and identified using clustering algorithms to identify anomalous sensor nodes. Additionally, the encrypted patrol strategy includes increasing the number of flights on the original patrol path or shortening the patrol interval from once every 2 days to once every 1 hour, thereby increasing the data acquisition density during high-risk periods.

[0094] This embodiment inputs the acceleration time-series data of the sensor over the past 7 days into a pre-trained LSTM model for inference, which can identify early creep signs of slopes that are difficult to detect by traditional threshold comparison methods. Combined with the abnormal clustering characteristics determined by the DBSCAN spatial clustering algorithm, the warning critical probability is set to 80%, which further reduces the false alarm rate. At the same time, the patrol frequency of the relay UAV is dynamically adjusted according to the warning level, thereby achieving optimized allocation of monitoring resources.

[0095] In one embodiment, the system further includes a closed-loop feedback mechanism for early warning: packaging the full-chain multimodal data of each early warning event into a digital case library, using the case library data to retrain the machine learning early warning model, optimizing the early warning threshold and recognition accuracy; and dynamically adjusting the threshold and logic of the early warning rules based on the false alarms and missed alarms of the early warning.

[0096] Specifically, the full-chain multimodal data includes raw waveform data collected by sensors, flight logs recorded by relay drones, early warning messages generated by the data processing center, and verification reports from on-site inspectors. The digital case library is managed using a relational database, establishing a multi-dimensional index of early warning events and time, space, and environmental factors.

[0097] Furthermore, the model retraining adopts an incremental learning strategy, automatically triggering the model fine-tuning process whenever a new typical case is added.

[0098] Furthermore, the dynamic adjustment of the early warning rules is based on confusion matrix analysis. If the false alarm rate is too high, the threshold is appropriately raised; if the false negative rate is too high, the threshold is appropriately lowered or the logical judgment conditions are optimized, so that the system has the ability to self-evolve.

[0099] This embodiment constructs a digital case library by packaging multimodal data across the entire chain, it uses an incremental learning strategy to iteratively optimize the machine learning model, and dynamically adjusts the warning rule thresholds and logic based on the confusion matrix analysis results. This enables the system to self-evolve and continuously optimize its performance, thereby improving the accuracy and reliability of the warning system and forming a virtuous cycle of "monitoring-warning-feedback-optimization".

[0100] Secondly, such as Figure 2 As shown, the wireless measurement method for slope monitoring based on UAV spraying relay includes the following steps:

[0101] Step S1: The drone spraying subsystem flies along a preset path, and the spraying mechanism releases multiple sensor nodes to deploy the sensor nodes on the slope surface.

[0102] Step S2: The UAV relay subsystem flies over the sensor deployment area according to the preset cruise path, collects data from the sensor nodes, and sends the aggregated data to the data processing center through the second communication module.

[0103] Step S3: Receive and process the aggregated data through the data processing center to conduct risk assessment and graded early warning.

[0104] This embodiment achieves rapid, full-area deployment of sensors on dangerous slopes by using drones for spraying, overcomes terrain signal blockages through aerial relays to ensure stable data transmission, and completes intelligent analysis and early warning with the help of a data processing center, effectively improving disaster response efficiency.

[0105] In one embodiment, step S1: The drone spraying subsystem flies along a preset path, and the spraying mechanism releases multiple sensor nodes to deploy the sensor nodes on the slope surface. This specifically includes the following steps:

[0106] Step S11: Construct a digital terrain model of the slope area at the ground control station, and automatically generate a bow-shaped traversal track covering the entire area based on the terrain undulation features, providing a path reference for the UAV's autonomous flight and targeted spraying.

[0107] Specifically, by importing the digital elevation model of the slope area, the software automatically plans a bow-shaped flight path covering the entire area based on the terrain elevation data, ensuring uniform deployment density.

[0108] Preferably, during the trajectory planning stage, the flight attitude and wind speed interference of the UAV are simulated simultaneously to smooth the trajectory inflection points, reduce attitude jitter during flight, and improve the positional accuracy of the drop point.

[0109] Step S12: Inject the pre-planned flight path file into the UAV flight control system, control the UAV to perform autonomous flight missions along the predetermined path, and trigger the execution command of the spraying mechanism based on real-time positioning information.

