A waterlogging event merging method, medium and device

Abstract

The present application relates to the technical field of smart city and flood prevention emergency command, and particularly relates to a waterlogging event merging method, medium and equipment, which maps waterlogging clue data to a global spatial grid index and only queries existing waterlogging events in the current grid and adjacent grids, thereby improving the merging efficiency of waterlogging events; the clues with an elevation difference exceeding a threshold or topological inaccessibility are screened and removed through a space-time distance threshold, thereby correctly distinguishing different stereoscopic levels and physically isolated independent waterlogging events while taking into account the space-time proximity, and then merging the clues falling into the same business semantic entity spatial fence into the same waterlogging event, so that the merging result is consistent with the flood prevention disposal unit boundary, and the merging accuracy and reliability are improved; the waterlogging event is newly created and associatedly stored in the grid unit, thereby having the ability to dynamically construct events and the ability to completely track the whole life cycle of waterlogging events from generation to recession.

Description

Technical Field

[0001] This invention relates to the fields of smart cities and flood control emergency command, and in particular to a method, medium and equipment for merging waterlogging events. Background Technology

[0002] With the advancement of smart city construction, visual AI technology has been widely applied in urban flood control. Image acquisition devices deployed at key locations can identify ground water accumulation in real time, forming a business chain for water accumulation identification, severity assessment, and reporting of water accumulation events. However, the same flood control unit often deploys multiple acquisition devices. Existing water accumulation event identification and merging methods process the water accumulation identification results pushed by each device independently, lacking an effective spatial aggregation mechanism. This leads to the system generating a large number of independent water accumulation alarms with similar geographical locations, resulting in duplicate reporting of water accumulation events and increasing the analytical burden on flood control commanders. Furthermore, existing merging methods often use simple two-dimensional planar distance thresholds for water accumulation event clustering. However, urban space has significant three-dimensional complexity, and simple two-dimensional distance threshold aggregation cannot identify complex scenes, easily leading to the incorrect merging of independent water accumulation events at different three-dimensional levels or in different traffic directions, reducing the accuracy of merging. Moreover, the basis of flood control scheduling in existing methods is business semantic units, such as the area within a residential community, one side of a tunnel, or the jurisdiction of a pumping station. The spatial forms of these units are often irregular and closely related to semantic elements such as road networks, administrative divisions, and facility boundaries. Clustering algorithms based purely on geometric distance cannot identify irregular entity boundaries, leading to inconsistencies between the merging results and the division of flood control and response units, making it difficult to directly support flood control command and dispatch.

[0003] Therefore, how to accurately aggregate fragmented flood warning and flood clue data into flood events that closely reflect actual flood control and response units has become an urgent problem to be solved. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention provides a method for merging water accumulation events, which includes the following steps: S1. For any acquired water accumulation clue data, determine the current grid of the current water accumulation clue data in the global spatial grid index according to the spatial coordinates, and query whether there are active water accumulation events in the current grid and the adjacent grids of the current grid. The active water accumulation event is a water accumulation event that is in progress and has not yet ended.

[0005] S2, if there is an active water accumulation event, the active water accumulation event is determined as the candidate water accumulation event corresponding to the current water accumulation clue data, and for any candidate water accumulation event, the spatiotemporal distance between the current water accumulation clue data and the current candidate water accumulation event is calculated.

[0006] S3. If the spatiotemporal distance is less than the preset initial screening threshold, the current water accumulation clue data is assigned to the preset clue set corresponding to the current candidate water accumulation event, and all water accumulation clue data are traversed. Based on all water accumulation clue data in the preset clue set, a candidate merging cluster corresponding to the current candidate water accumulation event is formed.

[0007] S4. Based on the preset elevation data and road network topology, remove water accumulation clue data from each candidate merge cluster that have an elevation difference greater than the preset three-dimensional layering threshold or are inaccessible by the road network topology, to obtain the corresponding spatiotemporal aggregated water accumulation clue dataset.

[0008] S5 calculates the spatial intersection of the spatial coordinates of each water accumulation clue data in each spatiotemporal aggregated water accumulation clue data set with the spatial fence of the business semantic entity, and merges all water accumulation clue data falling within the same spatial fence of the business semantic entity into the same water accumulation event.

[0009] S6. If there are no active water accumulation events, create a new water accumulation event based on the current water accumulation clue data, and store the new water accumulation event in the current grid.

[0010] The present invention also provides a non-transitory computer-readable storage medium storing at least one instruction or at least one program, wherein the at least one instruction or at least one program is loaded and executed by a processor to implement the above-described method for merging water accumulation events.

[0011] The present invention also provides an electronic device, including a processor and the aforementioned non-transitory computer-readable storage medium.

