Cloud image-based air cannon operation path planning method and system

Abstract

The application provides a cloud-based air cannon operation path planning method and system, relating to the technical field of artificial weather modification, comprising constructing a cloud layer space-time evolution field by analyzing cloud map data, predicting its evolution sequence and identifying a priority operation position. A time sequence constraint graph of the operation position is constructed to generate an operation sequence, and the disturbance effect of previous operations is considered to dynamically update the planning. Finally, a space-time constraint network is constructed for dynamic planning search to obtain an optimal operation path. The application realizes accurate and dynamic blocking of the propagation path of the cloud layer, and improves the timeliness and effect of artificial weather modification operations.

Description

Technical Field

[0001] This invention relates to the field of weather modification technology, and in particular to a method and system for planning the operational path of an air cannon based on cloud maps. Background Technology

[0002] In the field of weather modification, using aerial cannons and other equipment to intervene in cloud formations is a common practice. Current technologies typically rely on the analysis of static or quasi-static data such as weather radar images and satellite cloud images to plan operations. The conventional approach involves operators selecting one or more suitable operational locations based on experience and visual judgment of the current cloud formations, direction of movement, and speed, and then planning a rough path connecting these locations. The core objective of this path planning is to enable the equipment to move to the predetermined operational points, while the spatiotemporal constraints of movement between these points are relatively simple to consider.

[0003] This method, based on empirical judgment and static analysis, has significant limitations. Clouds are complex, dynamically evolving systems whose morphology, internal structure, and propagation paths constantly change over time. Conventional methods for analyzing cloud image data often remain at the level of describing the instantaneous state, lacking quantitative and refined predictions of the future spatiotemporal evolution trends of clouds. The selection of operational locations is often based on empirical identification of key areas within the current cloud layer, failing to accurately match the time required for the operational equipment to move from its current location to a candidate location with the time it takes for the cloud to dynamically propagate to that location. This results in the critical portion of the target cloud system having already moved out of its intended range by the time the operational equipment arrives, significantly weakening the operational effectiveness.

[0004] Furthermore, when planning continuous operational paths involving multiple work locations, each work point is typically treated as an independent node. Planning primarily considers the geographical distance or time cost of moving the work equipment, neglecting the disturbances the work itself causes to the cloud system. In reality, preceding operations alter local atmospheric conditions, influencing the subsequent evolution path and intensity of clouds. This dynamic feedback effect triggered by operations is generally overlooked in existing planning frameworks, resulting in temporally suboptimal planned operation sequences, and even subsequent work locations becoming invalid due to unexpected changes in cloud conditions. Therefore, a path planning method that deeply integrates cloud dynamic evolution prediction with the spatiotemporal constraints of the operational process is urgently needed to improve the scientific rigor and effectiveness of operations. Summary of the Invention

[0005] The present invention provides a method and system for air cannon operation path planning based on cloud maps, which can solve the problems in the prior art.

[0006] A first aspect of the present invention provides a method for planning the operational path of an air cannon based on cloud maps, comprising:

[0007] Obtain cloud map data for the target area;

[0008] Spatiotemporal coupling analysis of cloud image data is performed to extract dynamic propagation characteristic parameters of cloud layers. By tracking the cloud centroid trajectory and extrapolating the boundary diffusion situation, a spatiotemporal evolution field of cloud layers is constructed to obtain the spatial distribution evolution sequence of cloud layers in future time periods.

[0009] Based on the spatial distribution evolution sequence, cloud propagation path nodes are identified. The time-sensitivity matching relationship between candidate operation locations and propagation path nodes is calculated by combining the operation response delay of the air cannon operation device. Based on the time-sensitivity matching relationship, priority operation locations that have a spatial blocking effect on cloud propagation are identified.

[0010] A temporal constraint diagram of the operation positions is constructed based on the spatial connectivity and temporal relationship of the priority operation positions in the cloud propagation path. An operation sequence is generated based on the temporal constraint diagram. The perturbation effect of the preceding operation positions in the operation sequence on the spatiotemporal evolution field of the cloud is calculated. The timing matching relationship of the subsequent operation positions is recalculated based on the updated spatiotemporal evolution field of the cloud.

[0011] Each work location in the work sequence is taken as a path node. A spatiotemporal constraint network is constructed based on the matching relationship between movement time cost and timeliness. Dynamic programming search is performed on the spatiotemporal constraint network to obtain the work path of the air cannon working device passing through each path node in sequence.

[0012] Control the air cannon operating device to perform operations along the operating path.

[0013] In one optional embodiment, spatiotemporal coupling analysis is performed on cloud image data to extract dynamic propagation characteristic parameters of the cloud layer. By tracking the cloud centroid trajectory and extrapolating the boundary diffusion situation, a spatiotemporal evolution field of the cloud layer is constructed, resulting in a spatial distribution evolution sequence of the cloud layer in the future, including:

[0014] Spatial gradient field calculation is performed on the cloud image data to extract the spatial gradient vector field of cloud density distribution. Then, the spatial gradient vector field is decomposed into divergence and curl to obtain the divergence propagation component and rotation propagation component of the cloud layer.

[0015] The divergent propagation component and the rotational propagation component are vector-synthesized to obtain the dynamic propagation characteristics of the cloud layer.

[0016] Based on the aforementioned dynamic propagation characteristics, the centroid coordinates of the cloud layer are calculated. Time series analysis is then performed on the centroid coordinates, and the velocity vector and acceleration vector of the centroid displacement are fitted to obtain the cloud layer centroid trajectory.

[0017] Based on the dynamic propagation characteristics, the normal diffusion rate at the cloud boundary point is determined, and the cloud boundary is extrapolated along the cloud centroid trajectory with a time step to obtain the diffusion pattern of the cloud boundary;

[0018] The cloud centroid trajectory and the diffusion pattern of the cloud boundary are mapped in spatiotemporal coordinates to construct the spatiotemporal evolution field of the cloud. The spatiotemporal evolution field of the cloud is then extrapolated along the time axis to obtain the spatial distribution evolution sequence.

[0019] In one optional embodiment, cloud propagation path nodes are identified based on the spatial distribution evolution sequence. The timeliness matching relationship between candidate operational locations and propagation path nodes is calculated by combining the operational response delay of the air cannon operating device. Based on the timeliness matching relationship, priority operational locations that have a spatial blocking effect on cloud propagation are identified, including:

[0020] Feature direction extraction is performed on the spatial distribution evolution sequence to obtain the direction vector of cloud propagation. The spatiotemporal evolution field of the cloud is divided into equally spaced spatial slices along the direction vector. Local maxima of cloud density are identified in each spatial slice, and the local maxima are marked as propagation path nodes.

[0021] A spatiotemporal coordinate index is established for each propagation path node, the spatiotemporal coordinate index containing the spatial location coordinates of the propagation path node and the time of its appearance in the spatial distribution evolution sequence;

[0022] For each candidate operation location, the Euclidean distance between the candidate operation location and the spatial coordinates of each propagation path node is calculated. The Euclidean distance is then converted into the operation response delay of the air cannon operation device to obtain the operation time deviation of the candidate operation location to each propagation path node. A timeliness matching relationship is then constructed based on the operation time deviation and the occurrence time of the propagation path node.

[0023] Spatial coverage assessment of timeliness matching relationship is conducted, and priority operation locations are determined based on the assessment results.

[0024] In one optional embodiment, spatial coverage assessment of timeliness matching relationships is performed, and priority operation locations are determined based on the assessment results, including:

[0025] For each candidate job location, extract the propagation path nodes in its timeliness matching relationship where the job time deviation is less than a preset timeliness threshold, and form the propagation path nodes into a set of coverage nodes for that candidate job location.

[0026] Calculate the spatial span of the covered node set in the cloud propagation path, where the spatial span is the projection of the distance between the foremost and last nodes in the covered node set along the propagation direction;

[0027] The path coverage of the candidate job location is calculated based on the spatial span and the total length of the cloud propagation path.

[0028] The density coverage of the candidate job location is obtained by summing the weighted average cloud density of each propagation path node in the coverage node set.

[0029] The path coverage and density coverage are weighted and fused to obtain the spatial blocking effectiveness index of the candidate operation location. The candidate operation locations are sorted according to the spatial blocking effectiveness index, and the candidate operation location with the highest spatial blocking effectiveness index is selected as the priority operation location.

[0030] In one optional embodiment, a temporal constraint graph of the job positions is constructed based on the spatial connectivity and temporal relationship of the priority job positions in the cloud propagation path. A job sequence is generated based on the temporal constraint graph. The perturbation effect of the preceding job positions in the job sequence on the spatiotemporal evolution field of the cloud is calculated. The timeliness matching relationship of subsequent job positions is recalculated based on the updated spatiotemporal evolution field of the cloud, including:

[0031] The adjacency relationship of the priority work positions in the cloud propagation path is calculated to determine the spatial connectivity between each priority work position. Based on the temporal relationship of the nodes in the propagation path, the order of each priority work position in the cloud propagation process is determined, and directed connection edges are established between the priority work positions, with the direction of the directed connection edges pointing to the downstream work positions in the propagation path.

[0032] Using the priority job positions as graph nodes and the directed connecting edges as graph edges, a temporal constraint graph of job positions is constructed. The directed connecting edges are assigned time constraint weights, which correspond to the time interval between the completion of a job at a previous job position and the commencement of a job at a subsequent job position.

