Dynamic monitoring method and monitoring network system of unmanned aerial vehicle remote sensing and ground sensor

Abstract

This invention relates to the field of intelligent monitoring and discloses a dynamic monitoring method and monitoring network system for UAV remote sensing and ground sensors. The method includes generating a multidimensional temporal map of the monitoring area, extracting anomaly markers from the multidimensional temporal map, matching the geographic information of the anomaly markers, allocating inspection UAVs and ground sensors for task execution, assembling the ground sensors into a virtual sensor cluster and assigning monitoring tasks, forming a UAV swarm of inspection UAVs for inspection tasks, extracting the address information of each inspection UAV in the UAV swarm, planning the inspection paths of the inspection UAVs, the ground sensors executing monitoring tasks and generating real-time data packets, and the ground sensors waking up when the inspection UAV executes its inspection task and flies to the corresponding ground sensor, with the ground sensor sending real-time data packets to the inspection UAV.

Description

Technical Field

[0001] This invention relates to a dynamic monitoring method and monitoring network system for unmanned aerial vehicle (UAV) remote sensing and ground sensors, belonging to the field of intelligent monitoring technology. Background Technology

[0002] Existing forestry and water resource inspection and monitoring systems typically employ either periodic drone patrols along flight routes or static detection using fixed ground sensor networks. When anomalies are detected, signals are sent to a central gateway for alarm activation. This approach has inherent drawbacks, including the presence of spatiotemporal blind spots in drone patrols, where events occurring during flight intervals cannot be captured in a timely manner, resulting in incomplete monitoring. Fixed sensor networks have limited coverage, leading to inaccurate monitoring when events occur outside the sensor's range or near the sensor's monitoring edge. Furthermore, sensors require continuous monitoring and communication capabilities, and their built-in power supply makes them prone to power outages and offline operation, resulting in the loss of monitoring data for certain areas and hindering intelligent collaboration and adequate monitoring response.

[0003] Meanwhile, the data from drones and ground sensors are independent of each other and rely on a central gateway for linkage. There is a lack of deep fusion and joint reasoning between drones and sensors. Moreover, the central gateway relies on manual interpretation, resulting in a long signal chain from perception to response, low monitoring efficiency, and long signal processing time in emergency situations, making it impossible to make timely prevention and control responses. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic monitoring method and monitoring network system for UAV remote sensing and ground sensors. The system enables monitoring and perception by dynamically networking sensors and UAVs to form a virtual sensor cluster, dynamically planning the patrol path of the UAV, and wirelessly powering the sensors through the UAV.

[0005] To achieve the above objectives / to solve the above technical problems, the present invention is implemented using the following technical solution.

[0006] On one hand, the present invention provides a dynamic monitoring method for unmanned aerial vehicle (UAV) remote sensing and ground sensors, comprising: The monitoring data and remote sensing data are acquired and correlated to generate a multidimensional temporal map of the monitoring area. Anomaly markers are extracted from the multidimensional temporal map and matched with the geographic information of the anomaly markers. The inspection drones and ground sensors are allocated to perform the tasks. The ground sensors are assembled into a virtual sensor cluster and monitoring tasks are assigned. The inspection drones used to perform the inspection tasks are assembled into a drone cluster. The address information of each inspection drone in the drone cluster is extracted, and the inspection path of the inspection drones is planned. The ground sensor performs monitoring tasks and generates real-time data packets. When the inspection drone performs inspection tasks and flies to the corresponding ground sensor, it wakes up the ground sensor. The ground sensor sends real-time data packets to the inspection drone. The inspection drone parses the real-time data packets to update the multidimensional temporal map. Based on the updated multidimensional temporal map, it dynamically adjusts the inspection path to determine the next stage of the inspection path for the inspection drone to perform the inspection task. The specific inspection paths for the planned inspection drones include: Configure the affected area, and form an abnormal region with the location of the abnormal marker as the center and the affected area as the radius. The affected area represents the range of influence of the abnormal state corresponding to the abnormal marker in the natural environment. Ground sensors whose monitoring ranges intersect with the affected area are extracted, and the extracted ground sensors are grouped into a virtual sensor cluster; The process of assembling the virtual sensor cluster includes: For ground sensors whose monitoring range covers the affected area, mark the ground sensor as a stationary point, and mark the remaining ground sensors with monitoring overlap as fulcrum points; When there are no ground sensors whose monitoring range covers the affected area, ground sensors whose monitoring ranges intersect with the affected area and whose monitoring ranges intersect with each other are extracted to form a sensor set. The intersection area between the monitoring range of each ground sensor in the sensor set and the affected area is identified. The ground sensor with the largest intersection area is marked as a stationary point, and the other ground sensors in the sensor set are marked as fulcrum points. An edge is constructed between the stationary point and the pivot point. The monitoring task of the stationary point is assigned as real-time monitoring to monitor the anomaly markers and anomaly areas. The monitoring task of the pivot point is assigned as verification monitoring to verify the real-time monitoring results of the stationary point. The pivot point sends real-time data packets to the stationary point through the associated edge.

[0007] Preferably / furthermore, the planned inspection path for the inspection drone also includes: Topographic data and vegetation types of geographic locations are extracted from remote sensing data and monitoring data; The flight range is determined based on terrain data and vegetation type, and the flight attitude is determined by combining the flight range and vegetation type. Establish inspection boundaries and determine the inspection path of the inspection drone based on the inspection boundaries.

[0008] Preferably / further, the station receives at least one real-time data packet, and when the station receives multiple real-time data packets, it performs data analysis to determine the real-time data packet to be sent to the inspection drone; The system uses data analysis to reverse verify whether there are any anomalies in the real-time data packets sent by the sender's pivot. If an anomaly is detected, the system sends the analyzed real-time data packets to the pivot.