[0012] This invention has at least the following beneficial effects: By mapping water accumulation clue data to a global spatial grid index and querying only existing water accumulation events within the current grid and adjacent grids, the retrieval scope is reduced from global traversal to local neighborhood, improving the merging efficiency of water accumulation events; by calculating the spatiotemporal distance between the current water accumulation clue data and candidate water accumulation events, and grouping water accumulation clue data that meet the threshold conditions into the corresponding water accumulation clue dataset to form a candidate merging cluster, it simultaneously considers spatial proximity and temporal similarity, avoiding the mis-merging of water accumulation events that are far apart in time by relying solely on spatial distance, thus improving the accuracy of merging; by moving from the candidate merging cluster... Except for water accumulation clue data with elevation differences exceeding the threshold or topological inaccessibility, it correctly distinguishes independent water accumulation events at different three-dimensional levels and with physical isolation, solving the problem of mismerging in three-dimensional scenes and complex road network scenes; by merging water accumulation clue data falling within the same fence into the same water accumulation event, the merging result is consistent with the boundary of the actual flood control and disposal unit, further improving the accuracy and reliability of merging; by creating water accumulation events when no active water accumulation events exist and storing them in the corresponding grid unit, it has the ability to dynamically construct water accumulation events from scratch, as well as the ability to fully track the entire life cycle of water accumulation events from generation to dissipation. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a flowchart of a water accumulation event merging method provided in Embodiment 1 of the present invention. Detailed Implementation

[0015] 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.

[0016] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It is understood that, where appropriate, the terms used to distinguish similar objects can be interchanged so that the invention can also be implemented in other embodiments besides the illustrated or described embodiments. Furthermore, the terms "including," "having," and any variations are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0017] Example 1 This embodiment provides a method for merging water accumulation events, such as... Figure 1 As shown, the method for merging water accumulation events includes the following steps: S1. For any acquired water accumulation clue data, determine the current grid in the global spatial grid index based on the spatial coordinates, and query whether there are active water accumulation events in the current grid and its adjacent grids.

[0018] Among them, the water accumulation clue data is structured data obtained after standardizing the water accumulation identification results pushed by the front-end sensing device. It includes at least: a standard timestamp representing the time when the water accumulation was identified, the spatial coordinates of the location of the water accumulation in a unified spatial coordinate system, and water accumulation feature evaluation values ​​representing the severity of the water accumulation, such as water accumulation depth level and water accumulation area percentage.

[0019] Front-end sensing devices are visual recognition-capable acquisition devices deployed at key locations such as urban road sections prone to flooding, underpasses, and tunnel entrances and exits. These include cameras with AI analysis capabilities or edge computing nodes, used to acquire real-time images of the scene and output flooding identification results through built-in or back-end flooding identification algorithms. Those skilled in the art will recognize that any existing edge computing node with AI analysis capabilities falls within the scope of this invention. For example, the NVIDIA Jetson Orin NX can directly run the flooding identification algorithm locally on the front-end device, without needing to transmit the entire video stream back to the central server. By performing local AI inference on the real-time video stream captured by the camera, it can directly generate flooding identification results on-site, including information such as flooding depth level and area percentage, thereby significantly reducing network bandwidth consumption and central computing load. Further details will not be elaborated here.

[0020] The continuous geographic space is discretized into several non-overlapping grid cells according to preset grid hierarchy parameters, and a unique grid code is assigned to each grid cell to obtain a global spatial grid index. Each grid code is associated with a storage space used during system operation to store the water accumulation event identifier, the timestamp of the water accumulation event's center point, and the spatial coordinates when the spatial coordinates of a newly created water accumulation event fall within that grid range after being mapped by the code. In the initial state of index construction, the set of water accumulation event identifiers associated with each storage space is empty. Simultaneously, the index pre-stores the topological adjacency relationships between each grid code and its neighboring grid codes.

[0021] A water accumulation event goes through four states during its life cycle: newly discovered, developing, receding, and receding. Among them, an active water accumulation event is a water accumulation event that is currently in progress and has not yet ended, corresponding to the three types of water accumulation events: newly discovered, developing, and receding.

[0022] In urban flood control scenarios, there are a large number of existing active waterlogging events. If a naive global traversal method is used, the spatial distance to each existing waterlogging event must be calculated and sorted for each new waterlogging clue data arrival, resulting in high time complexity and failing to meet the processing requirements for rapid and continuous push of waterlogging clue data. Furthermore, if a waterlogging clue data falls near the boundary of the current grid, and the center point of a spatially adjacent waterlogging event happens to fall in an adjacent grid, querying only the current grid will miss this candidate waterlogging event.

[0023] Therefore, each existing water accumulation event is mounted into the storage space of the corresponding grid according to its spatial coordinates. When new water accumulation clue data arrives, only the grid code corresponding to its spatial coordinates needs to be calculated, which can reduce the search scope from the full water accumulation event database to the set of water accumulation events in the current grid and adjacent grids, improving efficiency while ensuring that no spatially adjacent water accumulation events are missed due to grid boundary effects.