[0110] Specifically, the planned flight path file is uploaded to the flight control system of the six-rotor UAV. After the UAV takes off, it enters the automatic navigation mode. The flight control computer compares the GPS positioning information with the coordinates of the preset drop point in real time and sends a pulse trigger signal to the drop mechanism every 20m of flight.

[0111] Preferably, the remaining battery power and attitude stability of the drone are monitored in real time during flight, and the flight attitude and speed are automatically adjusted when encountering sudden crosswinds to ensure that the timing of the trigger signal transmission is precisely matched with the spatial position.

[0112] Step S13: Responding to the trigger command issued by the flight control system, drive the spraying mechanism to perform a physical release action, and orderly deploy the sensor nodes to the slope surface to complete the full coverage under a single mission.

[0113] Specifically, after receiving a trigger signal, the stepper motor of the spraying mechanism drives the rotating wheel to rotate 60°, which moves a sensor node from the groove in the storage bin to the discharge port, and then falls to the slope surface under the action of gravity; all 50 sensor nodes can be deployed in a single flight mission.

[0114] Preferably, the distribution mechanism can also employ a low-pressure spring ejector or a pneumatic push rod array to adapt to the deployment requirements of sensor nodes of different specifications, thereby improving the system's compatibility with different application scenarios. This embodiment automatically plans a full-coverage bow-shaped trajectory using a digital elevation model, combined with precise drop control logic triggered by GPS positioning, achieving uniform sensor deployment in the target area with a deployment density meeting monitoring requirements. The compatible design of multiple distribution mechanisms enhances the adaptability of the solution, completely avoiding the operational risks of personnel entering steep and dangerous slopes, and laying a spatial foundation for subsequent acquisition of full-area monitoring data.

[0115] In one embodiment, step S2: The UAV relay subsystem flies over the sensor deployment area according to a preset cruise path, collects data from the sensor nodes, and sends the aggregated data to the data processing center through the second communication module. This specifically includes the following steps:

[0116] Step S21: Control the relay UAV to cruise along a low-altitude path covering the sensor deployment area, and continuously broadcast data collection beacons to the outside world through the airborne LoRa collection module to provide a unified trigger condition for data upload from ground sensor nodes.

[0117] Specifically, the relay UAV's cruising altitude is dynamically adjusted according to the slope's undulations and elevation differences, and a single flight can cover ground sensor nodes within a preset radius; the broadcast beacon frame contains the relay UAV's unique identifier and current timestamp, which is used by ground sensor nodes to identify valid information sources.

[0118] Preferably, the beacon broadcast period is set to 1 second to ensure that sensor nodes in the coverage edge area can also stably receive the trigger signal, thereby improving the beacon reach rate.

[0119] Step S22: Control the working mode of the ground sensor nodes so that they trigger the wake-up process after receiving a valid data collection beacon, complete the directional transmission of locally stored monitoring data, and avoid signal collision problems caused by concurrent uploads from multiple nodes.

[0120] Specifically, in step S1, the sensor nodes that were dropped to the ground have been awakened from deep sleep mode and entered the normal monitoring state after sensing the impact of the landing. The ground sensor nodes are in deep sleep mode during non-working periods to reduce power consumption. After being awakened, the stored vibration, acceleration and pressure data are sent to the relay drone after a random delay according to the preset backoff algorithm to avoid mutual interference between signals in the same frequency band.

[0121] Preferably, after the sensor node is woken up, it first performs a wireless link quality test. Data transmission is only initiated when the link signal-to-noise ratio is higher than a preset threshold, thereby further improving the success rate of data upload.

[0122] Step S23: Locally cache the monitoring data collected by the relay drone, and dynamically switch the backhaul channel according to the real-time communication link quality to ensure that the monitoring data is completely transmitted to the remote data processing center.

[0123] Specifically, the relay drone temporarily stores the received monitoring data in the local eMMC flash memory storage module; it obtains the real-time signal strength by reading the AGC register value at the bottom layer of the 4G DTU or satellite short message module; when the 4G signal strength is lower than the preset threshold, it automatically switches to the satellite communication link to complete the data backhaul; if the communication interruption exceeds the preset duration, the relay drone records the sequence number of the transmitted data, and performs breakpoint resume transmission based on the sequence number after the communication link is restored to avoid data loss.

[0124] Preferably, the cached monitoring data is lightly compressed and encoded before being transmitted back, reducing transmission bandwidth usage and improving transmission efficiency in weak network environments.