[0009] Preferably / further, the inspection drone is equipped with an optoelectronic pod and a side-looking radar to obtain the identification angle of the optoelectronic pod and the fan sweep angle of the side-looking radar, extract the slope and aspect from the terrain data, determine the flight layer in the abnormal area for the inspection drone to perform monitoring tasks according to the vegetation type, and fit the flight layer with the terrain data to determine the flight range. The slope normal of the terrain is determined by the slope and aspect. The identification angle and fan sweep angle are adjusted, the angle between the inspection drone and the slope normal is calculated, and the pitch angle of the inspection drone is adjusted to adjust the flight attitude of the inspection drone. Extract the coverage area of ​​the ground sensors used as the base, fuse terrain data to form a safety boundary and construct a constrained flight space, and optimize the flight space and form an inspection boundary through the mission performance of the ground sensors.

[0010] Preferably / furthermore, adjusting the pitch angle of the inspection drone includes: Compare the current heading and slope of the inspection drone; If the heading is consistent with the steepest ascent direction of the slope, then set the desired pitch angle equal to the slope value; If the heading is consistent with the contour line direction, then set the desired pitch angle to zero; For other headings, the desired pitch angle is determined by interpolation based on the angle between the heading and the steepest direction.

[0011] Preferably / furthermore, constructing the inspection boundary includes: Determine the link space for establishing a communication connection between the inspection drone and the ground sensors used as anchor points; Identify the distance between the flight level and the station; Based on the intersection of the distance and the link space, the altitude of the flight layer or the communication power of the station is adjusted to construct the inspection boundary.

[0012] Preferably / furthermore, planning the inspection path also includes: Identify blank areas within the monitoring area that cannot be covered by surface sensors, and prioritize inspections based on the remaining battery power or mission status of the inspection drone. Based on the characteristics of the blank area, the distribution of the station and pivot points, and the inspection priority, a specific inspection path is assigned to each inspection drone in the drone cluster.

[0013] Preferably / furthermore, waking up the ground sensor includes: After entering the communication range of the target ground sensor, the inspection drone sends a wake-up signal carrying the unique identifier of the target ground sensor. The target ground sensor receives and parses the wake-up signal through a low-power wake-up receiver. After the identifier is successfully matched, it is woken up from the sleep state and starts the main communication module. The target ground sensor establishes a communication link with the inspection drone and performs identity authentication.

[0014] On the other hand, the present invention provides a dynamic monitoring network system for UAV remote sensing and ground sensors, used to execute the above-mentioned dynamic monitoring method for UAV remote sensing and ground sensors, including: The integrated monitoring module is used to acquire monitoring data and remote sensing data, and perform correlation analysis to generate a multidimensional temporal map of the monitoring area; An anomaly identification module is used to extract anomaly markers from the multidimensional temporal map and match the geographic information of the anomaly markers; The task allocation module is used to allocate inspection drones and ground sensors to perform tasks, organize ground sensors into virtual sensor clusters and assign monitoring tasks, and build a drone cluster to perform inspection tasks. The path planning module is used to extract the address information of each inspection drone in the drone cluster and plan the inspection path in combination with the geographic information. The data interaction module is used to control the ground sensor to perform monitoring tasks and generate real-time data packets. When the inspection drone flies to the corresponding ground sensor, it wakes up the ground sensor and receives the real-time data packets sent by it. The map update module is used to parse the real-time data packets to update the multidimensional temporal map, and dynamically adjust the inspection path based on the updated map to determine the path for executing the next stage of the inspection task.

[0015] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: 1. By analyzing the correlation between monitoring data and remote sensing data, a multidimensional temporal map is generated. This dual data verification method eliminates the inaccuracy problem of single data monitoring and accurately identifies anomaly markers and corresponding geographic information. At the same time, a virtual sensor cluster is constructed, dividing ground sensors into stationary points and pivot points. Through a hierarchical monitoring mode of real-time monitoring at stationary points and verification monitoring at pivot points, the monitoring data is cross-linked and verified. Stationary points can also perform fusion analysis and reverse verification of pivot point data, promptly correcting the accuracy of pivot point monitoring and further improving data reliability. 2. By constructing a virtual sensor cluster, autonomous data interaction and fusion between ground sensors are realized. At the same time, a direct communication link between the UAV and the sensor is established, making the UAV a direct receiving and processing node for sensor data. This eliminates the intermediate link of manual interpretation by the central gateway and significantly shortens the signal chain from perception to response. 3. A dynamic adjustment closed loop for the entire process has been constructed. After receiving real-time data packets from ground sensors, the UAV will parse the data and update the multi-dimensional temporal map. Based on the updated map, the next stage of the inspection path will be dynamically adjusted, allowing the inspection task to be precisely adjusted according to the real-time situation of the monitored area. The virtual sensor cluster can be dynamically constructed according to the location of the anomaly marker. The allocation of the station and the pivot point will also be flexibly determined according to the monitoring association between the sensor and the anomaly area. It can quickly respond to sudden anomalies and complete the monitoring task allocation. In addition, the reverse verification mechanism of the station and the pivot point data can detect and correct the sensor monitoring anomalies in real time. The UAV's breakpoint resume mechanism ensures the integrity of data transmission, which greatly shortens the signal processing time of the system in the event of an emergency and enables timely prevention and control response. Attached Figure Description

[0016] Figure 1 The diagram shown is a system flowchart of the dynamic monitoring method of the present invention. Figure 2 The diagram shows a system flow chart for adjusting the pitch angle and flight attitude of an inspection drone. Figure 3 The diagram shows a system flow diagram of building a ground sensor cluster and generating data packets to be sent to the inspection drone. Figure 4 The diagram shows a flowchart of the system for determining the inspection path of an inspection drone. Figure 5 The diagram shown is a schematic of the dynamic monitoring network system of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0018] The term "and / or" simply describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0019] Example 1 refer to Figures 1 to 4As shown in this embodiment, a dynamic monitoring method for UAV remote sensing and ground sensors is introduced, including acquiring monitoring data and remote sensing data, performing correlation analysis to generate a multidimensional temporal map of the monitoring area, extracting anomaly markers from the multidimensional temporal map, and matching the geographic information of the anomaly markers; The inspection drones and ground sensors are allocated to perform the tasks. The ground sensors are assembled into a virtual sensor cluster and monitoring tasks are assigned. The inspection drones used to perform the inspection tasks are assembled into a drone cluster. The address information of each inspection drone in the drone cluster is extracted, and the inspection path of the inspection drones is planned. The ground sensor performs monitoring tasks and generates real-time data packets. When the inspection drone performs an inspection task and flies to the corresponding ground sensor, it wakes up the ground sensor. The ground sensor sends real-time data packets to the inspection drone. The inspection drone parses the real-time data packets to update the multidimensional temporal map. Based on the updated multidimensional temporal map, it dynamically adjusts the inspection path to determine the next stage of the inspection path for the inspection drone to perform the inspection task.