[0024] If a new water accumulation clue is spatially adjacent to a water accumulation event that has receded, it indicates that the water accumulation clue corresponds to a new water accumulation process. It should not be directly classified into a completed water accumulation event. Instead, subsequent steps should determine whether to reuse a historical water accumulation event or create a new water accumulation event. Therefore, this embodiment limits the query to active water accumulation events to exclude historical water accumulation events that have receded.

[0025] In one specific implementation, the construction of the global spatial grid index includes the following steps: S101, obtain the preset spatial grid encoding algorithm and preset grid level parameters.

[0026] S102, based on the deployment range and grid level parameters of several front-end sensing devices, determine the geographic spatial range covered by the global spatial grid index.

[0027] S103 uses a spatial grid coding algorithm to divide the geographic space into grids and generate several grid units, where each grid unit has a unique grid code.

[0028] S104, establish a corresponding storage space for each grid code. The storage space is used to associate and store the water accumulation event identifier, the center timestamp of the center point of the water accumulation event and the spatial coordinates of the newly created water accumulation event when the spatial coordinates of the newly created water accumulation event fall within the range of the grid code after being mapped by the spatial grid coding algorithm. In the initial state, the set of water accumulation event identifiers associated with each storage space is an empty set.

[0029] S105. Based on the neighborhood rules corresponding to the spatial grid coding algorithm, determine and store the topological adjacency relationship between each grid code and its adjacent grid codes to obtain the global spatial grid index.

[0030] The spatial grid coding algorithm is used to convert continuous geospatial coordinates into discrete grid cell codes, so that spatially adjacent points have the same or adjacent grid codes, thereby enabling fast indexing and neighborhood retrieval of spatial data. Those skilled in the art will recognize that any existing spatial grid coding algorithm falls within the protection scope of this invention, such as the H3 hexagonal grid coding algorithm and the Geohash rectangular grid coding algorithm, etc., and will not be elaborated upon here.

[0031] This embodiment determines the geographic spatial range covered by the global spatial grid index based on the deployment range and grid level parameters of several front-end sensing devices. This ensures that the global spatial grid index can cover all geographic areas that may generate water accumulation alarms, avoiding data mapping failures due to insufficient index coverage. Specifically, the installation location coordinates of all front-end sensing devices are collected to form a coordinate set. The minimum bounding rectangle containing all coordinates in this coordinate set is calculated, and its boundary is determined by the minimum and maximum longitude and latitude values. Finally, a preset margin is extended outward in each direction to obtain the geographic spatial range covered by the index. The specific value of the preset margin can be set by the implementer according to the actual situation. In this embodiment, it is set to 500 meters to cover the water accumulation areas that may occur outside the device deployment boundary.

[0032] A grid code is a unique string or numerical value that identifies a grid cell. Storage space is a data container allocated for each grid code to store information about water accumulation events associated with that grid cell. When the spatial coordinates of a newly created water accumulation event fall within the grid range corresponding to a certain grid code, the water accumulation event identifier, the center timestamp of the water accumulation event's center point, and the spatial coordinates are written to the storage space corresponding to that grid code. Subsequent queries can then quickly retrieve all existing water accumulation events within that grid range simply by using the grid code.

[0033] It should be noted that before the global spatial grid index is built and the system begins processing any water accumulation clue data, no water accumulation events are associated with and stored in the storage space corresponding to each grid code. Water accumulation events are dynamically written to the storage space of the corresponding grid during system operation, such as when a new water accumulation event is created.

[0034] By using neighborhood rules, all adjacent grid codes corresponding to each grid code can be pre-calculated and stored in the index. When processing water accumulation clue data, there is no need to calculate adjacent grids in real time; the list of adjacent grids can be obtained by directly querying the pre-stored adjacency relationships.

[0035] As described above, by mapping new water accumulation clue data to the global spatial grid index and querying only active water accumulation events within the current grid and adjacent grids, the retrieval scope of candidate water accumulation events is reduced from full traversal to local neighborhood search, thereby improving the real-time processing efficiency of water accumulation event merging.

[0036] In one specific embodiment, S1 includes the following steps: S111 receives several water accumulation identification results pushed by each front-end sensing device.

[0037] S112, standardize each water accumulation identification result to obtain corresponding water accumulation clue data. The water accumulation clue data includes at least a standard timestamp, spatial coordinates, and water accumulation feature evaluation values. The standard timestamp is obtained by formatting the identification timestamp into a preset standard time format. The spatial coordinates are obtained by converting the collection location coordinates into a unified spatial coordinate system. The water accumulation feature evaluation values ​​include the water accumulation depth level and water accumulation area percentage extracted from the water accumulation attribute information.

[0038] Among them, the water accumulation identification result is the original alarm information output by the front-end sensing device after analyzing the collected on-site images. It includes at least the device identifier that identifies which front-end sensing device reported the result, the identification timestamp that records the time when the water accumulation was identified by the algorithm, the collection location coordinates that represent the latitude and longitude of the device installation location, and water accumulation attribute information, such as water accumulation depth and area, which are descriptive fields output by the identification algorithm.