[0033] Perform topological sorting on the time-series constraint graph and generate a job sequence based on the sorting results;

[0034] For the preceding operation position in the operation sequence, the disturbance effect of the simulated air gun operation device on the spatiotemporal evolution field of the cloud layer during the operation at the preceding operation position is calculated. The changes in cloud density distribution and the deflection of propagation direction caused by the disturbance effect are calculated, and the spatiotemporal evolution field of the cloud layer is updated. Based on the updated spatiotemporal evolution field of the cloud layer, the timeliness matching relationship of the subsequent operation positions is recalculated.

[0035] In one optional embodiment, the simulated air cannon operating device performs operations at the preceding operating position, causing disturbance effects on the spatiotemporal evolution field of clouds. The calculation of the cloud density distribution changes and propagation direction deflection caused by the disturbance effects includes:

[0036] Using the location of the preceding operation as the disturbance source point, calculate the arrival time of the wavefront at different spatial locations as the disturbance wave propagates radially outward in the spatiotemporal evolution field of the cloud.

[0037] Based on the operational energy release characteristics of the air cannon operating device, the initial disturbance intensity at the disturbance source point is calculated, and the initial disturbance intensity is spatially attenuated and corrected according to the propagation distance of the disturbance wave and the arrival time of the wavefront to obtain the disturbance intensity distribution at each spatial location.

[0038] The disturbance intensity at each spatial location is applied to the cloud density value at the corresponding spatial location in the spatiotemporal evolution field of the cloud, the reduction of cloud density by the disturbance intensity is calculated, and the cloud density distribution is updated according to the reduction.

[0039] Based on the vector angle between the propagation direction of the disturbance wave and the original cloud propagation direction, the deflection moment generated by the disturbance wave on the cloud propagation direction is calculated, and the cloud propagation direction is updated based on the deflection moment.

[0040] In one optional embodiment, each operation location in the operation sequence is taken as a path node. A spatiotemporal constraint network is constructed based on the matching relationship between movement time cost and timeliness. Dynamic programming search is performed on the spatiotemporal constraint network to obtain the operation path of the air cannon operation device passing through each path node in sequence, including:

[0041] The three-dimensional spatial coordinates of each operation position in the operation sequence are extracted, and the operation position is used as a path node, with the current position of the air cannon operation device as the starting path node;

[0042] Calculate the three-dimensional spatial distance between each path node, and based on the movement speed characteristics of the air cannon operating device under different terrain conditions, convert the three-dimensional spatial distance into the movement time cost between the path nodes;

[0043] The operation time window corresponding to each path node is extracted from the timeliness matching relationship. The operation time window includes the operation start time and operation end time of the path node. The movement time cost is coupled with the operation time window in a time sequence to construct a spatiotemporal constraint network. The connection edges between each path node in the spatiotemporal constraint network are set with time constraint weights. The time constraint weights record the time boundary conditions corresponding to the arrival of the subsequent path node.

[0044] Dynamic programming search is performed on the spatiotemporal constrained network. Starting from the initial path node, the cumulative time cost to reach each path node is calculated under the constraint of time constraint weights. The path node sequence with the minimum cumulative time cost is selected to generate the operation path of the air cannon operation device passing through each path node in sequence.

[0045] A second aspect of the present invention provides an air gun operation path planning system based on cloud maps, comprising:

[0046] The cloud map acquisition unit is used to acquire cloud map data of the target area;

[0047] Evolutionary building units are used to perform spatiotemporal coupling analysis on cloud image data, extract dynamic propagation characteristic parameters of cloud layers, and construct a spatiotemporal evolution field of cloud layers through cloud centroid trajectory tracking and boundary diffusion situation inference, thereby obtaining the spatial distribution evolution sequence of cloud layers in future time periods.

[0048] The node identification unit is used to identify cloud propagation path nodes based on the spatial distribution evolution sequence, calculate the time-matching relationship between candidate operation positions and propagation path nodes by combining the operation response delay of the air cannon operation device, and identify priority operation positions that have a spatial blocking effect on cloud propagation based on the time-matching relationship.

[0049] The temporal constraint unit is used to construct a temporal constraint diagram of the operation positions based on the spatial connectivity and temporal relationship of the priority operation positions in the cloud propagation path, generate an operation sequence based on the temporal constraint diagram, calculate the perturbation effect of the preceding operation positions in the operation sequence on the spatiotemporal evolution field of the cloud, and recalculate the timing matching relationship of the subsequent operation positions based on the updated spatiotemporal evolution field of the cloud.

[0050] The path planning unit is used to take each operation position in the operation sequence as a path node, construct a spatiotemporal constraint network based on the matching relationship between movement time cost and timeliness, and perform dynamic planning search on the spatiotemporal constraint network to obtain the operation path of the air cannon operation device passing through each path node in sequence.

[0051] The operation execution unit is used to control the air cannon operating device to perform operations along the operating path.

[0052] A third aspect of the present invention provides an electronic device, comprising:

[0053] processor;

[0054] Memory used to store processor-executable instructions;

[0055] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0056] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0057] In this embodiment of the invention, by performing spatiotemporal coupling analysis on cloud image data, the dynamic propagation characteristics of clouds can be extracted, and their spatiotemporal evolution field can be constructed to accurately predict the spatial distribution evolution of clouds in future time periods. By calculating the timeliness matching relationship between the air cannon operation response delay and the cloud propagation path nodes, it is ensured that the selected operation location is effective within the time window, enabling intervention to be completed before the clouds reach critical nodes, thus improving the timeliness and targeting of the operation. By dynamically updating the cloud state and re-evaluating the matching relationship of subsequent operation points, adaptive adjustment of the operation sequence is achieved. This closed-loop feedback mechanism enhances the robustness of the planning system, enabling it to cope with changes in the cloud field environment due to human intervention and avoiding failure problems that may be caused by static planning. By constructing a spatiotemporal constraint network that includes movement time costs and timeliness matching, and using dynamic programming for search, an operation path that is optimized in both time and space can be generated. This path guides the air cannon operation device to pass through each operation point in the most efficient order, minimizing movement time while meeting strict timing requirements, thereby significantly improving overall operation efficiency and resource utilization, and realizing systematic and efficient intervention of clouds in the target area. Attached Figure Description

[0058] Figure 1 This is a flowchart illustrating the air cannon operation path planning method based on cloud maps.

[0059] Figure 2 This is a flowchart illustrating the evolution logic of cloud disturbances. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0062] Figure 1 This is a flowchart illustrating the air cannon operation path planning method based on cloud maps according to an embodiment of the present invention, as shown below. Figure 1 As shown, the air gun operation path planning method based on cloud maps includes:

[0063] Obtain cloud map data for the target area;

[0064] Spatiotemporal coupling analysis of cloud image data is performed to extract dynamic propagation characteristic parameters of cloud layers. By tracking the cloud centroid trajectory and extrapolating the boundary diffusion situation, a spatiotemporal evolution field of cloud layers is constructed to obtain the spatial distribution evolution sequence of cloud layers in future time periods.

[0065] Based on the spatial distribution evolution sequence, cloud propagation path nodes are identified. The time-sensitivity matching relationship between candidate operation locations and propagation path nodes is calculated by combining the operation response delay of the air cannon operation device. Based on the time-sensitivity matching relationship, priority operation locations that have a spatial blocking effect on cloud propagation are identified.

[0066] A temporal constraint diagram of the operation positions is constructed based on the spatial connectivity and temporal relationship of the priority operation positions in the cloud propagation path. An operation sequence is generated based on the temporal constraint diagram. The perturbation effect of the preceding operation positions in the operation sequence on the spatiotemporal evolution field of the cloud is calculated. The timing matching relationship of the subsequent operation positions is recalculated based on the updated spatiotemporal evolution field of the cloud.

[0067] Each work location in the work sequence is taken as a path node. A spatiotemporal constraint network is constructed based on the matching relationship between movement time cost and timeliness. Dynamic programming search is performed on the spatiotemporal constraint network to obtain the work path of the air cannon working device passing through each path node in sequence.

[0068] Control the air cannon operating device to perform operations along the operating path.

[0069] In one optional embodiment, spatiotemporal coupling analysis is performed on cloud image data to extract dynamic propagation characteristic parameters of the cloud layer. By tracking the cloud centroid trajectory and extrapolating the boundary diffusion situation, a spatiotemporal evolution field of the cloud layer is constructed, resulting in a spatial distribution evolution sequence of the cloud layer in the future, including:

[0070] Spatial gradient field calculation is performed on the cloud image data to extract the spatial gradient vector field of cloud density distribution. Then, the spatial gradient vector field is decomposed into divergence and curl to obtain the divergence propagation component and rotation propagation component of the cloud layer.

[0071] The divergent propagation component and the rotational propagation component are vector-synthesized to obtain the dynamic propagation characteristics of the cloud layer.

[0072] Based on the aforementioned dynamic propagation characteristics, the centroid coordinates of the cloud layer are calculated. Time series analysis is then performed on the centroid coordinates, and the velocity vector and acceleration vector of the centroid displacement are fitted to obtain the cloud layer centroid trajectory.

[0073] Based on the dynamic propagation characteristics, the normal diffusion rate at the cloud boundary point is determined, and the cloud boundary is extrapolated along the cloud centroid trajectory with a time step to obtain the diffusion pattern of the cloud boundary;

[0074] The cloud centroid trajectory and the diffusion pattern of the cloud boundary are mapped in spatiotemporal coordinates to construct the spatiotemporal evolution field of the cloud. The spatiotemporal evolution field of the cloud is then extrapolated along the time axis to obtain the spatial distribution evolution sequence.