[0020] Specifically, monitoring data and remote sensing data are correlated and analyzed through spatiotemporal alignment and feature fusion to obtain a multidimensional temporal map. Spatiotemporal processing and physical constraints are used to obtain the causal relationship between the monitoring data and remote sensing data, in order to determine whether there are any anomalies in the monitored area. The monitoring data includes identification information, temperature information, and smoke information; the remote sensing data includes geographic information and remote sensing images. The identification information of the ground sensors is extracted to determine their geographical location. The geographic information of the inspection drone is also extracted. The identification information of the ground sensors and the geographic information of the inspection drone are spatiotemporally processed to identify the ground sensors and inspection drones performing monitoring in the same area, and the ground sensors and inspection drones are spatiotemporally bound together. Temperature and smoke information from ground sensors under spatiotemporal binding are extracted, and remote sensing images from inspection drones are also extracted. The monitoring range of the ground sensors is configured, which represents the effective monitoring area of ​​the ground sensors. The ground sensors are also configured with temperature thresholds. When the temperature information exceeds the temperature threshold, a temperature anomaly is generated and an anomaly signal marked as 1 is generated. When an anomaly signal is detected, it is identified whether the smoke information is 1. When both the anomaly signal and the smoke information are 1, it is determined that an anomaly has occurred in the monitoring area, and the data is fed back to the monitoring data and marked as 1. The system synchronously analyzes remote sensing images, extracts multiple frames of real-time remote sensing images acquired by the inspection drone, compares them with standard remote sensing images, identifies the presence of smoke features in the remote sensing images, and marks the remote sensing image frame as 1 when smoke features are present, indicating that the inspection drone's remote sensing data detected an anomaly in the corresponding monitoring area. The system synchronously feeds back to the remote sensing data and marks the remote sensing data as 1. The correlation analysis results of the monitoring data and remote sensing data are used to form a multidimensional temporal map. Data in the multidimensional temporal map where both the monitoring data and remote sensing data are 1 are marked as anomalies. The identification information of the ground sensors is extracted, and geographic information is generated based on the monitoring area of ​​the corresponding ground sensors.

[0021] It should be noted that the monitoring data identifies whether there are anomalies within the monitoring area of ​​the ground sensor. Through correlation analysis of remote sensing data and monitoring data, the inspection drones monitoring the same area as the ground sensor are identified, and it is determined whether there are anomalies in the remote sensing data. By combining the multidimensional temporal map determined by the correlation analysis of monitoring data and remote sensing data, it is determined whether there are anomalies in the monitoring area, eliminating the inaccuracy of single data. The accuracy of monitoring is improved by correlation verification, and the geographical information of anomalies is accurately detected.

[0022] Furthermore, the specific inspection paths for the planned inspection drones include: A virtual sensor cluster is constructed, and ground sensors whose monitoring range and affected range have overlapping monitoring ranges are extracted. Based on the coverage relationship between the monitoring range and affected range and the mutual intersection relationship between the monitoring ranges of multiple ground sensors, stationary points and pivot points are determined, and edges between nodes are constructed. The monitoring task of the stationary points is assigned as real-time monitoring, which is used to monitor the monitoring area in real time. The monitoring task of the pivot points is assigned as verification monitoring, which is used to verify the results of the real-time monitoring of the stationary points. The pivot points send real-time data packets to the stationary points through the associated edges.

[0023] Topographic data and vegetation type of geographic location are extracted from remote sensing data and monitoring data. Flight range is determined based on topographic data and vegetation type, and flight attitude is determined by combining flight range and vegetation type. Establish inspection boundaries and determine the inspection path of the inspection drone based on the inspection boundaries.

[0024] Specifically, the configuration allocation mechanism includes establishing a virtual sensing cluster and allocating stationary points and pivots when anomaly markers exist in the multidimensional temporal spectrum, specifically including: When extracting anomaly markers from a multidimensional temporal map, an anomaly region is formed with the location of the anomaly marker as the center and the affected area as the radius. The affected area represents the range of influence of the anomaly state corresponding to the anomaly marker in the natural environment. The monitoring association between ground sensors and the affected area is identified, and ground sensors with monitoring associations are extracted. The monitoring association represents the monitoring overlap between the monitoring range of the ground sensors and the affected area.

[0025] For ground sensors whose monitoring range covers the affected area, the ground sensor is marked as a stationary point, and the remaining ground sensors with monitoring associations are marked as pivot points; wherein, the stationary point performs real-time monitoring of anomaly markers and anomaly areas, and the pivot point performs verification monitoring of the real-time monitoring results of the stationary point.

[0026] When no ground sensor has a monitoring range covering the affected area, ground sensors whose monitoring ranges intersect with the affected area and whose monitoring ranges also intersect with each other are extracted to form a sensor set. The intersection area between the monitoring range of each ground sensor in the sensor set and the affected area is identified, and the ground sensor with the largest intersection area is marked as a stationary point, while the other ground sensors in the sensor set are marked as pivot points. Thus, when a single ground sensor cannot cover an abnormal area, the stationary point undertaking the main real-time monitoring task is determined by the cross-linked monitoring of the abnormal area by multiple ground sensors, and the other pivot points provide verification monitoring data.

[0027] Configure a transmission threshold, which represents the maximum communication distance between a ground sensor acting as a fulcrum and a ground sensor acting as a stationary point, allowing for a direct communication connection. Determine whether the communication distance between the marked fulcrum and the stationary point is within the transmission threshold range. If it is within the transmission threshold range, construct an edge between the marked fulcrum and the stationary point. If it exceeds the transmission threshold range, determine whether the marked fulcrum has a fulcrum with an edge connected to the stationary point in its respective intersection association or composite association. If a fulcrum with an edge connected to the stationary point exists, construct an edge between the marked fulcrum and its adjacent fulcrum in the intersection association. If the marked fulcrum is in a composite association, exclude the marked fulcrum, extract the fulcrum with the largest intersection area among the other fulcrums and the affected area, and construct an edge between the marked fulcrum and the fulcrum with the largest intersection area.