[0039] Because the water accumulation recognition results output by front-end sensing devices from different manufacturers and models vary in terms of field naming, data format, and coordinate system selection, standardization is required for unified use. Specifically, the recognition timestamp field in the water accumulation recognition results is extracted, parsed and reformatted according to a preset standard time format, and converted to a standard time zone to obtain a standard timestamp. The acquisition location coordinates in the water accumulation recognition results are extracted, and a coordinate transformation function is called to convert the original coordinates into longitude and latitude values ​​in a unified spatial coordinate system to obtain spatial coordinates. The description field in the water accumulation attribute information is parsed to extract data related to water accumulation depth and area. For water accumulation depth, it is mapped to discrete water accumulation depth levels; for water accumulation area, the pixel ratio of the water accumulation area in the detection image is calculated to obtain the water accumulation area ratio. The obtained standard timestamp, spatial coordinates, and water accumulation feature evaluation values ​​are then packaged into a single water accumulation clue data according to a preset data structure template.

[0040] The above-mentioned standardization of water accumulation identification results provides a stable and reliable data foundation for subsequent steps such as spatial retrieval and spatiotemporal distance calculation.

[0041] In one specific embodiment, S1 further includes the following steps: S121 transforms the spatial coordinates of the current water accumulation clue data to the target spatial coordinate system corresponding to the grid level parameters, thus obtaining the target coordinates.

[0042] S122, The target coordinates are encoded and calculated according to the grid coding algorithm to obtain the grid code corresponding to the current water accumulation clue data.

[0043] S123, determine the grid cell corresponding to the grid code as the current grid corresponding to the current water accumulation clue data.

[0044] The target spatial coordinate system is a standard coordinate system used for grid coding calculations and is bound to the spatial grid coding algorithm. This is to avoid the coordinates of the same physical location being mapped to different grid codes on different devices due to coordinate system mismatch. For example, the H3 hexagonal grid coding algorithm requires the input coordinates to be latitude and longitude coordinates on the WGS84 ellipsoid, while the Geohash rectangular grid coding algorithm is usually based on the WGS84 or Web Mercator coordinate system.

[0045] As described above, by converting the spatial coordinates of the water accumulation clue data into a target coordinate system that matches the encoding algorithm and performing encoding calculations, and then mapping the encoding to the grid cells in the index, the consistency of the mapping results of the water accumulation clue data at the same location at different times is ensured, providing a stable and accurate retrieval entry point for subsequent neighborhood retrieval.

[0046] In one specific embodiment, S1 further includes the following steps: S131, based on the grid code of the current grid and the grid codes of adjacent grids, query the water accumulation event identifier associated with each grid code in the global spatial grid index to obtain all initial water accumulation events that fall into the current grid and adjacent grids.

[0047] S132, based on the water accumulation event identifier of each initial water accumulation event, query the current status corresponding to each initial water accumulation event recorded in the preset water accumulation event status maintenance module, wherein the current status is newly discovered, developing, receding, or already receded.

[0048] S133, an initial water accumulation event that is currently in a newly discovered, developing, or receding state is identified as an active water accumulation event.

[0049] The water accumulation event identifier is a unique identifier assigned by the system when a water accumulation event is first created, used to uniquely identify a water accumulation event throughout its entire lifecycle. A query code list is formed by combining the current grid's grid code and the grid codes of all adjacent grids. Then, using each grid code in the query code list as a key, the list of water accumulation event identifiers recorded in the storage space corresponding to that grid code is read from the global spatial grid index. Finally, all read water accumulation event identifiers are deduplicated and summarized to form the initial water accumulation event set.

[0050] The pre-defined water accumulation event status maintenance module is a system component independent of the global spatial grid index, specifically responsible for recording and managing the current status of each water accumulation event. The current status is the lifecycle stage of a water accumulation event at the current moment in the system, and is divided into four types: Newly discovered, corresponding to the first creation of the water accumulation event, with no other water accumulation clue data yet included; Developing, corresponding to the inclusion of new water accumulation clue data, and the water accumulation event is still ongoing; Receding, corresponding to the water accumulation beginning to weaken, but still in an active period; and Receded, corresponding to the complete receding of the water accumulation, and the water accumulation event has ended.

[0051] The business semantic constraint for flooding events is as follows: if a new flooding clue is spatially adjacent to a flooding event that has already receded, it indicates that the flooding clue corresponds to a new round of flooding, meaning that the previous round of flooding has completely subsided and then re-occurred. It should not be directly classified as a terminated flooding event; instead, it should be determined whether it is a reuse of a historical flooding event based on the same meteorological cycle, or a newly created flooding event. Therefore, based on the current state, active flooding events are selected from the initial flooding events, while historical flooding events that have already receded are excluded.