[0125] This embodiment uses a relay drone to broadcast beacons at low altitudes, triggering ground sensors to wake up and upload data on demand, which significantly reduces the power consumption of ground nodes. The airborne relay design effectively overcomes the limitations of complex terrain on wireless signals, and the adaptive communication switching and breakpoint resume mechanism ensures the complete transmission of data in harsh environments. A single flight can complete the collection of sensor data over a large area, improving the reliability and efficiency of data transmission.

[0126] In one embodiment, step S3: receiving and processing the aggregated data through the data processing center to perform risk assessment and graded early warning, specifically includes the following steps:

[0127] Step S31: After the server daemon in the data center detects the data inflow into the network port, it starts a data parsing thread to verify, decompress, and format the raw byte stream, remove abnormal and dirty data, and obtain a clean time-series monitoring dataset.

[0128] Specifically, missing data is filled using linear interpolation, and abnormal jump data is smoothed using filtering to ensure the continuity and reliability of the time-series monitoring dataset.

[0129] Step S32: Extract multidimensional feature values ​​characterizing slope stability from the clean time-series monitoring dataset, including vibration frequency spectrum, peak acceleration, pressure change gradient and root mean square value. The pressure change gradient is obtained by calculating the pressure difference per unit time and is used to reflect the stress change trend of the slope.

[0130] Step S33: Input the time-series monitoring dataset after feature extraction into the pre-trained machine learning model for risk reasoning, combine spatial clustering analysis to determine the possibility of slope instability, and output the corresponding level of early warning decision.

[0131] Specifically, the acceleration time-series data of each sensor over the past 7 days is used as input parameters and input into a pre-trained LSTM model for inference calculation. The LSTM model is trained based on a large amount of historical slope monitoring data and can identify abnormal patterns of slope creep.

[0132] Furthermore, the DBSCAN density clustering algorithm is used to determine the spatial clustering of the inference results, thereby filtering out anomalous sensor nodes with spatial correlation and eliminating false alarms caused by isolated noise points.

[0133] Specifically, the warning levels are divided into three levels: a yellow warning is generated when more than 30% of the sensors report anomalies in the same area, and a red warning is generated when more than 70% of the sensors report anomalies in the same area.

[0134] Preferably, while pushing out early warning information, on-site management personnel are notified through one or more of the following methods: SMS, platform pop-up window, and audible and visual alarm, clearly indicating the location, risk level, and handling suggestions of potential disasters.

[0135] This embodiment, through multi-dimensional feature extraction combined with inference calculation of a pre-trained LSTM model, can identify early signs of slope instability that are difficult to capture by traditional thresholding methods; the spatial clustering judgment logic effectively reduces the probability of false alarms from a single sensor, and the 80% critical probability setting balances the sensitivity and accuracy of the early warning; in actual testing, it successfully issued an accurate early warning 36 hours before significant slope displacement, providing a sufficient window for personnel evacuation and emergency response, and improving the foresight and practicality of slope disaster early warning.

[0136] In one specific embodiment, this embodiment is applied to a steep rock slope along a mountain highway. The slope has recently experienced frequent rainfall and is at risk of landslides. It is difficult for personnel to reach and has complex terrain. The system achieves monitoring through the coordinated operation of four subsystems: impact-resistant spraying sensors, UAV spraying system, UAV relay system, and data processing center.

[0137] The impact-resistant spill sensor, serving as the data sensing terminal, is a spherical shape approximately 10cm in diameter. Its outer shell is made of impact-resistant polycarbonate material, covered with a 5mm thick rubber buffer layer. The sensor is in deep sleep mode within the storage hopper. When spilled and landed, a triaxial accelerometer or attitude sensor / impact switch detects an impact signal or sudden attitude change far exceeding gravitational acceleration. This signal triggers the MCU's interrupt pin, waking the sensor. Upon waking, the sensor automatically collects slope vibration, acceleration, and pressure data at a default frequency of once per 10 minutes. When a vibration event is detected where the acceleration value continuously exceeds a preset threshold, it automatically switches to a high-speed acquisition mode of once per second for 30 seconds. The collected data is packaged by the MCU and transmitted periodically via a LoRa module at a cycle of once per hour. In vibration event trigger mode, data is transmitted instantly, ensuring the timeliness of critical data.