[0028] It should be noted that when anomaly markers are present, the ground sensors participating in anomaly area monitoring are determined by identifying the monitoring correlation between the monitoring range and the affected area of ​​the ground sensors. Stationary points and pivot points are determined by judging whether the monitoring range covers the affected area, or, if not, comparing the intersection area between multiple ground sensors and the affected area. Then, based on the transmission threshold and the adjacent monitoring relationships between pivot points, and combined with the intersection area of ​​other pivot points and the affected area, edges for sending real-time data packets are constructed. Ground sensors acting as stationary points integrate real-time data and handle data interaction with the inspection drone. Ground sensors acting as pivot points send verification monitoring data and cross-reference the real-time monitoring results of the stationary points for verification, thus providing a ground node foundation for subsequent inspection boundary construction and inspection path planning.

[0029] Specifically, the allocation mechanism also includes setting up virtual sensor clusters and assigning anchor points and pivot points during routine inspections to improve the monitoring of the monitored area by ground sensors and inspection drones during routine inspections. This includes: The monitoring area is divided into grids. Ground sensors whose monitoring range falls entirely within the grid are selected, while those intersecting with other grid areas are removed. This is because after grid division, ground sensors intersecting with other grid areas may simultaneously monitor at least two adjacent grid areas, potentially misleading data from a single grid area. Therefore, ground sensors completely falling within the grid are considered valid, reducing the possibility of data errors. From the valid ground sensors, one is randomly selected as a stationary point, and the others serve as pivot points. A polling time is configured, and the stationary point is randomly polled according to the polling time. Edges are established between pivot points and stationary points, and all pivot points communicate with stationary points in a single manner, forming point-to-point edges. Pivot points whose communication distance with stationary points exceeds the transmission threshold are identified and marked as isolated pivot points. No edges for communication are built between isolated pivot points and stationary points or pivot points. When cooperating with UAV inspections, the inspection UAV establishes communication connections with stationary points and also establishes separate communication connections with isolated pivot points, thereby improving the accuracy of data transmission between the inspection UAV and ground sensors during daily inspections.

[0030] It should be noted that since the monitoring area is mainly forested, it is often large and has significant elevation differences. Therefore, each grid cell has a relatively large area and may contain multiple ground sensors. To address this, removing ground sensors that intersect with adjacent grid cells can improve the collection of monitoring data within each grid cell. For areas where ground sensors are not monitoring within a grid cell, patrol drones can be deployed for inspection. A polling time is set to randomly select a station within the grid cell to eliminate errors in data transmission between a single station and a pivot point. When the relationships between a station, pivot, and edge are unique, the patrol drone's path will also be unique, resulting in a fixed inspection area and insufficient inspection of edge areas. By randomly polling station cells, the patrol drone's path can be dynamically adjusted periodically, improving the accuracy of regional monitoring using both monitoring and remote sensing data.

[0031] Furthermore, the inspection drone is equipped with an optoelectronic pod and a side-looking radar to obtain the identification angle of the optoelectronic pod and the fan-scan angle of the side-looking radar, extract the slope and aspect from the terrain data, determine the flight layer in the abnormal area for the inspection drone to perform monitoring tasks based on the vegetation type, and fit the flight layer with the terrain data to determine the flight range. The slope and aspect are used to determine the slope normal of the terrain. The identification angle and fan sweep angle are adjusted, the angle between the inspection drone and the slope normal is calculated, and the pitch angle of the inspection drone is adjusted to regulate the flight attitude of the inspection drone.

[0032] Specifically, vegetation type includes vegetation height and density. Height is used to distinguish vegetation types, including but not limited to shrubs, trees, and arbores. Since different vegetation types have different heights, combining this with density allows for a generalization of vegetation types, adjusting the flight layer determination for the inspection drone under different vegetation types. The lower limit of the flight layer is set as the average height of the vegetation canopy plus a safety redundancy distance to ensure the inspection drone does not collide with the canopy top. The upper limit of the flight layer is set to be subject to airspace control and communication constraints. Slope characterizes the degree of inclination of the slope relative to the horizontal plane, and aspect characterizes the projection direction of the slope normal onto the horizontal plane. During flight, the onboard mission computer acquires the slope and aspect at the current projection point of the inspection drone in real time and calculates the normal vector of the slope at the projection point based on geometric relationships. This normal vector is perpendicular to the slope and points towards the sky, with its horizontal component pointing downhill and its vertical component determined by the slope magnitude. The flight layer is fitted with terrain data to generate a three-dimensional surface as the expected flight range of the inspection drone. Within the flight range, the inspection drone dynamically adjusts its flight altitude according to the terrain undulations to maintain a constant relative distance from the ground surface in order to obtain accurate remote sensing data.

[0033] Based on the geometric relationship between the real-time heading of the inspection drone and the terrain data, the adjustment amount of the expected pitch angle of the inspection drone and the target pointing angle of the electro-optical pod and the side-looking radar are calculated.

[0034] To determine the desired pitch angle of the inspection drone, compare the drone's current heading with its slope: If the heading is consistent with the steepest upward direction of the slope, the desired pitch angle is equal to the slope value, so that the aircraft pitches up to adapt to the upward slope. If the heading is along the contour lines, the desired pitch angle is zero, and the aircraft remains level. For other headings, the desired pitch angle is calculated by cosine interpolation of the angle between the heading and the steepest direction to achieve a smooth transition.

[0035] For the optoelectronic pod, the intersection of the inspection drone's direct bottom and the slope is taken as the monitoring target point. Combining the inspection drone's current spatial position, the adjusted pitch angle, and the geographic coordinates of the target point, the relative direction vector of the target point in the inspection drone's body coordinate system is calculated through coordinate transformation. The azimuth and pitch angles of this vector in the body coordinate system are the target angles that the optoelectronic pod needs to rotate to, ensuring that the camera's optical axis in the optoelectronic pod is always perpendicular to the slope to obtain orthophotos.