[0052] As described above, by using grid codes as keys to batch query the index to obtain the initial set of spatially adjacent water accumulation events, and then querying the status of each water accumulation event across modules and filtering out active water accumulation events, the process of determining candidate water accumulation events is completed collaboratively by three decoupled sub-steps: spatial retrieval, status query, and status filtering. This ensures both the efficiency of retrieval and the elimination of interference from receding water accumulation events through status filtering, thus ensuring the correctness of the merging logic.

[0053] S2, if there is an active water accumulation event, the active water accumulation event is determined as the candidate water accumulation event corresponding to the current water accumulation clue data, and for any candidate water accumulation event, the spatiotemporal distance between the current water accumulation clue data and the current candidate water accumulation event is calculated.

[0054] Candidate water accumulation events are identified as existing water accumulation events that may belong to the same water accumulation event as the current water accumulation clue data, after spatial neighborhood retrieval and activity status filtering. One water accumulation clue data may correspond to multiple candidate water accumulation events, and the spatiotemporal distance of each needs to be calculated to determine which water accumulation event it belongs to or whether it is a newly created water accumulation event.

[0055] Since multiple observations belonging to the same flooding event should be spatially proximate and temporally continuous, the candidate flooding events only achieve spatial coarse screening through grid indexing. This ensures that the candidate flooding events and flooding clue data are spatially proximate, but it cannot distinguish the following two situations: first, the flooding clue data and the flooding event are spatially proximate but temporally distant, such as the flooding event from the previous rainfall that has subsided several hours ago; second, the flooding clue data and the flooding event are spatially proximate but belong to different road sections, such as details that cannot be distinguished by grid coarse screening. Therefore, this embodiment introduces spatiotemporal distance, and the spatial distance and temporal difference are weighted and fused into a single-generation value, providing a quantitative basis for merging judgment.

[0056] In one specific embodiment, S2 includes the following steps: S21, obtain the spatial coordinates of the center point of the current candidate water accumulation event, and calculate the spatial distance between the spatial coordinates of the current water accumulation clue data and the spatial coordinates of the center point of the water accumulation event.

[0057] S22, obtain the center timestamp of the center point of the current candidate water accumulation event, and calculate the time difference between the standard timestamp and the center timestamp of the current water accumulation clue data.

[0058] S23, the spatial distance and time difference are weighted and summed according to the preset weights to obtain the spatiotemporal distance.

[0059] The spatial coordinates of the center point of a flooding event are used to uniquely represent the spatial reference position of a candidate flooding event at the current moment. When a flooding event is first created, the spatial coordinates of the flooding clue data that created the event are used as the initial center point spatial coordinates. When new flooding clue data is added to the event, the position of the center point of the current candidate flooding event is re-evaluated. Specifically, for each flooding clue data added to the event, a weighted score is given based on its image quality factor and flooding feature evaluation value. For example, the image clarity, viewing angle coverage, flooding depth level, and flooding area ratio of the snapshot are weighted and scored. The flooding clue data with the highest comprehensive score is recommended as the center point of the flooding event. The spatial coordinates of the center point of the flooding event are used as the updated center point spatial coordinates, and the standard timestamp of the center point of the flooding event is used as the updated center timestamp.

[0060] The spatial coordinates of the epicenter of a flooding event represent the most representative spatial location of that event. By calculating the geometric distance between the spatial coordinates of the flooding clue data and the coordinates of the epicenter, a measure of the spatial dimensional difference can be obtained. Specifically, the shortest arc distance between two points on the Earth's ellipsoid can be calculated using the spherical distance formula, and this distance can be used as the spatial distance; alternatively, planar Euclidean distance can be used to calculate the spatial distance.

[0061] The center timestamp is a unique time value that represents the current time reference position of a candidate flooding event. When a flooding event is first created, the timestamp of the flooding clue data that created the event is used as the initial center timestamp; when new flooding clue data is added to the event, the timestamp of the latest added flooding clue data is updated as the center timestamp.

[0062] By calculating the difference between the timestamp of the water accumulation clue data and the central timestamp, and converting it into minutes, a measure of the difference in the time dimension can be obtained.

[0063] The preset weights include spatial distance weights and time difference weights. The relative importance of spatial and time factors can be adjusted according to the business scenario. For example, spatial weights can be increased in scenarios with high spatial concentration and large water accumulation areas; time weights can be increased in scenarios with long rainfall duration and frequent repeated water accumulation.

[0064] Spatiotemporal distance is used to quantify the comprehensive proximity of the current water accumulation clue data and the candidate water accumulation event in both spatial and temporal dimensions. Correspondingly, the smaller the spatiotemporal distance, the stronger the spatiotemporal correlation between the water accumulation clue data and the water accumulation event, and the more likely the water accumulation clue data belongs to the water accumulation event.

[0065] The above-mentioned method calculates the spatial distance and time difference between the current water accumulation clue data and the candidate water accumulation events, and then sums them according to preset weights to obtain the spatiotemporal distance. This expands the merging judgment from a single spatial proximity to a comprehensive spatiotemporal proximity assessment. While retaining the sensitivity of spatial screening, it introduces time constraints to avoid the erroneous merging of water accumulation clue data that are not continuous in time, and provides an adjustable comprehensive quantitative basis for subsequent threshold screening.