[0075] In one specific implementation, after acquiring cloud image data of the target area, in-depth spatiotemporal coupling analysis is performed to reveal the evolutionary patterns of the clouds. Cloud image data typically contains spatial distribution information of clouds at a specific moment, and its data format is a rasterized two-dimensional matrix or three-dimensional voxel data. Each data point records physical quantities such as cloud density, reflectivity, or optical thickness at the corresponding spatial location. To extract the dynamic propagation characteristic parameters of the clouds, spatial gradient field calculations are performed on the cloud image data. Specifically, numerical differentiation methods are used to calculate the rate of change of cloud density distribution in each spatial direction. In two-dimensional cloud images, partial derivatives in the horizontal and vertical directions are calculated separately, while in three-dimensional cloud images, partial derivatives in three orthogonal directions need to be calculated. The gradient vector formed by these partial derivatives can characterize the spatial trend of cloud density change. The direction of the gradient vector points to the direction of the fastest increase in cloud density, and its magnitude reflects the drastic degree of density change.

[0076] After extracting the spatial gradient vector field, divergence and curl decomposition are performed to separate propagation components with different properties. Divergence calculation involves summing the partial derivatives of the gradient vector field components in each spatial dimension. A positive divergence value indicates a divergent trend of clouds at that location, meaning the clouds are spreading outwards from that point; a negative divergence value indicates a convergent trend, meaning the clouds are converging towards that point. Curl calculation is obtained by a specific combination of the partial derivatives of the gradient vector field. In two dimensions, curl is a scalar; in three dimensions, it is a vector, reflecting the rotational motion characteristics of the clouds. The field distribution obtained through divergence calculation represents the divergent propagation components of the clouds, reflecting the radial diffusion or contraction characteristics of the clouds due to internal dynamic processes. The rotational propagation components obtained from curl decomposition reveal the vortex motion or tangential flow characteristics of the clouds. These rotational characteristics are usually related to atmospheric circulation, topographic effects, or local airflow disturbances.

[0077] When vector synthesizing the divergent and rotational propagation components, the two components are superimposed at each spatial location according to the vector addition rule. The divergent propagation component is represented as a radially outward vector field, oriented along the density gradient direction, and its magnitude is related to the divergence value; the rotational propagation component is represented as a tangential vector field, oriented perpendicular to the gradient direction, and its magnitude is related to the curl value. The combined vector field obtained after their synthesis represents the dynamic propagation characteristics of the cloud layer. This characteristic field comprehensively describes the cloud's motion trend at various spatial locations, including both radial motion (diffusion and contraction) and tangential motion (rotation and translation). This decomposition and synthesis method can extract quantifiable dynamic parameters from the complex cloud evolution process, providing a physical basis for subsequent cloud evolution prediction.

[0078] Based on the extracted dynamic propagation characteristics, the centroid coordinates of the cloud layer are calculated. The density of each spatial location within the cloud layer is used as a weight to perform a weighted average of the spatial coordinates. In a two-dimensional cloud image, the x-coordinate of the centroid is equal to the sum of the products of the x-coordinates of all pixels and their corresponding densities, divided by the total density; the y-coordinate is calculated in the same way. In a three-dimensional cloud image, all three coordinate components of the centroid are obtained through a similar weighted average. A time series of the centroid coordinates can be obtained from cloud image data over multiple consecutive time periods. When analyzing this time series, the difference between the centroid positions at adjacent time points is calculated, and divided by the time interval to obtain the velocity vector of the centroid displacement. The velocity vector reflects the overall direction and speed of cloud movement; its direction indicates the main propagation direction, and its magnitude characterizes the cloud's speed. Differential calculation of the velocity vector time series yields the acceleration vector, which reveals the changing trend of cloud speed; positive acceleration indicates accelerated cloud propagation, while negative acceleration indicates deceleration. By fitting the velocity vector and acceleration vector of the center of mass displacement, a dynamic model of the cloud center of mass motion can be established. This model can predict the position of the cloud center of mass at future moments, thus obtaining the trajectory of the cloud center of mass.

[0079] While obtaining the cloud centroid trajectory, the diffusion pattern of the cloud boundary is analyzed. Cloud boundary identification is achieved by setting a density threshold; areas with density above the threshold are defined as the cloud interior, and areas with density below the threshold are defined as the cloud exterior; the boundary between these two is the cloud boundary. Cloud boundaries are typically irregular curves or surfaces; the spatial coordinates of each point on the boundary are extracted. Based on the dynamic propagation characteristics, the normal diffusion rate of the cloud boundary points is determined. The normal direction of a boundary point is defined as the outward normal direction of the cloud boundary at that point, i.e., perpendicular to the boundary tangent and pointing outward. The projection of the dynamic propagation characteristic vector at that point onto the normal direction is the normal diffusion rate; a positive value indicates outward expansion of the boundary, and a negative value indicates inward contraction. When extrapolating the cloud boundary along the cloud centroid trajectory over time steps, each boundary point is moved one time step along its normal direction according to the corresponding diffusion rate, thus obtaining the boundary position at the next moment. By continuously performing simulations over multiple time steps, the morphological changes of the cloud boundary at various future moments are obtained. These changes reflect the cloud diffusion pattern, including the expansion, contraction, or deformation of the cloud coverage area.

[0080] By mapping the cloud centroid trajectory to the cloud boundary diffusion pattern using spatiotemporal coordinates, both the trajectory of the centroid position and the evolution of the boundary morphology over time are simultaneously expressed in a unified spatiotemporal coordinate system. This mapping constructs a cloud spatiotemporal evolution field, a four-dimensional structure containing three-dimensional spatial coordinates and one-dimensional time coordinates. Each spatiotemporal point in the field corresponds to the cloud state at a specific time and spatial location. The cloud spatiotemporal evolution field not only contains spatial distribution information of the clouds but also integrates their motion trends and evolutionary patterns, comprehensively depicting the evolution of clouds from their current state to future states. When extrapolating the cloud spatiotemporal evolution field along the time axis, based on the established centroid motion dynamics model and boundary diffusion rate model, the current cloud state is used as the initial condition, and the cloud state at each future time step is iteratively calculated according to a certain time step. The calculation at each time step includes updating the centroid position, updating the boundary morphology, and updating the internal density distribution. Through this iterative extrapolation, the spatial distribution prediction results of clouds at various times in the future are obtained. This series of spatial distribution states arranged in chronological order constitutes the spatial distribution evolution sequence.

[0081] The spatial distribution evolution sequence presents the continuous evolution of clouds in the form of discrete time snapshots, with each snapshot corresponding to the spatial distribution state of clouds at a specific moment. The time interval between adjacent snapshots in the sequence depends on the predicted time step; a smaller time step provides a more detailed description of the evolution process, but also increases the computational load. This sequence provides crucial predictive information for subsequent operational path planning, enabling operational decisions to be based on the future state of the clouds rather than solely on the current state, thus achieving proactive and forward-looking intervention. Through the aforementioned spatiotemporal coupling analysis process, static cloud image observation data is transformed into a dynamic cloud evolution prediction model, providing a scientific basis for the timing and location deployment of air gun operations. This physics-based cloud evolution modeling method fully utilizes the dynamic laws of cloud propagation, exhibiting stronger reliability and adaptability compared to simple statistical extrapolation or empirical prediction, and can adapt to the complexity and diversity of cloud evolution under different meteorological conditions and terrain environments.

[0082] In one optional embodiment, cloud propagation path nodes are identified based on the spatial distribution evolution sequence. The timeliness matching relationship between candidate operational locations and propagation path nodes is calculated by combining the operational response delay of the air cannon operating device. Based on the timeliness matching relationship, priority operational locations that have a spatial blocking effect on cloud propagation are identified, including:

[0083] Feature direction extraction is performed on the spatial distribution evolution sequence to obtain the direction vector of cloud propagation. The spatiotemporal evolution field of the cloud is divided into equally spaced spatial slices along the direction vector. Local maxima of cloud density are identified in each spatial slice, and the local maxima are marked as propagation path nodes.

[0084] A spatiotemporal coordinate index is established for each propagation path node, the spatiotemporal coordinate index containing the spatial location coordinates of the propagation path node and the time of its appearance in the spatial distribution evolution sequence;

[0085] For each candidate operation location, the Euclidean distance between the candidate operation location and the spatial coordinates of each propagation path node is calculated. The Euclidean distance is then converted into the operation response delay of the air cannon operation device to obtain the operation time deviation of the candidate operation location to each propagation path node. A timeliness matching relationship is then constructed based on the operation time deviation and the occurrence time of the propagation path node.

[0086] Spatial coverage assessment of timeliness matching relationship is conducted, and priority operation locations are determined based on the assessment results.

[0087] In one specific implementation, when identifying cloud propagation path nodes, deep feature extraction is performed on the obtained spatial distribution evolution sequence. By calculating the displacement vector of the cloud centroid position at adjacent time points and combining it with the maximum gradient direction of cloud boundary diffusion, a weighted average method is used to fuse these two types of directional information to obtain a direction vector that can characterize the main propagation trend of the cloud. In practice, the cloud centroid positions at 5 to 10 consecutive time points in the spatial distribution evolution sequence are linearly fitted, and the direction of the fitted line is the main propagation direction. This direction vector is usually expressed in the form of azimuth and elevation angles, with the azimuth angle ranging from 0° to 360° and the elevation angle ranging from -90° to 90°.