[0036] For side-looking radar, based on the current altitude, attitude, and slope normal direction of the inspection drone, the theoretical intersection range of the radar beam and the slope is calculated, and the expected values ​​of the radar scanning center angle and scanning width are determined so that the beam coverage area matches the slope to be measured.

[0037] It should be noted that the inspection drone's flight control system uses the difference between the desired pitch angle and the current pitch angle as the control variable. By adjusting the drone's wings, it achieves a smooth attitude transition, ensuring that the drone always flies in a near-parallel attitude above the slope. The gimbal servo mechanism of the electro-optical pod drives the camera to rotate according to the calculated target angle, ensuring that the target point always falls within the center of the field of view, and continuously updates the pointing at a preset frequency to ensure that image acquisition always follows terrain changes. The side-looking radar control system adjusts the beam's start and end angles in real time based on the calculated scanning parameters, ensuring that the scanning range accurately covers the target slope section.

[0038] The regulation of these three execution channels is not independent of each other, but is subject to the synergistic effect of multiple geometric constraints: Changes in the UAV's pitch angle alter its relative height and slant range to the slope, thus affecting the calculation benchmark for side-looking radar scanning parameters and the geometric model of the electro-optical pod's pointing. Angle adjustments for the electro-optical pod and side-looking radar must be performed within their respective mechanical or electronic limits. If these limits are exceeded, the UAV's flight path is adjusted first, including but not limited to changing flight altitude or lateral offset, to meet the detection requirements of the electro-optical pod and test radar. The onboard mission computer runs an integrated collaborative control algorithm, taking real-time terrain data, the UAV's motion status, and current sensor parameters as input. Through iterative calculation, it outputs control commands, which are sent to the flight control system, the electro-optical pod gimbal, and the radar control system, forming a closed-loop process of parameter calculation and adjustment under terrain perception. The entire adjustment process continues in flight, enabling the UAV to adaptively traverse undulating terrain while ensuring that the electro-optical pod and test radar always collect monitoring data with optimal pointing, providing continuous and accurate dynamic perception.

[0039] Among them, the angle between the heading and the steepest ascent direction is calculated:

[0040] The included angle is normalized to Within the range, but the periodicity of the cosine function allows us to directly use the difference. The current heading angle is the projected azimuth angle of the inspection drone's flight direction on the horizontal plane. It represents the steepest ascent direction angle, measured clockwise or counterclockwise with true north as the reference reference.

[0041] The desired pitch angle is determined by the cosine difference:

[0042] in, Slope is a measure of the angle at which a slope is tilted relative to the horizontal plane. , The pitch angle represents the desired pitch angle. When the inspection drone is adjusted to the corresponding desired pitch angle, a positive value indicates that the nose is tilted up and rising along the slope; a negative value indicates that the nose is tilted down and falling along the slope. The unit is consistent with the slope.

[0043] Furthermore, the link space for communication between the inspection drone and the station is configured in a spherical shape. When the link spaces of the inspection drone and the station intersect, the communication connection between the inspection drone and the station can be completed, the distance between the station and the flight layer can be identified, and the inspection boundary can be constructed. Specifically, this includes: When the inspection drone is inspecting at the lower limit of the flight layer, if there is an intersection of the link space between the flight layer and the station, the flight layer shall be used as the inspection boundary. If there is no intersection between the flight layer and the station, identify the distance difference between the link space of the inspection drone and the station. If the distance difference is within the safe redundancy distance, lower the lower limit of the flight layer by the distance difference to adjust the flight layer, and use the adjusted flight layer as the inspection boundary. If the distance difference is outside the safe redundancy distance, maintain the lower limit of the flight layer, and use the flight layer as the inspection boundary. Increase the communication power of the station to improve the link space range of the station, so that communication connection can be established between the station and the inspection drone. While ensuring the flight safety margin of the inspection drone, adjust the ground sensors of the station to establish communication connection.

[0044] Specifically, determining the inspection path of the inspection drone based on the inspection boundary includes: Extract blank areas within the monitoring area of ​​the ground sensor. Blank areas represent areas not covered by the monitoring range of the ground sensor. Also extract stationary points and isolated pivot points. The inspection priority of inspection drones is determined based on their remaining battery power and whether they are performing a task. When inspecting a station, drones stored in the mother bin have a higher priority than those performing an inspection task. The number and area of ​​blank areas, as well as the number of station and isolated support points, are obtained to determine the number of drones to perform the inspection and to form a drone swarm.

[0045] Prioritize assigning inspection drones located in the mother warehouse to perform inspections along the station points, determine the starting and ending points of the station points on the inspection path, mark the mother warehouse located at the station point ending point, mark the station points on the inspection path, determine the shortest patrol flight path between station points, control the inspection drones to take off from the mother warehouse closest to the station point starting point, inspect the station points sequentially along the shortest patrol flight path to the station point ending point, and return to the mother warehouse closest to the station point ending point to complete the inspection of the station points; When the inspection drones located in the mother warehouse are insufficient to inspect all the stations, the inspection drones that are currently performing inspection tasks and have sufficient power are assigned to inspect the stations. The number of stations that can be inspected is determined based on the power level, and station markers are formed. The shortest patrol path between station markers is determined, and the inspection drone is controlled to fly to the nearest station marker. After inspecting the station along the shortest patrol path, it finally returns to the mother warehouse. Preferably, when performing inspections of the station, a wake-up power threshold is configured. When allocating inspection drones to the station, the wake-up power threshold is used for allocation so that when the inspection drone flies to the station, there is enough power to wake up the station and enable data communication.

[0046] After determining the inspection drones at the inspection stations, a corresponding number of inspection drones are matched according to the number of stations. The inspection drones are assigned to inspect the fulcrums determined by the edge relationship of each station. The fulcrum with the shortest edge is the inspection starting point, and the fulcrum closest to the mother warehouse is the inspection ending point. The shortest path between the fulcrums is identified, and the fulcrum path is formed in sequence. The inspection drones are controlled to perform inspections of the areas where the fulcrums are located along the fulcrum path.