[0066] S3. If the spatiotemporal distance is less than the preset initial screening threshold, the current water accumulation clue data is assigned to the preset clue set corresponding to the current candidate water accumulation event, and all water accumulation clue data are traversed. Based on all water accumulation clue data in the preset clue set, a candidate merging cluster corresponding to the current candidate water accumulation event is formed.

[0067] The preset initial screening threshold is a pre-defined upper limit for spatiotemporal distance based on business experience. A current waterlogging clue is included in the waterlogging clue dataset of a candidate waterlogging event only if the spatiotemporal distance between the current waterlogging clue and that event is less than the preset initial screening threshold. If the spatiotemporal distance is greater than or equal to the preset initial screening threshold, the current waterlogging clue is not included in the current candidate waterlogging event, and the process continues to iterate through the remaining candidate waterlogging events, repeating the above judgment.

[0068] The preset clue set is a collection of all water accumulation clue data associated with the current candidate water accumulation event before the execution of this merging operation. The water accumulation clue data in this preset clue set comes from previous merging operations, and each time a merging operation is performed, a new water accumulation clue data is added to the set. It should be noted that the preset clue set is dynamically accumulated, and initially it only contains the water accumulation clue data that created the water accumulation event.

[0069] The water accumulation clue data is sorted by time. For each water accumulation clue, its spatiotemporal distance to each candidate water accumulation event is calculated. The first candidate water accumulation event that meets the threshold condition is identified as the merging target. Then, through the merging operation, the current water accumulation clue data is associated with the corresponding candidate water accumulation event, and the identifier of the water accumulation clue data is added to the preset clue set of the candidate water accumulation event. The merging operation does not create new water accumulation events, nor does it change other attributes of the water accumulation events; it only expands the range of water accumulation clue data associated with the water accumulation event.

[0070] After traversing all the water accumulation clue data, determine all the water accumulation clue data that need to be included in the preset clue set corresponding to the current candidate water accumulation event and perform the inclusion operation to obtain the corresponding candidate merging cluster.

[0071] It should be noted that by integrating all water accumulation clue data in the candidate merged cluster, the location of the water accumulation event center point is reassessed, making the verification operation more complete and the judgment more accurate.

[0072] As described above, by comparing the spatiotemporal distance with a preset initial screening threshold, the current water accumulation clue data is assigned to the preset clue set of the corresponding candidate water accumulation event and a candidate merging cluster is formed when the threshold condition is met. This completes the merging decision under the dual constraints of spatiotemporal constraints, ensuring the accuracy of the merging.

[0073] S4. Based on the preset elevation data and road network topology, remove water accumulation clue data from each candidate merge cluster that have an elevation difference greater than the preset three-dimensional layering threshold or are inaccessible by the road network topology, to obtain the corresponding spatiotemporal aggregated water accumulation clue dataset.

[0074] In one specific embodiment, S4 includes the following steps: S41, for any candidate merge cluster, obtain the elevation value of each water accumulation clue data in the current candidate merge cluster, as well as the elevation value of the water accumulation event center point of the candidate water accumulation corresponding to the current candidate merge cluster, based on the preset elevation data.

[0075] S42, for any water accumulation clue data in the current candidate merge cluster, calculate the elevation difference between the elevation value of the current water accumulation clue data and the elevation value of the corresponding water accumulation event center point.

[0076] S43. If there is no connected path between the spatial coordinates of the current waterlogging clue data and the corresponding waterlogging event center point within a preset number of topological steps, or if the spatial coordinates of the current waterlogging clue data and the corresponding waterlogging event center point belong to the opposite lane with physical isolation, then the current waterlogging clue data is determined to be unreachable in the road network topology.

[0077] S44. Remove the corresponding water accumulation clue data from the current candidate merged clusters that have an elevation difference greater than the preset three-dimensional layering threshold or are inaccessible by the road network topology, and obtain the corresponding spatiotemporal aggregated water accumulation clue dataset.

[0078] Urban spaces exhibit significant three-dimensional complexity: elevated roads and ground-level roads may have completely overlapping latitude and longitude coordinates on a two-dimensional plane, but their elevation values ​​differ by several meters to tens of meters; although opposite lanes on the same road may be only a few meters apart in planar distance, the presence of a central median prevents water accumulation on one side from spreading to the other. If only two-dimensional spatiotemporal distance is relied upon for merging, water accumulation clues in these scenarios will be incorrectly merged.

[0079] Therefore, this embodiment further introduces three-dimensional elevation information and road network topology constraints to perform refined verification and filtering of water accumulation clue data in candidate merge clusters, and removes water accumulation clue data that, although adjacent on the two-dimensional plane, should belong to different water accumulation events due to different three-dimensional levels or physical isolation of the road network.