[0088] After obtaining the direction vector, the spatiotemporal evolution field of the cloud is segmented into equally spaced spatial slices along this direction vector. Specifically, a series of parallel slices are established with the direction vector as the normal vector. The spacing between the slices is determined according to the spatial scale of the target area and the cloud thickness, typically set to 300 to 800 meters. Within each spatial slice, the three-dimensional cloud density distribution data of the region where the slice intersects with the spatiotemporal evolution field of the cloud is extracted. Local extremum detection is performed on the density data to identify the local maxima of cloud density. The local extremum detection uses a sliding window method, with the window size set to a three-dimensional spatial range of 150 meters × 150 meters × 100 meters. Density peak points are searched within the window. When the density value of a point is greater than the density values ​​of all surrounding points and exceeds a set threshold, the point is marked as a propagation path node. This threshold is typically set to 1.5 to 2.0 times the average cloud density to ensure that the identified nodes truly correspond to key locations in cloud propagation.

[0089] When establishing a spatiotemporal coordinate index for each propagation path node, a four-dimensional coordinate system is used for description. The spatial location coordinates employ a three-dimensional Cartesian coordinate system with the target region center as the origin, recording the node's longitude, latitude, and altitude information, with coordinate accuracy maintained at the meter level. The occurrence time records the first time the propagation path node is identified in the spatial distribution evolution sequence, with time accuracy down to the second level. The spatiotemporal coordinate index is stored in the form of a data structure, with each node containing attribute information such as node number, three-dimensional spatial coordinates, occurrence time, cloud density value, and the slice number to which it belongs. This index structure facilitates subsequent rapid retrieval and matching calculations.

[0090] When calculating the timeliness matching relationship between candidate operation locations and propagation path nodes, the spatial distribution of candidate operation locations is determined. Candidate operation locations are typically a pre-defined set of locations based on factors such as terrain, accessibility, and deployment conditions of air cannons. These locations cover key areas within the target region that interfere with cloud propagation. For each candidate operation location, the Euclidean distance between its spatial coordinates and those of all propagation path nodes is calculated using the following formula: ,in , , These represent the coordinate differences between the candidate job location and the propagation path node in the three coordinate axes.

[0091] When performing spatiotemporal transformation on the calculated Euclidean distance, the operational response delay of the air cannon is considered. The operational response delay includes the time from receiving the command to completing operational preparation, the operational execution time, and the physical propagation time of the operational effect from the launch position to the target position. Assuming the speed of sound in the atmosphere is approximately 340 meters per second, the physical propagation time from the candidate operational position to the propagation path node is... The preparation and execution times of the device are determined according to the equipment model, and are typically between 30 and 120 seconds, denoted as... The calculation of the operation time deviation considers the difference between the time when the shock wave arrives at the propagation path node and the actual time when the node appears, if the operation is carried out at the candidate operation location. ,in This refers to the moment when a node appears in the propagation path. When the value is close to zero, it indicates that performing the operation at the candidate job location can accurately hit the propagation path node and achieve the best intervention effect.

[0092] A timeliness matching relationship is constructed based on the job time deviation and the occurrence time of propagation path nodes. This relationship quantifies the degree to which candidate job locations grasp the timing of operations at each propagation path node, and is characterized by a matching degree function. When the absolute value of the job time deviation is less than a set time window threshold, the candidate job location is considered to have a good timeliness matching relationship with the propagation path node. The time window threshold is typically set between 60 and 180 seconds, with the specific value adjusted according to the cloud propagation speed and job accuracy requirements. The matching degree function can employ a Gaussian decay function, ensuring that the smaller the time deviation, the higher the matching degree; when the time deviation exceeds the threshold, the matching degree rapidly decreases to near zero.

[0093] In one optional embodiment, spatial coverage assessment of timeliness matching relationships is performed, and priority operation locations are determined based on the assessment results, including:

[0094] For each candidate job location, extract the propagation path nodes in its timeliness matching relationship where the job time deviation is less than a preset timeliness threshold, and form the propagation path nodes into a set of coverage nodes for that candidate job location.

[0095] Calculate the spatial span of the covered node set in the cloud propagation path, where the spatial span is the projection of the distance between the foremost and last nodes in the covered node set along the propagation direction;

[0096] The path coverage of the candidate job location is calculated based on the spatial span and the total length of the cloud propagation path.

[0097] The density coverage of the candidate job location is obtained by summing the weighted average cloud density of each propagation path node in the coverage node set.

[0098] The path coverage and density coverage are weighted and fused to obtain the spatial blocking effectiveness index of the candidate operation location. The candidate operation locations are sorted according to the spatial blocking effectiveness index, and the candidate operation location with the highest spatial blocking effectiveness index is selected as the priority operation location.

[0099] In one specific implementation, when determining priority operation locations, the spatial blocking effectiveness of candidate operation locations is quantitatively evaluated. For each candidate operation location, propagation path nodes that meet the timeliness conditions are extracted from its timeliness matching data. The timeliness matching relationship records the time deviation of the candidate operation location from each propagation path node. When the operation time deviation is less than a preset timeliness threshold, it indicates that the node is within the effective operation window of the candidate operation location. The preset timeliness threshold is determined based on the cloud propagation speed and the response characteristics of the air cannon operation device, and is typically set to 3 to 8 minutes. The propagation path nodes that meet the timeliness conditions are grouped into a coverage node set for the candidate operation location, which represents the cloud propagation area that the location can effectively intervene in.

[0100] Spatial span parameters are calculated based on the set of covered nodes. The foremost and last nodes of the covered node set are identified along the cloud propagation path. A projected coordinate system is established along the cloud propagation direction, and the distance projection between the two endpoint nodes is calculated. When the cloud propagation direction is tortuous, the cumulative arc length of the propagation path is used as the metric for the projected distance. Spatial span reflects the extension length of the operational influence range of the candidate operation location along the propagation direction; a larger value indicates that a single operation can cover a longer propagation path segment.

[0101] Path coverage is calculated as the ratio of spatial span to the total length of the cloud propagation path. The total length of the cloud propagation path is obtained by measuring the propagation trajectory of the cloud's spatiotemporal evolution field, extending from the current cloud position along the propagation direction to the end of the prediction period, and the total length is obtained by summing the distances between each propagation path node. The formula for calculating path coverage is as follows: ,in For spatial span, This represents the total length of the cloud propagation path. The path coverage value ranges from 0 to 1; the closer the value is to 1, the greater the proportion of the propagation path the candidate job location can cover.

[0102] Cloud density data for each propagation path node in the coverage node set is extracted. Cloud density is obtained through cloud image data analysis and reflects the mass distribution intensity of the cloud at that node location, expressed in grams per cubic meter. For each propagation path node in the coverage node set, its corresponding cloud density value is obtained, and a weighting coefficient is assigned based on the node's relative position in the propagation path. Nodes closer to the cloud propagation front are assigned higher weights because operations at these locations can block cloud propagation earlier. The weighting coefficients are linearly distributed according to the node's normalized position in the propagation path, with front-end nodes having a weighting coefficient close to 1 and end-end nodes having a weighting coefficient close to 0.3. The density coverage of the candidate operation location is obtained by summing the products of the cloud density and their weighting coefficients for each propagation path node. This parameter comprehensively reflects the intensity of cloud quality distribution within the coverage area.

[0103] The spatial blocking effectiveness index is calculated by weighting and fusing path coverage and density coverage. The fusion formula uses a linear weighting form. ,in and To integrate the weighting coefficients, The value ranges from 0.4 to 0.6. The value range is from 0.4 to 0.6, and satisfies... In practical applications, when focusing on the breadth of the task coverage, increasing... The value; when focusing on the reduction effect of cloud density, increase The value. In scenarios where cloud density is relatively uniform, it is usually set to... It is 0.55. The value is 0.45. In scenarios where cloud density distribution varies significantly, the value is set to... It is 0.45. The value is set to 0.55 to ensure priority processing of high-density cloud areas.

[0104] After calculating the spatial blocking effectiveness index for each candidate operational location, they are sorted in descending order of their numerical values. A higher spatial blocking effectiveness index indicates better spatial coverage and density reduction capabilities in blocking cloud propagation. Based on actual operational needs and the limited number of air cannon operational devices, the top few candidate operational locations are selected as priority operational locations from the sorted results. When there is only one operational device, the candidate operational location with the highest spatial blocking effectiveness index is selected; when there are multiple operational devices, the corresponding number of candidate operational locations are selected sequentially according to the sorting order. The spatial independence between candidate operational locations must also be considered during the selection process to avoid over-concentration of priority operational locations leading to operational redundancy. Generally, the straight-line distance between adjacent priority operational locations is required to be greater than the product of cloud propagation speed and operational response time.

[0105] In actual calculations, for complex cloud formations with multiple branch propagation paths, the set of covered nodes and related parameters are calculated separately for each branch path. For a candidate job location, if its set of covered nodes spans multiple branch paths, the spatial span and density coverage of each branch path are calculated separately. These are then weighted and summed according to the cloud quality proportion of each branch path to obtain the comprehensive spatial span and comprehensive density coverage of the candidate job location. The cloud quality proportion is calculated by spatially integrating the cloud density along each branch path; branch paths with higher cloud quality have a higher weight in the comprehensive calculation.