[0047] The blank areas are continuously divided into grids based on the maximum inspection area of ​​the inspection drones. During division, an S-shaped path is used to divide the blank areas, extracting continuous grid regions. The starting and ending points of each grid region are determined, along with the number and area of ​​each grid region. Inspection drones closest to the lowest priority level are selected from the inspection priority list and assigned to fly sequentially from the starting point to the ending point along the divided path. The flight distance of each assigned inspection drone within a grid region is determined based on its remaining battery power. This process ensures that only drones within each grid region are assigned to perform inspections, thus completing the inspection of the blank areas.

[0048] It should be noted that during routine inspections, inspections of outposts, support points, and blank areas are conducted simultaneously to improve monitoring efficiency. The remaining battery power determines the patrol distance of the inspection drones, ensuring sufficient power for each area's inspection. When patrolling outposts, the drones need to wake up and establish communication with the outpost; therefore, the allocation of drones for outpost patrols is based on both the wake-up battery threshold and remaining battery power. When blank areas are discontinuous, inspections are assigned to continuous blank areas to avoid conflicts between inspection drones patrolling across blank areas and those patrolling outposts or support points, minimizing the impact of overlapping areas during drone patrols.

[0049] Furthermore, the station receives at least one real-time data packet. When the station receives multiple real-time data packets, it performs data analysis and determines the real-time data packet to be sent to the inspection drone. The system uses data analysis to reverse verify whether there are any anomalies in the real-time data packets sent by the sender's pivot. When an anomaly is detected, the system sends the analyzed real-time data packets to the pivot to verify and correct the pivot's detection accuracy.

[0050] The station receives real-time data packets sent by each support point, performs timestamp alignment and outlier removal on the data packets, extracts the historical accuracy factor, real-time consistency factor, energy state factor, and sensor health factor for each support point, and comprehensively calculates the dynamic fusion weight of each support point. The historical accuracy factor is determined based on the long-term deviation statistics between the support point and the station's verification value. The real-time consistency factor is determined based on the closeness of the support point's data to the data of other support points at the current moment. The energy state factor is determined based on the remaining power of the support point. The sensor health factor is determined based on the self-check status of the support point. The dynamic fusion weight of each support point is normalized.

[0051] Based on the normalized dynamic fusion weights, the monitoring values ​​of each support point are weighted and averaged to obtain the fused monitoring value. The fusion confidence score is calculated by the deviation between the monitoring values ​​of each support point and the fused monitoring value. The fused monitoring value, fusion confidence score, support point identifiers participating in the fusion, and support point data statistics are encapsulated to generate a summary data packet to be sent to the inspection drone.

[0052] The standardized deviation between the monitored value and the fused monitored value of each pivot point is calculated. When the standardized deviation exceeds a preset threshold, the data of that pivot point is determined to be abnormal. Based on the continuous trend of the deviation, the abnormality type is classified as sudden abnormality, trend abnormality, or fault abnormality. For pivot points determined to have abnormal data, the stationary point generates a reverse verification data packet. The reverse verification data packet contains at least the pivot point identifier, the abnormality type identifier, the current fused monitored value, and the suggested calibration parameters. The stationary point sends the reverse verification data packet to the corresponding pivot point through the edge of the virtual sensor cluster. After receiving the reverse verification data packet, the pivot point adjusts its own measurement model according to the calibration parameters, updates its local statistical records, and replies with processing confirmation information to the stationary point.

[0053] The system maintains a database of the historical performance of each fulcrum. After each fusion, it updates the historical accuracy factor of the fulcrum based on its standardized deviation. It uses an exponentially weighted moving average method to achieve dynamic evolution and updates the measurement standard deviation of the fulcrum.

[0054] It should be noted that the stationary point merges and summarizes the real-time data packets sent by the monitoring point to achieve a comprehensive assessment of the status within the monitoring area. This summarized data packet is then sent to the inspection drone, reducing data transmission redundancy and improving the communication efficiency between the ground sensors and the inspection drone. The monitoring status of the monitoring point is dynamically adjusted using reverse verification, further improving the efficiency and accuracy of monitoring.

[0055] Furthermore, when the inspection drone flies to the corresponding ground sensor during its inspection mission, it will wake up the ground sensor in the following ways: The inspection drone flies to the link space range of the target ground sensor according to the planned path, and sends a wake-up signal at a preset frequency and power through the airborne radio frequency transmitter. The wake-up signal contains the unique identifier of the target ground sensor. The ground sensor has a built-in ultra-low power wake-up receiver that continuously monitors the ambient radio frequency signals. When it receives the wake-up signal and parses out the identifier that matches the identifier stored in its own memory, it generates a wake-up signal to trigger the ground sensor's main controller and main communication module to power on and start from sleep mode. After the ground sensor is activated, it establishes a communication link with the inspection drone through the main communication module and performs two-way authentication. The data transmission method of the communication link is wireless data transmission. After authentication, the ground sensors will send the cached real-time data packets to the inspection drone via the communication link; After receiving and confirming the data packet, the inspection drone sends a hibernation command to the ground sensor. The ground sensor then shuts down its main communication module and non-essential peripherals, re-entering a hibernation state that only retains the wake-up receiver.

[0056] Specifically, the wake-up signal adopts an on / off keying modulation method and includes a preamble, a synchronization word, at least one target sensor identifier, and a cyclic redundancy check field. When the inspection drone sends the wake-up signal, it dynamically adjusts the transmission power according to the estimated distance to the target ground sensor to ensure effective wake-up only within the preset communication range and avoids accidental triggering of irrelevant sensors at a distance.

[0057] Two-way authentication specifically includes: The ground sensor sends its own identifier and a random number to the inspection drone. The inspection drone encrypts the random number using a pre-shared key and returns it. The ground sensor decrypts and verifies the data. If the verification is successful, the communication link is confirmed to be secure, and data transmission continues.

[0058] During the process of receiving real-time data packets, the inspection drone supports a breakpoint resume mechanism. The ground sensor records the sequence number of the sent data packets, and the inspection drone records the sequence number of the received data packets. When communication is unexpectedly interrupted and the connection is re-established, the unfinished data packets are resumed from the point of interruption.