[0080] Specifically, elevation constraints ensure that water accumulation clue data and water accumulation events within the same merge cluster are at the same 3D level, while topology constraints ensure that water accumulation clue data and water accumulation events are actually reachable from each other in the road network. The two constraints are OR-based: violating either one results in removal, ensuring the strictness of the verification and guaranteeing that the spatiotemporally aggregated water accumulation clue dataset has high consistency in both 3D space and the road network level.

[0081] Elevation data, obtained through geographic information systems or elevation services, records the altitude values ​​of any location in geographic space and is stored in the form of a digital elevation model. The elevation value of a corresponding location can be queried using latitude and longitude coordinates. In the scenario of merging water accumulation events, the coordinates of elevated roads and ground roads may completely overlap or be extremely close, but their elevation values ​​differ significantly.

[0082] Specifically, each water accumulation clue data in the current candidate merge cluster is traversed, and its spatial coordinates (longitude and latitude) are extracted. A preset digital elevation model (DEM) query interface is called, using longitude and latitude as input parameters, to obtain the elevation value of that coordinate location, which is recorded as the elevation value of the water accumulation clue data, i.e., the corresponding altitude value. Simultaneously, the spatial coordinates of the water accumulation event center point of the candidate water accumulation event corresponding to the current candidate merge cluster are extracted, and the elevation value of the water accumulation event center point is obtained in the same way. For each water accumulation clue data in the current candidate merge cluster, its elevation value is compared with the elevation value of the corresponding candidate water accumulation event center point, and the absolute difference between the two is calculated. This yields the elevation difference corresponding to the water accumulation clue data, used to quantify the difference between the water accumulation clue data and the water accumulation event in the vertical dimension: the smaller the difference, the more likely they are at the same three-dimensional level; the larger the difference, the more likely they belong to different three-dimensional levels.

[0083] The preset three-dimensional layering threshold is a critical elevation difference value set by the system to distinguish different three-dimensional layers. It can be set based on the usual elevation difference between urban elevated roads and ground roads. If the elevation difference between the water accumulation clue data and the center point of the water accumulation event is greater than the preset three-dimensional layering threshold, then the two are determined to be in different three-dimensional layers and should not be classified into the same water accumulation event.

[0084] Road network topology is a graph-structured data describing the connectivity between road segments in an urban road network. The road network is abstracted as a directed or undirected graph, where nodes represent road intersections or road segment endpoints, edges represent road segments, and the direction of the edges indicates the permitted direction of travel. This topology is used to support topological connectivity judgments, determining whether a reachable path exists between two spatial coordinate points in the road network.

[0085] In this embodiment, although the waterlogging clue data and the waterlogging event center point are adjacent on the plane, they may be separated by insurmountable obstacles such as rivers, railways, or enclosed walls in the road network map, resulting in no connecting path between them and therefore they should not be merged into the same waterlogging event. Specifically, the spatial coordinates of the waterlogging clue data and the spatial coordinates of the waterlogging event center point are mapped to the nearest road segment node in the road network connectivity map, and a search is conducted within a preset topology step count to determine if a connecting path exists between them; if not, it is determined to be topologically unreachable. The preset topology step count is used in the road network connectivity map to limit the maximum number of search steps for connectivity determination. For example, if the preset topology step count is 3, it only determines whether a connecting path exists between the two coordinate points within 3 road segments, controlling the search range while ensuring the rationality of the determination and avoiding the performance overhead caused by full-map search in a large road network. A connecting path refers to a continuous path in the road network topology map, from the road segment where the current waterlogging clue data spatial coordinates are located to the road segment where the candidate waterlogging event center point is located, formed by sequentially passing through several road segments in the allowed traffic direction. If there is at least one connected path and the number of road segments traversed is within the preset topology steps, then the two are determined to be topologically reachable.

[0086] However, even if a connection path exists, if the waterlogging clue data and the waterlogging event center point are located in opposite lanes of the same road, and that road has central physical barriers such as a median strip, guardrails, or green belts, they are not accessible in practical terms and should not be grouped into the same waterlogging event. Specifically, the lane direction attributes of the road segment containing the spatial coordinates of the waterlogging clue data and the road segment containing the spatial coordinates of the waterlogging event center point are obtained; if they are located in different lanes of the same road section, and that road segment is marked as having a central barrier, then it is determined to be topologically unreachable.

[0087] Traverse each water accumulation clue data in the current candidate merge cluster, and construct the spatiotemporal aggregated water accumulation clue dataset corresponding to the candidate water accumulation event based on the water accumulation clue data retained in the candidate merge cluster.

[0088] The above-mentioned approach, by introducing elevation difference calculation and road network topology accessibility judgment as dual verifications on the basis of spatiotemporal initial screening, and removing water accumulation clue data that does not meet any of the criteria from the candidate merging clusters, upgrades the merging process from two-dimensional planar clustering to three-dimensional and refined merging under road network constraints. This effectively solves the problem of erroneous merging in complex urban scenarios such as overlapping elevated roads and ground surfaces and physically isolated lanes, and ensures that the final set of water accumulation clue data entering business semantic merging has high consistency in the four dimensions of space, time, elevation, and topology.