[0106] When cloud propagation paths change dynamically, the extraction of the coverage node set is based on real-time updated cloud spatiotemporal evolution data. As time progresses, the spatial location and cloud density of cloud propagation path nodes change. Therefore, when calculating the timeliness matching relationship, the arrival time of candidate operation locations is matched with the corresponding state of the propagation path nodes. Spatiotemporal interpolation methods are used to obtain the cloud density values ​​of propagation path nodes at any given time, ensuring the spatiotemporal consistency of density coverage calculation. For clouds with faster propagation speeds, the timeliness threshold is set relatively low to ensure timely operation intervention; for clouds with slower propagation speeds, the timeliness threshold is appropriately relaxed to allow for greater flexibility in operation time.

[0107] The calculation results of the space blocking effectiveness index are also affected by cloud thickness and vertical structure. When cloud thickness is high, the penetration capability of a single operation is limited. Therefore, the weighting coefficient of density coverage is adjusted by incorporating vertical cloud layering information. Through vertical profile analysis of cloud image data, the density distribution of clouds at different altitudes is obtained. Weighted calculations are then performed on cloud density within the effective operating height range of the operational equipment to improve the accuracy of space blocking effectiveness assessment.

[0108] In one optional embodiment, a temporal constraint graph of the job positions is constructed based on the spatial connectivity and temporal relationship of the priority job positions in the cloud propagation path. A job sequence is generated based on the temporal constraint graph. The perturbation effect of the preceding job positions in the job sequence on the spatiotemporal evolution field of the cloud is calculated. The timeliness matching relationship of subsequent job positions is recalculated based on the updated spatiotemporal evolution field of the cloud, including:

[0109] The adjacency relationship of the priority work positions in the cloud propagation path is calculated to determine the spatial connectivity between each priority work position. Based on the temporal relationship of the nodes in the propagation path, the order of each priority work position in the cloud propagation process is determined, and directed connection edges are established between the priority work positions, with the direction of the directed connection edges pointing to the downstream work positions in the propagation path.

[0110] Using the priority job positions as graph nodes and the directed connecting edges as graph edges, a temporal constraint graph of job positions is constructed. The directed connecting edges are assigned time constraint weights, which correspond to the time interval between the completion of a job at a previous job position and the commencement of a job at a subsequent job position.

[0111] Perform topological sorting on the time-series constraint graph and generate a job sequence based on the sorting results;

[0112] For the preceding operation position in the operation sequence, the disturbance effect of the simulated air gun operation device on the spatiotemporal evolution field of the cloud layer during the operation at the preceding operation position is calculated. The changes in cloud density distribution and the deflection of propagation direction caused by the disturbance effect are calculated, and the spatiotemporal evolution field of the cloud layer is updated. Based on the updated spatiotemporal evolution field of the cloud layer, the timeliness matching relationship of the subsequent operation positions is recalculated.

[0113] In one specific implementation, after obtaining multiple priority job locations, these discrete spatial points are organized into a job sequence with logical relationships. Adjacency relationships are calculated for the spatial distribution of each priority job location along the cloud propagation path. By constructing a spatial adjacency matrix, for any two priority job locations... and Calculate the spatial distance between the two. .when Less than the preset adjacency threshold When spatial connectivity exists between two locations, it is determined. This adjacency threshold is determined based on the width of the cloud propagation path and the effective operating radius of the air cannon, and is typically set to 0.5 to 1.5 times the width of the cloud propagation path. By determining spatial connectivity, priority operating location groups that are adjacent in the cloud propagation path and can form a synergistic blocking effect can be identified.

[0114] Based on the established spatial connectivity, the order of priority operation locations during cloud propagation is determined according to the temporal relationships of the propagation path nodes. For each priority operation location, the timestamp information of its corresponding propagation path nodes is extracted. If the priority operation location... Corresponding propagation path node timestamp Earlier than the priority operation position Corresponding timestamp Then determine During cloud propagation, located in The upstream position is determined by comparing the temporal relationships of all priority work positions. For pairs of priority work positions that are spatially connected and have a sequential relationship, a directed connection edge is established between them. The starting point of this directed connection edge is the work position upstream of the propagation path, and the ending point is the work position downstream, with the edge pointing in the direction of cloud propagation. In this way, the dispersed priority work positions are connected by directed edges, forming a directional positional relationship network.

[0115] Using priority operation locations as graph nodes and directed edges as graph edges, a temporal constraint graph of operation locations is constructed. This graph is a directed acyclic graph (DAG), where each node represents a location requiring air cannon operation, and each directed edge represents the temporal dependency between two operation locations. For each directed edge in the graph, a time constraint weight is calculated and assigned. This time constraint weight represents the time interval required from the completion of the operation at the preceding location to the commencement of the operation at the subsequent location. The calculation of the time constraint weight comprehensively considers three factors: the natural propagation time of the cloud layer from the preceding location to the subsequent location, the duration of the operation effect at the preceding location, and the movement time of the air cannon device from the preceding location to the subsequent location. Specifically, the time constraint weight... Determined by: calculating cloud layers from location propagation to location transmission time This time is calculated based on the spatial distance between the two locations and the cloud propagation speed; the duration of the effect of the preceding operation at the location is estimated. This time is related to the operational intensity of the air cannon and the cloud density; finally, the movement time of the air cannon operating device is calculated. The time is determined based on the device's moving speed and the distance between its positions. The time constraint weight is the maximum value among these three time parameters to ensure that subsequent operations are carried out only after the effects of preceding operations have been fully realized and the cloud cover has reached its designated position.

[0116] After constructing the temporal constraint graph, a topological sort is performed on the directed acyclic graph. The topological sorting algorithm calculates the in-degree value of each node, prioritizes nodes with an in-degree of zero, adds them to the sorted sequence, deletes the node and all its outgoing edges, updates the in-degree values ​​of the remaining nodes, and repeats this process until all nodes are added to the sequence. The topological sorting result provides a reasonable temporal order for each operation location, ensuring that the execution of any operation location occurs only after all its preceding operations have been completed. A operation sequence is generated based on the topological sorting result, which clarifies the sequential order in which the air cannon operation device performs its operations. In cases where multiple nodes have an in-degree of zero, the location with the earlier timestamp of the corresponding propagation path node is prioritized to ensure that the operation sequence is consistent with the temporal order of cloud propagation.

[0117] After generating the operation sequence, the mutual influence between operation locations is considered. For the preceding operation location in the operation sequence, the disturbance effect of the air cannon operation on the spatiotemporal evolution field of the clouds at that location is simulated. The air cannon operation generates a shock wave by releasing high-pressure gas, which generates dynamic disturbances in the clouds. The disturbance effect is simulated by introducing a local dynamic disturbance source into the spatiotemporal evolution field of the clouds. A disturbance influence domain is established around the preceding operation location, the extent of which is determined based on the intensity of the air cannon operation and the propagation distance of the shock wave. Within the influence domain, the change in cloud density distribution caused by the disturbance is calculated. The air cannon operation reduces the cloud density around the operation location, and the magnitude of the density reduction is negatively correlated with the spatial distance from the operation location. The spatial distribution characteristics of the density change are described by a Gaussian decay function, with the density reduction being the most significant at the operation location, and the density change gradually decreasing with increasing distance.

[0118] Besides density changes, disturbance effects also cause deflection of cloud propagation direction. The shock wave generated by the air cannon applies directional momentum to the cloud, causing it to deflect from its original propagation direction. The deflection angle is calculated considering the angle between the shock wave's direction and the cloud's original propagation direction, as well as the ratio of the shock wave intensity to the cloud's propagation momentum. The deflection effect is most significant when the shock wave direction is perpendicular to the cloud's propagation direction; when they are parallel, it mainly affects the propagation speed rather than the direction. By using a vector superposition method, the additional momentum caused by the shock wave is combined with the original propagation momentum of the cloud to obtain the disturbed cloud propagation direction.

[0119] Based on the calculated changes in density distribution and the deflection of propagation direction, the spatiotemporal evolution field of the cloud layer is updated. The update process is achieved by modifying the density matrix and propagation vector field within the evolution field. Within the influence domain of the preceding operation location, the original density value is subtracted from the density reduction caused by the disturbance to obtain a new density distribution. For the propagation direction, at each spatial point within the influence domain, the deflection effect of the shock wave at that point is calculated based on the distance and direction from the operation location, updating the direction and magnitude of the propagation vector. The updated spatiotemporal evolution field of the cloud layer reflects the impact of the preceding operation on the subsequent evolution of the cloud layer.

[0120] Based on the updated cloud spatiotemporal evolution field, the generation process of the cloud spatial distribution evolution sequence is re-executed to obtain the cloud propagation situation considering the influence of preceding operations. Using the updated evolution sequence, the timeliness matching relationship of subsequent operation locations is recalculated. Since preceding operations change the cloud propagation path and arrival time, the previously calculated subsequent operation locations are no longer optimal. Through recalculation, the subsequent operation locations are adjusted and optimized. For each operation location in the operation sequence, after the completion of its preceding operations, a process of updating the cloud spatiotemporal evolution field and recalculating the subsequent operation location is performed, forming an iterative optimization mechanism. This mechanism ensures that the operation sequence can be dynamically adjusted according to the actual operation results, so that the selection of each operation location is based on the latest cloud propagation situation, improving the accuracy of the overall operation path planning and the operation effect. Through the construction of the temporal constraint diagram, the generation of the operation sequence, and the iterative calculation of the disturbance effect, collaborative optimization among multiple operation locations is achieved, ensuring that the operation path can exert the maximum blocking effect in the dynamic process of cloud propagation.

[0121] like Figure 2 The diagram shown illustrates the logical flowchart of cloud disturbance evolution.