[0059] Furthermore, a power threshold is configured for the ground sensors. When the power of the ground sensors is lower than the power threshold, a power replenishment command is sent. When the inspection drone is performing an inspection task, it is also used to receive the power replenishment command sent by the ground sensors, update the inspection path, descend to the vicinity of the ground sensors, and perform wireless power replenishment through contact, reducing the occurrence of power outages and offline situations of the ground sensors.

[0060] Example 2 Based on the same inventive concept as Embodiment 1, this embodiment introduces a dynamic monitoring network system for UAV remote sensing and ground sensors, which includes: The integrated monitoring module is used to acquire monitoring data and remote sensing data, and perform correlation analysis to generate a multidimensional temporal map of the monitoring area; An anomaly identification module is used to extract anomaly markers from the multidimensional temporal map and match the geographic information of the anomaly markers; The task allocation module is used to allocate inspection drones and ground sensors to perform tasks, organize ground sensors into virtual sensor clusters and assign monitoring tasks, and build a drone cluster to perform inspection tasks. The path planning module is used to extract the address information of each inspection drone in the drone cluster and plan the inspection path in combination with the geographic information. The data interaction module is used to control the ground sensor to perform monitoring tasks and generate real-time data packets. When the inspection drone flies to the corresponding ground sensor, it wakes up the ground sensor and receives the real-time data packets sent by it. The map update module is used to parse the real-time data packets to update the multidimensional temporal map, and dynamically adjust the inspection path based on the updated map to determine the path for executing the next stage of the inspection task.

[0061] In summary, this invention generates multi-dimensional temporal maps through the correlation analysis of monitoring data and remote sensing data. This dual data verification eliminates the accuracy issues associated with single-data monitoring, accurately identifying anomaly markers and corresponding geographic information. Simultaneously, a virtual sensor cluster is constructed, dividing ground sensors into stationary points and pivot points. Through a hierarchical monitoring mode of real-time monitoring at stationary points and verification monitoring at pivot points, cross-linking and verification of monitoring data are achieved. Stationary points can also perform fusion analysis and reverse verification of pivot point data, promptly correcting the accuracy of pivot point monitoring and further improving data reliability. For areas not covered by ground sensors, drone inspection paths are planned, overcoming the limitations of limited coverage in traditional fixed sensor networks and the spatiotemporal blind spots inherent in periodic drone inspections, thus achieving full coverage of the monitoring area.

[0062] By constructing virtual sensor clusters, autonomous data interaction and fusion between ground sensors are achieved. Simultaneously, a direct communication link is established between UAVs and sensors, making UAVs direct receivers and processors of sensor data. This eliminates the intermediate step of manual interpretation through a central gateway, significantly shortening the signal chain from perception to response. Furthermore, through spatiotemporal binding of sensors and UAVs, collaborative inspections of UAV swarms based on task priorities, and dynamic formation of virtual sensor clusters in response to anomalies, deep fusion and joint inference between UAV remote sensing and ground sensing are realized. This solves the problems of independent data and poor coordination between the two, improving the overall response efficiency of the monitoring system to anomalies.

[0063] The flight path and attitude of the inspection drone are adaptively adjusted based on the terrain data and vegetation type of the monitored area. By fitting the flight range determined by the terrain and vegetation, and combining the pitch angle dynamically adjusted by the slope and aspect, the drone can fly in accordance with undulating terrain. At the same time, the monitoring angles of the electro-optical pod and side-looking radar are adjusted in coordination to ensure that remote sensing data is always collected from the best perspective. The inspection boundary constructed based on the station communication link space and the flight layer, as well as the inspection priority divided by the drone's remaining battery power and mission status, makes the inspection path planning more reasonable and avoids the problem of insufficient inspection caused by fixed paths. The flight layer is equipped with a safe redundancy distance for the vegetation canopy, and the safety boundary formed by integrating the terrain creates a constrained flight space, effectively preventing the drone from colliding with the vegetation and ensuring the safety of the inspection flight.

[0064] The ground sensor employs a low-power operating mechanism that uses a patrol drone for wake-up. Normally, it only listens for signals via a low-power wake-up receiver, while the main communication module remains in sleep mode. It only activates after the drone arrives and sends a matching wake-up signal, completely resolving the rapid power consumption issue caused by continuous monitoring and communication in traditional sensors. Simultaneously, a ground sensor power threshold is configured. When the sensor's power is insufficient, the drone can receive a power-up command and update its path for wireless power replenishment, effectively preventing data loss due to sensor offline status and significantly extending the ground sensor's uninterrupted monitoring endurance.

[0065] The monitoring system constructs a dynamic adjustment closed loop throughout the entire process. After receiving real-time data packets from ground sensors, the UAV parses the data and updates the multi-dimensional temporal map. Based on the updated map, it dynamically adjusts the next stage of the inspection path, allowing the inspection task to be precisely adjusted according to the real-time situation of the monitored area. Virtual sensor clusters can be dynamically constructed based on the location of anomaly markers, and the allocation of anchor points and pivot points is flexibly determined based on the monitoring correlation between sensors and anomaly areas, enabling rapid response to sudden anomalies and completion of monitoring task allocation. Through the reverse verification mechanism of anchor point data against pivot point data, sensor monitoring anomalies can be detected and corrected in real time. The UAV's breakpoint resume mechanism ensures the integrity of data transmission, significantly shortening the signal processing time in emergency situations and enabling timely prevention and control responses.