[0138] Furthermore, the drone-based spraying system serves as the deployment platform, employing a hexacopter drone with a maximum payload capacity of 10kg, its bottom suspending a dedicated spraying mechanism. During implementation, a bow-shaped flight path covering the entire 500m × 200m rectangular monitoring area is generated, along with 50 evenly distributed spraying points. After the flight path file is uploaded to the flight control system, the drone automatically takes off. Every 20m of flight, the flight control computer sends a pulse trigger signal to the spraying mechanism. A servo motor drives the hatch to open, and a stepper motor rotates a rotary wheel 60°, displacing individual sensors from their recesses to the discharge port, where they fall under gravity. A single flight completes the deployment of all 50 sensor nodes, eliminating the need for personnel to engage in any hazardous activities.

[0139] The drone relay system serves as a data transmission bridge, employing a quadcopter drone with longer endurance. Its payload is a relay communication gateway, integrating a high-sensitivity LoRa data collection module, a 4G DTU module, and an embedded computer. After sensor deployment, the relay drone flies along a low-altitude path covering the deployment area, with a flight altitude set at approximately 50 meters. During flight, the onboard LoRa data collection module continuously broadcasts a data collection beacon containing a unique identifier and timestamp. Ground sensors receive the beacon and upload data with random delays based on a backoff algorithm. A single flight can cover sensor nodes within a radius of approximately 300 meters. The collected data is temporarily stored in the embedded computer and then transmitted back to the data processing center via a 4G network. When the 4G signal strength falls below a preset threshold, it automatically switches to a satellite link. In the event of a communication interruption, the sequence number of the transmitted data is recorded, and transmission is resumed once the link is restored, effectively ensuring data integrity.

[0140] The data processing center is implemented by a server running dedicated data analysis software. The server receives data streams via a Socket interface, performs parsing, cleaning, and interpolation, and extracts the root mean square (RMS) value of acceleration data, dominant frequency components, duration of vibration events, maximum amplitude, and 24-hour trend of air pressure data. A safe threshold of 0.5g for vibration RMS is set. Acceleration time-series data from each sensor over the past 7 days are input into a pre-trained LSTM model for inference. This model, trained on a large amount of historical data, can identify abnormal slope creep patterns. When the high-risk probability value output by the model exceeds the critical warning probability of 80% and the DBSCAN algorithm determines that abnormal sensors are spatially clustered, a corresponding level of warning information is generated. The warning information is simultaneously notified to on-site management personnel via SMS, platform pop-ups, and audible and visual alarms, achieving closed-loop management from monitoring to warning.

[0141] This embodiment, through the coordinated operation of the four subsystems mentioned above, completed the sensor deployment on the dangerous slope within 2 hours and stably acquired monitoring data through twice-daily drone relay patrols. On the 15th day of operation, the system successfully issued an accurate early warning based on the LSTM model 36 hours before significant slope displacement occurred, gaining valuable time for personnel and equipment evacuation and traffic control, fully verifying the practicality, reliability, and advanced nature of this solution.

[0142] Thirdly, embodiments of this application also provide a readable storage medium.

[0143] This application stores a slope monitoring wireless measurement program based on UAV spraying relay on a readable storage medium, wherein when the UAV spraying relay-based slope monitoring wireless measurement program is executed by a processor, it implements the steps of the UAV spraying relay-based slope monitoring wireless measurement method as described above.

[0144] The method implemented when the slope monitoring wireless measurement program based on UAV spraying relay is executed can be referred to in the various embodiments of the slope monitoring wireless measurement method based on UAV spraying relay in this application, and will not be repeated here.

[0145] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A wireless measurement system for slope monitoring based on UAV spraying relay, characterized in that, It includes an impact-resistant spray sensor subsystem, a drone spraying subsystem, a drone relay subsystem, and a data processing center subsystem; The impact-resistant spray sensor subsystem includes multiple sensor nodes, which are configured to wake up after sensing an impact, collect and transmit vibration, acceleration and pressure data of the slope. The drone spraying subsystem is configured to fly according to a preset path, control the spraying mechanism to release multiple sensor nodes, and deploy the sensor nodes on the slope surface. The UAV relay subsystem includes a relay UAV and a relay communication gateway mounted thereon. The relay communication gateway includes a first communication module and a second communication module, and is configured to fly over the sensor deployment area according to a preset cruise path, collect data from the sensor nodes, and send the aggregated data to the data processing center subsystem through the second communication module. The data processing center subsystem is configured to receive and process the aggregated data, and to perform risk assessment and graded early warning.

2. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 1, characterized in that, The drone spraying subsystem is configured to fly along a bow-shaped flight path and trigger a spraying action once every preset flight distance based on GPS positioning information. The outer shell of the sensor node is spherical or streamlined, made of engineering plastics or composite materials, and covered with a flexible buffer layer. A buffer mechanism is provided between the outer shell and the internal module. The buffer mechanism is at least one of a suspension structure, a silicone potting layer, or a spring damping system.

3. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 2, characterized in that, The scattering mechanism includes a storage bin, a controllable discharge gate driven by a servo motor, and a dispensing mechanism driven by a stepper motor. The sorting mechanism is configured to detach the sensor node from the storage bin and drop it by rotating it at a preset angle; The UAV spraying subsystem is configured to deploy sensors to a preset target area in a single flight mission.

4. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 1, characterized in that, The relay drone is configured to fly along a low-altitude path covering the area where the sensors are deployed. The first communication module is a LoRa collection module with a receiving sensitivity higher than a preset sensitivity threshold. It is configured to continuously broadcast data collection beacons and listen to the channel, and receive data sent by ground sensor nodes in response to the beacons. The relay communication gateway also includes an embedded computer configured to temporarily store and convert the collected data.

5. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 4, characterized in that, The second communication module is a 4G DTU module or a satellite short message module, configured to adaptively switch between cellular network and satellite communication based on real-time signal strength; When communication between the relay drone and the data processing center is interrupted for more than a preset time, the relay drone will temporarily store the aggregated data in the local storage module and resume transmission from the point of interruption after communication is restored.

6. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 1, characterized in that, The sensor node is equipped with an accelerometer or a gyroscope and is configured to generate a wake-up signal when the sensed acceleration value or attitude angle change exceeds a preset threshold. The sensor node is configured to perform normal data acquisition at a first sampling frequency. When the acceleration value of the vibration event detected exceeds a preset threshold, it automatically switches to a second sampling frequency for high-speed data acquisition. After high-speed data acquisition continues for a preset duration, it returns to the normal data acquisition mode.

7. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 1, characterized in that, The data processing center subsystem is configured as follows: Feature values ​​characterizing slope stability are extracted from vibration signals, acceleration signals, and pressure signals. These feature values ​​include vibration frequency spectrum, peak acceleration, pressure change gradient, and root mean square value. By comparing the extracted feature values ​​with the preset safety threshold, when it is determined that more than a preset proportion of sensors in the same area continuously report abnormalities, a corresponding level of warning information is generated.

8. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 7, characterized in that, The data processing center subsystem is configured to use machine learning models to train and analyze historical and real-time data to identify abnormal patterns that characterize slope instability. The machine learning models include Long Short-Term Memory Network (LSTM), Support Vector Machine (SVM), or Random Forest. Input the acceleration time series data of each sensor for a preset period of time into the pre-trained LSTM model to predict the slope instability trend. When the high-risk probability value output by the model exceeds the critical warning probability and shows spatial clustering anomalies, a high-level warning is generated. The patrol frequency of relay drones is dynamically adjusted according to the warning level, and encrypted patrols are initiated when the warning level reaches a preset high level.

9. The slope monitoring wireless measurement system based on UAV spraying relay as described in claim 8, characterized in that, The system also includes an early warning closed-loop feedback mechanism: The entire chain of multimodal data for each early warning event is packaged into a digital case library. The case library data is used to retrain the machine learning early warning model to optimize the early warning threshold and recognition accuracy. The thresholds and logic of the early warning rules are dynamically adjusted based on the false alarms and missed alarms.

10. A wireless measurement method for slope monitoring based on UAV spraying relay, characterized in that, Includes the following steps: The drone spraying subsystem flies along a preset path and controls the spraying mechanism to release multiple sensor nodes, which are then deployed on the slope surface. The drone relay subsystem flies over the sensor deployment area according to a preset cruise path, collects data from the sensor nodes, and sends the aggregated data to the data processing center through the second communication module. The data processing center receives and processes the aggregated data to conduct risk assessment and tiered early warning.

Images