[0122] In one optional embodiment, the simulated air cannon operating device performs operations at the preceding operating position, causing disturbance effects on the spatiotemporal evolution field of clouds. The calculation of the cloud density distribution changes and propagation direction deflection caused by the disturbance effects includes:

[0123] Using the location of the preceding operation as the disturbance source point, calculate the arrival time of the wavefront at different spatial locations as the disturbance wave propagates radially outward in the spatiotemporal evolution field of the cloud.

[0124] Based on the operational energy release characteristics of the air cannon operating device, the initial disturbance intensity at the disturbance source point is calculated, and the initial disturbance intensity is spatially attenuated and corrected according to the propagation distance of the disturbance wave and the arrival time of the wavefront to obtain the disturbance intensity distribution at each spatial location.

[0125] The disturbance intensity at each spatial location is applied to the cloud density value at the corresponding spatial location in the spatiotemporal evolution field of the cloud, the reduction of cloud density by the disturbance intensity is calculated, and the cloud density distribution is updated according to the reduction.

[0126] Based on the vector angle between the propagation direction of the disturbance wave and the original cloud propagation direction, the deflection moment generated by the disturbance wave on the cloud propagation direction is calculated, and the cloud propagation direction is updated based on the deflection moment.

[0127] In one specific implementation, after the preceding operation is performed at the location, the resulting disturbance effect is quantitatively analyzed to assess its impact on the spatiotemporal evolution field of clouds. When calculating the disturbance effect, the location of the preceding operation is considered the disturbance source point; the disturbance wave propagating outward from this point affects the cloud structure in the surrounding space. The propagation characteristics of the disturbance wave in the spatiotemporal evolution field of clouds are similar to the propagation process of mechanical waves in a medium, and its wavefront propagation velocity is affected by physical parameters such as cloud density, temperature, and humidity.

[0128] To address the propagation process of disturbance waves, a propagation model is established to calculate the arrival time of the wavefront at various spatial locations. In the implementation, the spatiotemporal evolution field of the cloud layer is discretized into a three-dimensional spatial grid, with each grid point corresponding to a spatial coordinate. Starting from the disturbance source point, the propagation speed of the disturbance wave varies along different radial directions. This difference is mainly caused by the density gradient and wind field distribution within the cloud layer. By calculating the local propagation speed of the disturbance wave at each grid point and combining this with the distance between that grid point and the disturbance source point, the cumulative time required for the disturbance wave to propagate from the source point to that location, i.e., the wavefront arrival time, can be obtained.

[0129] In the calculation process, ray tracing is used to simulate the propagation path of the disturbance wave. Starting from the disturbance source point, multiple rays are set in three-dimensional space, each representing a propagation direction. Along each ray, based on the physical parameters of each location in the spatiotemporal evolution field of the cloud, the propagation distance and propagation time of the disturbance wave are calculated step by step. When a ray passes through a grid point, the cumulative propagation time of that ray is assigned to that grid point as the time for the wavefront to arrive at that location. For cases where the same grid point is passed by multiple rays, the value with the shortest arrival time is selected as the final result. In this way, the wavefront arrival time distribution of all spatial locations in the entire spatiotemporal evolution field of the cloud is obtained.

[0130] After obtaining the wavefront arrival time distribution, the disturbance intensity at each spatial location is further calculated. The disturbance intensity reflects the ability of the disturbance wave to affect the cloud structure at that location, and its magnitude is closely related to the energy released by the air gun operating device and the propagation distance of the disturbance wave. At the disturbance source point, the initial disturbance intensity is calculated based on the technical parameters of the air gun operating device, including ammunition load, firing pressure, and spray angle. The initial disturbance intensity characterizes the energy density of the disturbance wave at the source point, and its magnitude directly affects the range of subsequent disturbance effects.

[0131] Disturbance waves experience energy attenuation during propagation, primarily caused by two mechanisms: geometric diffusion and medium absorption. Geometric diffusion attenuation occurs because the energy of the disturbance wave diffuses outward in three-dimensional space, leading to a decrease in energy density per unit area with increasing propagation distance. Medium absorption attenuation is caused by the energy loss of the disturbance wave due to components such as water vapor and aerosols in clouds. When calculating spatial attenuation corrections, the effects of both attenuation mechanisms are considered. For geometric diffusion attenuation, a spherical wave diffusion model is used, where the disturbance intensity is inversely proportional to the square of the propagation distance. For medium absorption attenuation, based on the density distribution and composition characteristics of clouds, an absorption coefficient parameter is introduced, and the disturbance intensity exhibits an exponential decrease with propagation distance.

[0132] After spatial attenuation correction of the initial disturbance intensity, the disturbance intensity distribution at each spatial location is obtained. This distribution exhibits a spatial pattern of gradually weakening outward from the disturbance source point. The disturbance intensity is relatively high in the vicinity of the source point, while it approaches zero in areas far from the source point. The spatial morphology of the disturbance intensity distribution is also affected by the wind field within the cloud layer. In the downwind direction, the disturbance wave propagates further and the disturbance intensity attenuates more slowly, while in the upwind direction, the disturbance intensity attenuates more quickly.

[0133] After obtaining the perturbation intensity distribution, it is applied to the cloud density value at the corresponding spatial location in the cloud spatiotemporal evolution field to calculate the impact of the perturbation on cloud density. The perturbation wave alters the cloud density distribution by generating pressure pulses and turbulence effects. In areas with high perturbation intensity, the pressure fluctuations generated by the perturbation wave cause the water vapor condensation nuclei in the cloud to be broken up, and the aggregation structure between cloud droplets to be disrupted, thereby reducing the cloud density at that location. The magnitude of the reduction is related to the perturbation intensity and the original cloud density at that location; the greater the perturbation intensity, the more significant the reduction effect on cloud density.

[0134] In the specific calculations, a response relationship between perturbation intensity and density reduction is established. This relationship considers the physical characteristics of the cloud layer, including cloud type, water content, and cloud droplet size distribution. The density reduction effect produced by the same perturbation intensity varies for different types of clouds. For example, for thick clouds such as cumulus, due to their relatively stable internal structure, a larger perturbation intensity is needed to produce a significant density reduction; while for thinner clouds such as stratus, a smaller perturbation intensity can produce a significant effect. By calculating the density reduction corresponding to the perturbation intensity at each spatial location and subtracting this reduction from the original cloud density value, the cloud density distribution is updated.

[0135] Besides affecting cloud density, disturbance waves also influence the propagation direction of clouds. Disturbance waves have a definite propagation direction, pointing from the disturbance source to various spatial locations. When a disturbance wave encounters a moving cloud, there is usually an angle between the wave's propagation direction and the cloud's original propagation direction. This vector angle reflects the degree of deviation between the two directions of motion. The disturbance wave alters the cloud's direction of motion by applying a lateral force, producing a deflection effect.

[0136] Vector mechanics is used to calculate the deflection moment. The propagation direction of the disturbance wave and the original propagation direction of the cloud are both represented as unit vectors. The direction and magnitude of the deflection moment are obtained by calculating the cross product of these two vectors. The magnitude of the deflection moment is proportional to the sine of the distance between the two direction vectors. The deflection moment reaches its maximum when the two directions are perpendicular, and is zero when they are parallel or antiparallel. The direction of the deflection moment is perpendicular to the plane formed by the propagation directions of the disturbance wave and the cloud, indicating the direction in which the cloud's propagation direction will be deflected.

[0137] The cloud propagation direction is updated based on the calculated deflection moment. The effect of the deflection moment is related to the cloud's mass and inertia; clouds with larger masses and faster speeds are less affected by the deflection moment, while clouds with smaller masses and slower speeds are more easily deflected. When updating the propagation direction, the deflection moment is applied as an external force to the cloud's equation of motion, and the new propagation direction is obtained by solving the equations. The updated propagation direction is characterized by a deflection of a certain angle from the original propagation direction towards the direction of the disturbance wave propagation. The magnitude of the deflection angle is determined by both the deflection moment and the cloud's dynamic parameters.

[0138] Through the above calculation process, a comprehensive quantitative assessment of the disturbance effect of the preceding operation locations was completed, obtaining updated results for cloud density distribution and propagation direction after the disturbance. These updated data will replace the original data for the corresponding regions in the cloud spatiotemporal evolution field, forming a new cloud spatiotemporal evolution field state. Based on the updated cloud spatiotemporal evolution field, the timing matching relationship of subsequent operation locations can be recalculated, ensuring that the overall optimization effect of the operation sequence conforms to the actual cloud evolution trend.

[0139] In one optional embodiment, each operation location in the operation sequence is taken as a path node. A spatiotemporal constraint network is constructed based on the matching relationship between movement time cost and timeliness. Dynamic programming search is performed on the spatiotemporal constraint network to obtain the operation path of the air cannon operation device passing through each path node in sequence, including:

[0140] The three-dimensional spatial coordinates of each operation position in the operation sequence are extracted, and the operation position is used as a path node, with the current position of the air cannon operation device as the starting path node;

[0141] Calculate the three-dimensional spatial distance between each path node, and based on the movement speed characteristics of the air cannon operating device under different terrain conditions, convert the three-dimensional spatial distance into the movement time cost between the path nodes;

[0142] The operation time window corresponding to each path node is extracted from the timeliness matching relationship. The operation time window includes the operation start time and operation end time of the path node. The movement time cost is coupled with the operation time window in a time sequence to construct a spatiotemporal constraint network. The connection edges between each path node in the spatiotemporal constraint network are set with time constraint weights. The time constraint weights record the time boundary conditions corresponding to the arrival of the subsequent path node.