[0066] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0067] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0068] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0069] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0070] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A dynamic monitoring method using UAV remote sensing and ground sensors, characterized in that, include: The monitoring data and remote sensing data are acquired and correlated to generate a multidimensional temporal map of the monitoring area. Anomaly markers are extracted from the multidimensional temporal map and matched with the geographic information of the anomaly markers. The inspection drones and ground sensors are allocated to perform the tasks. The ground sensors are assembled into a virtual sensor cluster and monitoring tasks are assigned. The inspection drones used to perform the inspection tasks are assembled into a drone cluster. The address information of each inspection drone in the drone cluster is extracted, and the inspection path of the inspection drones is planned. The ground sensor performs monitoring tasks and generates real-time data packets. When the inspection drone performs inspection tasks and flies to the corresponding ground sensor, it wakes up the ground sensor. The ground sensor sends real-time data packets to the inspection drone. The inspection drone parses the real-time data packets to update the multidimensional temporal map. Based on the updated multidimensional temporal map, it dynamically adjusts the inspection path to determine the next stage of the inspection path for the inspection drone to perform the inspection task. The specific inspection paths for the planned inspection drones include: Configure the affected area, and form an abnormal region with the location of the abnormal marker as the center and the affected area as the radius. The affected area represents the range of influence of the abnormal state corresponding to the abnormal marker in the natural environment. Ground sensors whose monitoring ranges intersect with the affected area are extracted, and the extracted ground sensors are grouped into a virtual sensor cluster; The process of assembling the virtual sensor cluster includes: For ground sensors whose monitoring range covers the affected area, mark the ground sensor as a stationary point, and mark the remaining ground sensors with monitoring overlap as fulcrum points; When there are no ground sensors whose monitoring range covers the affected area, ground sensors whose monitoring ranges intersect with the affected area and whose monitoring ranges intersect with each other are extracted to form a sensor set. The intersection area between the monitoring range of each ground sensor in the sensor set and the affected area is identified. The ground sensor with the largest intersection area is marked as a stationary point, and the other ground sensors in the sensor set are marked as fulcrum points. An edge is constructed between the stationary point and the pivot point. The monitoring task of the stationary point is assigned as real-time monitoring to monitor the anomaly markers and anomaly areas. The monitoring task of the pivot point is assigned as verification monitoring to verify the real-time monitoring results of the stationary point. The pivot point sends real-time data packets to the stationary point through the associated edge.

2. The dynamic monitoring method for UAV remote sensing and ground sensors according to claim 1, characterized in that, The planned inspection routes for inspection drones also include: Topographic data and vegetation types of geographic locations are extracted from remote sensing data and monitoring data; The flight range is determined based on terrain data and vegetation type, and the flight attitude is determined by combining the flight range and vegetation type. Establish inspection boundaries and determine the inspection path of the inspection drone based on the inspection boundaries.

3. The dynamic monitoring method for UAV remote sensing and ground sensors according to claim 1, characterized in that, The station receives at least one real-time data packet. When the station receives multiple real-time data packets, it performs data analysis and determines the real-time data packet to be sent to the inspection drone. The system uses data analysis to reverse verify whether there are any anomalies in the real-time data packets sent by the sender's pivot. If an anomaly is detected, the system sends the analyzed real-time data packets to the pivot.

4. The dynamic monitoring method for UAV remote sensing and ground sensors according to claim 2, characterized in that, The inspection drone is equipped with an optoelectronic pod and a side-looking radar. It obtains the recognition angle of the optoelectronic pod and the fan sweep angle of the side-looking radar, extracts the slope and aspect from the terrain data, determines the flight layer in the abnormal area for the inspection drone to perform monitoring tasks based on the vegetation type, and fits the flight layer with the terrain data to determine the flight range. The slope normal of the terrain is determined by the slope and aspect. The identification angle and fan sweep angle are adjusted, the angle between the inspection drone and the slope normal is calculated, and the pitch angle of the inspection drone is adjusted to adjust the flight attitude of the inspection drone. Extract the coverage area of ​​the ground sensors used as the base, fuse terrain data to form a safety boundary and construct a constrained flight space, and optimize the flight space and form an inspection boundary through the mission performance of the ground sensors.

5. The dynamic monitoring method for UAV remote sensing and ground sensors according to claim 4, characterized in that, Adjusting the pitch angle of the inspection drone includes: Compare the current heading and slope of the inspection drone; If the heading is consistent with the steepest ascent direction of the slope, then set the desired pitch angle equal to the slope value; If the heading is consistent with the contour line direction, then set the desired pitch angle to zero; For other headings, the desired pitch angle is determined by interpolation based on the angle between the heading and the steepest direction.

6. The dynamic monitoring method for UAV remote sensing and ground sensors according to claim 5, characterized in that, Constructing the inspection boundary includes: Determine the link space for establishing a communication connection between the inspection drone and the ground sensors used as anchor points; Identify the distance between the flight level and the station; Based on the intersection of the distance and the link space, the altitude of the flight layer or the communication power of the station is adjusted to construct the inspection boundary.

7. The dynamic monitoring method for UAV remote sensing and ground sensors according to claim 6, characterized in that, The planned inspection route also includes: Identify blank areas within the monitoring area that cannot be covered by surface sensors, and prioritize inspections based on the remaining battery power or mission status of the inspection drone. Based on the characteristics of the blank area, the distribution of the station and support points, and the inspection priority, a specific inspection path is assigned to each inspection drone in the drone cluster.

8. The dynamic monitoring method for UAV remote sensing and ground sensors according to claim 1, characterized in that, Waking up the ground sensor includes: After entering the communication range of the target ground sensor, the inspection drone sends a wake-up signal carrying the unique identifier of the target ground sensor. The target ground sensor receives and parses the wake-up signal through a low-power wake-up receiver. After the identifier is successfully matched, it is woken up from the sleep state and starts the main communication module. The target ground sensor establishes a communication link with the inspection drone and performs identity authentication.

9. A dynamic monitoring network system for unmanned aerial vehicle (UAV) remote sensing and ground sensors, used to perform the method as described in any one of claims 1 to 8, characterized in that, include: The integrated monitoring module is used to acquire monitoring data and remote sensing data, and perform correlation analysis to generate a multidimensional temporal map of the monitoring area; An anomaly identification module is used to extract anomaly markers from the multidimensional temporal map and match the geographic information of the anomaly markers; The task allocation module is used to allocate inspection drones and ground sensors to perform tasks, organize ground sensors into virtual sensor clusters and assign monitoring tasks, and build a drone cluster to perform inspection tasks. The path planning module is used to extract the address information of each inspection drone in the drone cluster and plan the inspection path in combination with the geographic information. The data interaction module is used to control the ground sensor to perform monitoring tasks and generate real-time data packets. When the inspection drone flies to the corresponding ground sensor, it wakes up the ground sensor and receives the real-time data packets sent by it. The map update module is used to parse the real-time data packets to update the multidimensional temporal map, and dynamically adjust the inspection path based on the updated map to determine the path for executing the next stage of the inspection task.

Images