[0143] Dynamic programming search is performed on the spatiotemporal constrained network. Starting from the initial path node, the cumulative time cost to reach each path node is calculated under the constraint of time constraint weights. The path node sequence with the minimum cumulative time cost is selected to generate the operation path of the air cannon operation device passing through each path node in sequence.

[0144] In one specific implementation, when converting a job sequence into an actual execution path, multiple factors such as spatial distance, time constraints, and equipment mobility are comprehensively considered. The three-dimensional spatial coordinates of each job position are extracted sequentially from the job sequence. ,in and Indicates the geographical coordinates of the work location in the horizontal plane. This indicates the required equipment working height for this work location. These work locations are defined as a set of path nodes. Simultaneously, obtain the current position coordinates of the air cannon operating device. Set it as the starting path node This process establishes a spatial mapping relationship between the current state of the equipment and the target task to be performed, providing basic node information for subsequent path planning.

[0145] For any two nodes in the path node set and Calculate the three-dimensional spatial distance between them. This distance includes not only horizontal planar displacement but also the vertical adjustment distance of the equipment. The straight-line distance between nodes is calculated using the Euclidean distance formula. However, in actual movement, it is often impossible to travel along a straight path. Therefore, distances need to be corrected based on terrain data of the target area. When the path traverses mountains, hills, or complex terrain, a terrain correction factor is introduced. Adjust the straight-line distance to obtain the actual distance traveled. The terrain correction factor is dynamically determined based on the terrain complexity of the area traversed by the route. In plains, the factor is close to 1, while in mountainous or obstacle-heavy areas, the factor reaches 1.5 to 2.0.

[0146] The movement speed of air cannon operating devices varies significantly under different terrain conditions. On flat roads, the equipment can move at a relatively fast speed, while in rugged terrain or areas requiring frequent adjustments to the equipment's attitude, the movement speed decreases significantly. A mapping relationship between terrain type and movement speed was established, classifying terrain into categories such as paved roads, dirt roads, grasslands, and slopes, with each terrain category corresponding to an average movement speed. For the node Move to node For each path segment, the applicable movement speed is determined based on the main terrain types traversed, and the movement time cost is calculated in conjunction with the actual movement distance. When a route traverses multiple terrain types, a weighted average is calculated based on the proportion of each terrain type to the total route length to obtain the overall travel speed before calculating the travel time cost. Furthermore, the equipment's docking and preparation time at nodes must also be considered. This time includes operations such as equipment positioning, attitude adjustment, and pre-operation system checks, which are added to the movement time cost to form the complete inter-node transfer time cost. .

[0147] Extract the job time window corresponding to each path node from the timeliness matching relationship. The job time window is determined by the job start time. With the deadline for the work In essence, the time window is defined as the physical requirement that equipment must arrive at the node and complete the operation within a specified timeframe to effectively intervene in cloud propagation. The start time of the operation is typically determined by the estimated time it will take for the cloud to reach that location, while the end time is determined by the point at which the intervention effect significantly diminishes after the cloud continues to propagate. The width of the time window reflects the flexibility of the operation timing; nodes with wider windows have greater time flexibility in path planning, while nodes with narrower windows have strict requirements on arrival time.

[0148] By temporally coupling the travel time cost with the operation time window, a spatiotemporal constraint network is constructed. ,in For a set of path nodes, The set of connecting edges between nodes. This is the set of time constraint weights for the edges. For nodes... and Connecting edges between Its time constraint weight Recorded from Move to Time cost and arrival The time boundary conditions that must be met subsequently. Specifically, if the device at time... Leave node Then the node is reached. The time is To ensure the task is effective and meets the time constraints, the following conditions must be met. When the expected arrival time is earlier than the start time of the operation, the equipment at the node... Waiting until The task can only begin after the expected arrival time is later than the task deadline; if the expected arrival time is later than the task deadline, the path node sequence is infeasible and is pruned during the search process. Time constraint weights encode these time relationships into the network structure, enabling subsequent search algorithms to automatically identify and avoid path combinations that violate time constraints.

[0149] The construction of spatiotemporally constrained networks also considers the reachability between nodes. Not all pairs of nodes are directly connected; only physically reachable and temporally feasible pairs are connected. This is determined through pre-calculation, and if a node... Start, even if you move to the node at the fastest speed. Its arrival time is still later than The deadline for the assignment is not... and Establish connecting edges between them. This preprocessing can significantly reduce the search space and improve the computational efficiency of path planning.

[0150] Dynamic programming search is performed based on a spatiotemporally constrained network. Starting from the initial path node... Departure, at the current moment As the initial time state, the dynamic programming process maintains a state table, recording the minimum cumulative time cost to reach each path node and its corresponding predecessor node. During initialization, the cumulative time cost of the starting node is set to zero, and the cumulative time cost of all other nodes is set to infinity. Nodes are expanded sequentially using either breadth-first search or a priority queue. For the currently visited node... and its arrival time ,traverse all from Reachable subsequent nodes Calculation via arrive The moment This expression reflects the time adjustment mechanism for waiting if arriving early. Further calculation of the cumulative time cost is then performed. ,in To reach the node The cumulative time cost includes the waiting time.

[0151] like and Arrival less than the current record The cumulative time cost then updates the node. The state is updated to its accumulated time cost. Arrival time updated to And record the predecessor node as This update process continues to propagate until the states of all reachable nodes converge. Dynamic programming search guarantees finding the path with the minimum cumulative time cost under time constraints. This path is the optimal job path that visits all job positions sequentially from the starting position and satisfies the job time window requirements of each node.

[0152] During the search process, if certain path nodes are found to be unreachable before their operation deadline, the path branches formed by that node and its subsequent dependent nodes are pruned to avoid invalid calculations. Finally, by backtracking the predecessor node pointers, starting from the last operation location node and tracing back to the starting node, a complete sequence of path nodes is generated. This sequence clearly defines the order in which the air cannon operating device will pass through each operation location from its current position, along with its estimated arrival time, forming an operational path that directly guides the equipment's execution. The time information attached to each node in the path provides a precise time reference for equipment scheduling, ensuring that the equipment arrives at the appropriate location at the appropriate time, achieving effective spatial blocking of cloud propagation.

[0153] A second aspect of the present invention provides an air gun operation path planning system based on cloud maps, comprising:

[0154] The cloud map acquisition unit is used to acquire cloud map data of the target area;

[0155] Evolutionary building units are used to perform spatiotemporal coupling analysis on cloud image data, extract dynamic propagation characteristic parameters of cloud layers, and construct a spatiotemporal evolution field of cloud layers through cloud centroid trajectory tracking and boundary diffusion situation inference, thereby obtaining the spatial distribution evolution sequence of cloud layers in future time periods.

[0156] The node identification unit is used to identify cloud propagation path nodes based on the spatial distribution evolution sequence, calculate the time-matching relationship between candidate operation positions and propagation path nodes by combining the operation response delay of the air cannon operation device, and identify priority operation positions that have a spatial blocking effect on cloud propagation based on the time-matching relationship.

[0157] The temporal constraint unit is used to construct a temporal constraint diagram of the operation positions based on the spatial connectivity and temporal relationship of the priority operation positions in the cloud propagation path, generate an operation sequence based on the temporal constraint diagram, calculate the perturbation effect of the preceding operation positions in the operation sequence on the spatiotemporal evolution field of the cloud, and recalculate the timing matching relationship of the subsequent operation positions based on the updated spatiotemporal evolution field of the cloud.

[0158] The path planning unit is used to take each operation position in the operation sequence as a path node, construct a spatiotemporal constraint network based on the matching relationship between movement time cost and timeliness, and perform dynamic planning search on the spatiotemporal constraint network to obtain the operation path of the air cannon operation device passing through each path node in sequence.

[0159] The operation execution unit is used to control the air cannon operating device to perform operations along the operating path.

[0160] A third aspect of the present invention provides an electronic device, comprising:

[0161] processor;

[0162] Memory used to store processor-executable instructions;

[0163] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0164] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0165] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0166] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for planning the operational path of an air cannon based on cloud maps, characterized in that, include: Obtain cloud map data for the target area; Spatiotemporal coupling analysis of cloud image data is performed to extract dynamic propagation characteristic parameters of cloud layers. By tracking the cloud centroid trajectory and extrapolating the boundary diffusion situation, a spatiotemporal evolution field of cloud layers is constructed to obtain the spatial distribution evolution sequence of cloud layers in future time periods. Based on the spatial distribution evolution sequence, cloud propagation path nodes are identified. The time-sensitivity matching relationship between candidate operation locations and propagation path nodes is calculated by combining the operation response delay of the air cannon operation device. Based on the time-sensitivity matching relationship, priority operation locations that have a spatial blocking effect on cloud propagation are identified. A temporal constraint diagram of the operation positions is constructed based on the spatial connectivity and temporal relationship of the priority operation positions in the cloud propagation path. An operation sequence is generated based on the temporal constraint diagram. The perturbation effect of the preceding operation positions in the operation sequence on the spatiotemporal evolution field of the cloud is calculated. The timing matching relationship of the subsequent operation positions is recalculated based on the updated spatiotemporal evolution field of the cloud. Each work location in the work sequence is taken as a path node. A spatiotemporal constraint network is constructed based on the matching relationship between movement time cost and timeliness. Dynamic programming search is performed on the spatiotemporal constraint network to obtain the work path of the air cannon working device passing through each path node in sequence. Control the air cannon operating device to perform operations along the operating path.

2. The method according to claim 1, characterized in that, Spatiotemporal coupling analysis of cloud image data is performed to extract dynamic propagation characteristic parameters of the cloud layer. By tracking the cloud centroid trajectory and extrapolating the boundary diffusion pattern, a spatiotemporal evolution field of the cloud layer is constructed, yielding the spatial distribution evolution sequence of the cloud layer in the future, including: Spatial gradient field calculation is performed on the cloud image data to extract the spatial gradient vector field of cloud density distribution. Then, the spatial gradient vector field is decomposed into divergence and curl to obtain the divergence propagation component and rotation propagation component of the cloud layer. The dynamic propagation characteristics of the cloud layer are obtained by vector synthesis of the divergent propagation component and the rotational propagation component. Based on the aforementioned dynamic propagation characteristics, the centroid coordinates of the cloud layer are calculated. Time series analysis is then performed on the centroid coordinates, and the velocity vector and acceleration vector of the centroid displacement are fitted to obtain the cloud layer centroid trajectory. Based on the dynamic propagation characteristics, the normal diffusion rate at the cloud boundary point is determined, and the cloud boundary is extrapolated along the cloud centroid trajectory with a time step to obtain the diffusion pattern of the cloud boundary; The cloud centroid trajectory and the diffusion pattern of the cloud boundary are mapped in spatiotemporal coordinates to construct the spatiotemporal evolution field of the cloud. The spatiotemporal evolution field of the cloud is then extrapolated along the time axis to obtain the spatial distribution evolution sequence.

3. The method according to claim 1, characterized in that, Based on the spatial distribution evolution sequence, cloud propagation path nodes are identified. The time-sensitivity matching relationship between candidate operational locations and propagation path nodes is calculated by combining the operational response delay of the air cannon operating device. Based on this time-sensitivity matching relationship, priority operational locations that spatially block cloud propagation are identified, including: Feature direction extraction is performed on the spatial distribution evolution sequence to obtain the direction vector of cloud propagation. The spatiotemporal evolution field of the cloud is divided into equally spaced spatial slices along the direction vector. Local maxima of cloud density are identified in each spatial slice, and the local maxima are marked as propagation path nodes. A spatiotemporal coordinate index is established for each propagation path node, the spatiotemporal coordinate index containing the spatial location coordinates of the propagation path node and the time of its appearance in the spatial distribution evolution sequence; For each candidate operation location, the Euclidean distance between the candidate operation location and the spatial coordinates of each propagation path node is calculated. The Euclidean distance is then converted into the operation response delay of the air cannon operation device to obtain the operation time deviation of the candidate operation location to each propagation path node. A timeliness matching relationship is then constructed based on the operation time deviation and the occurrence time of the propagation path node. Spatial coverage assessment of timeliness matching relationship is conducted, and priority operation locations are determined based on the assessment results.

4. The method according to claim 3, characterized in that, A spatial coverage assessment of the timeliness matching relationship is conducted, and priority operation locations are determined based on the assessment results, including: For each candidate job location, extract the propagation path nodes in its timeliness matching relationship where the job time deviation is less than a preset timeliness threshold, and form the propagation path nodes into a set of coverage nodes for that candidate job location. Calculate the spatial span of the covered node set in the cloud propagation path, where the spatial span is the projection of the distance between the foremost and last nodes in the covered node set along the propagation direction; The path coverage of the candidate job location is calculated based on the spatial span and the total length of the cloud propagation path. The density coverage of the candidate job location is obtained by summing the weighted average cloud density of each propagation path node in the coverage node set. The path coverage and density coverage are weighted and fused to obtain the spatial blocking effectiveness index of the candidate operation location. The candidate operation locations are sorted according to the spatial blocking effectiveness index, and the candidate operation location with the highest spatial blocking effectiveness index is selected as the priority operation location.

5. The method according to claim 1, characterized in that, A temporal constraint map of the priority operation locations is constructed based on the spatial connectivity and temporal relationship of the locations within the cloud propagation path. An operation sequence is generated based on this temporal constraint map. The perturbation effect of preceding operation locations on the spatiotemporal evolution field of the cloud is calculated for each preceding operation location in the sequence. Based on the updated spatiotemporal evolution field of the cloud, the timing matching relationship of subsequent operation locations is recalculated, including: The adjacency relationship of the priority work positions in the cloud propagation path is calculated to determine the spatial connectivity between each priority work position. Based on the temporal relationship of the nodes in the propagation path, the order of each priority work position in the cloud propagation process is determined, and directed connection edges are established between the priority work positions, with the direction of the directed connection edges pointing to the downstream work positions in the propagation path. Using the priority job positions as graph nodes and the directed connecting edges as graph edges, a temporal constraint graph of job positions is constructed. The directed connecting edges are assigned time constraint weights, which correspond to the time interval between the completion of a job at a previous job position and the commencement of a job at a subsequent job position. Perform topological sorting on the timing constraint graph, and generate a job sequence based on the topological sorting result; For the preceding operation position in the operation sequence, the disturbance effect of the simulated air gun operation device on the spatiotemporal evolution field of the cloud layer during the operation at the preceding operation position is calculated. The changes in cloud density distribution and the deflection of propagation direction caused by the disturbance effect are calculated, and the spatiotemporal evolution field of the cloud layer is updated. Based on the updated spatiotemporal evolution field of the cloud layer, the timeliness matching relationship of the subsequent operation positions is recalculated.

6. The method according to claim 5, characterized in that, The simulation of the air cannon's operation at the previous operational position caused disturbances to the spatiotemporal evolution of clouds. The calculation of the cloud density distribution changes and propagation direction deflections caused by the disturbances included: Using the location of the preceding operation as the disturbance source point, calculate the arrival time of the wavefront at different spatial locations as the disturbance wave propagates radially outward in the spatiotemporal evolution field of the cloud. Based on the operational energy release characteristics of the air cannon operating device, the initial disturbance intensity at the disturbance source point is calculated, and the initial disturbance intensity is spatially attenuated and corrected according to the propagation distance of the disturbance wave and the arrival time of the wavefront to obtain the disturbance intensity distribution at each spatial location. The disturbance intensity at each spatial location is applied to the cloud density value at the corresponding spatial location in the spatiotemporal evolution field of the cloud, the reduction of cloud density by the disturbance intensity is calculated, and the cloud density distribution is updated according to the reduction. Based on the vector angle between the propagation direction of the disturbance wave and the original cloud propagation direction, the deflection moment generated by the disturbance wave on the cloud propagation direction is calculated, and the cloud propagation direction is updated based on the deflection moment.

7. The method according to claim 1, characterized in that, Taking each operation location in the operation sequence as a path node, a spatiotemporal constraint network is constructed based on the matching relationship between movement time cost and timeliness. Dynamic programming search is then performed on the spatiotemporal constraint network to obtain the operation path of the air cannon operating device sequentially passing through each path node, including: The three-dimensional spatial coordinates of each operation position in the operation sequence are extracted, and the operation position is used as a path node, with the current position of the air cannon operation device as the starting path node; Calculate the three-dimensional spatial distance between each path node, and based on the movement speed characteristics of the air cannon operating device under different terrain conditions, convert the three-dimensional spatial distance into the movement time cost between the path nodes; The operation time window corresponding to each path node is extracted from the timeliness matching relationship. The operation time window includes the operation start time and operation end time of the path node. The movement time cost is coupled with the operation time window in a time sequence to construct a spatiotemporal constraint network. The connection edges between each path node in the spatiotemporal constraint network are set with time constraint weights. The time constraint weights record the time boundary conditions corresponding to the arrival of the subsequent path node. Dynamic programming search is performed on the spatiotemporal constrained network. Starting from the initial path node, the cumulative time cost to reach each path node is calculated under the constraint of time constraint weights. The path node sequence with the minimum cumulative time cost is selected to generate the operation path of the air cannon operation device passing through each path node in sequence.

8. A cloud-based air cannon operation path planning system, used to implement the method as described in any one of claims 1-7, characterized in that, include: The cloud map acquisition unit is used to acquire cloud map data of the target area; Evolutionary building units are used to perform spatiotemporal coupling analysis on cloud image data, extract dynamic propagation characteristic parameters of cloud layers, and construct a spatiotemporal evolution field of cloud layers through cloud centroid trajectory tracking and boundary diffusion situation inference, thereby obtaining the spatial distribution evolution sequence of cloud layers in future time periods. The node identification unit is used to identify cloud propagation path nodes based on the spatial distribution evolution sequence, calculate the time-matching relationship between candidate operation positions and propagation path nodes by combining the operation response delay of the air cannon operation device, and identify priority operation positions that have a spatial blocking effect on cloud propagation based on the time-matching relationship. The temporal constraint unit is used to construct a temporal constraint diagram of the operation positions based on the spatial connectivity and temporal relationship of the priority operation positions in the cloud propagation path, generate an operation sequence based on the temporal constraint diagram, calculate the perturbation effect of the preceding operation positions in the operation sequence on the spatiotemporal evolution field of the cloud, and recalculate the timing matching relationship of the subsequent operation positions based on the updated spatiotemporal evolution field of the cloud. The path planning unit is used to take each operation position in the operation sequence as a path node, construct a spatiotemporal constraint network based on the matching relationship between movement time cost and timeliness, and perform dynamic planning search on the spatiotemporal constraint network to obtain the operation path of the air cannon operation device passing through each path node in sequence. The operation execution unit is used to control the air cannon operating device to perform operations along the operating path.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images