[0089] S5 calculates the spatial intersection of the spatial coordinates of each water accumulation clue data in each spatiotemporal aggregated water accumulation clue data set with the spatial fence of the business semantic entity, and merges all water accumulation clue data falling within the same spatial fence of the business semantic entity into the same water accumulation event.

[0090] Among them, the business semantic entity spatial fence is a virtual geographic fence that corresponds to the spatial boundary of the actual handling unit in flood control command and dispatch. It can divide the geographic space into units with clear flood control business meaning. Unlike purely geometric buffers or grids, the business semantic entity spatial fence directly corresponds to the actual management objects such as a certain community, an underpass, or the jurisdiction of a certain pumping station. Its boundary is usually determined based on semantic information such as administrative divisions, road red lines, and facility management scope, rather than a fixed radius or regular shape.

[0091] In one specific implementation, the types of business semantic entity spatial fences include at least area fences, linear facility fences, and node facility fences.

[0092] Among them, the planar area fence is a two-dimensional area enclosed by a pre-defined polygonal boundary. It is used to represent flood control units with relatively large areas and irregular boundaries, such as residential communities, school campuses, large squares, and parking lots. Its boundaries are usually derived from administrative divisions, land parcel boundaries, or manual mapping in GIS layers. It can identify and merge water accumulation clue data from different locations within the same community and from different camera coverage areas.

[0093] Linear facility fences are pre-defined strip or line segments along roads or linear facilities. They are used to represent long-distance, narrow-strip flood control units, such as specific sections of a road, one side of a tunnel, ramps, etc. They usually include facility names and linear station information along the facility's direction. They can identify and merge water accumulation clue data from different locations or different camera perspectives on the same road segment.

[0094] A node facility fence is a buffer zone generated with a certain radius or preset shape, centered on the coordinates of a key flood control facility. It is used to represent flood control units centered on that facility, such as pumping stations, underpasses, and culverts. It can aggregate water accumulation clue data from multiple angles and devices surrounding a key facility to ensure that the water accumulation status of the facility is comprehensively assessed.

[0095] In this embodiment, although the spatiotemporally aggregated floodwater clue dataset obtained after elevation and topology verification is consistent across the four dimensions of space, time, elevation, and road network, it is still organized by candidate floodwater events. A single candidate floodwater event may span multiple business semantic units, or multiple candidate floodwater events may correspond to the same business semantic unit. Therefore, by introducing a business semantic entity spatial fence, the floodwater clue data is regrouped according to actual flood control and response units, ensuring that the merged floodwater events are fully aligned with the cognitive and operational units of command and dispatch personnel.

[0096] Correspondingly, the spatial coordinates of each water accumulation clue data point are spatially correlated with all pre-defined business semantic entity spatial fences one by one. If a water accumulation clue data point is located inside a fence, then the water accumulation clue data point is merged with the water accumulation events associated with that fence into the same water accumulation event. When water accumulation clue data points from multiple candidate merge clusters converge into the same fence, even if these water accumulation clue data points originally belonged to different candidate merge clusters, they are forcibly merged into the same water accumulation event. Conversely, if water accumulation clue data points from a candidate merge cluster are split into multiple fences because they fall within different fences, they will eventually be merged into the water accumulation events corresponding to their respective fences to correct cross-semantic unit errors that may have occurred in the previous clustering.

[0097] Furthermore, the spatial coordinates of all water accumulation clue data in each spatiotemporal aggregated water accumulation clue dataset are extracted. Then, the spatial coordinates are sequentially intersected with each fence in the pre-defined business semantic entity spatial fence layer.

[0098] For area fencing, typically used in large, irregularly shaped areas such as residential communities and campuses, where the internal spaces of the same management unit are interconnected and the causes of water accumulation are consistent, all water accumulation clue data within the same area are grouped into a unified water accumulation event. This facilitates resource allocation by flood control personnel on a community-by-community basis. Correspondingly, the spatial coordinates of the water accumulation clue data are extracted, and the GIS function for determining whether a point is inside a polygon is called. If the function returns true, the area identifier associated with that polygon is obtained, and the water accumulation clue data is assigned to the water accumulation event corresponding to that identifier.

[0099] For linear fencing, typically corresponding to linear structures such as roads and tunnels distributed along their direction, it's important to consider that different points on the same road segment might observe the same water accumulation event. However, points that are too far apart should be classified as independent water accumulation events. Therefore, the spatial coordinates of the water accumulation clue data are extracted, and its projection position and linear station number on the linear structure are calculated. If the distance between the water accumulation clue data and the station number of an existing water accumulation event on the same structure does not exceed a preset continuous threshold, it is included in that water accumulation event; otherwise, a new water accumulation event can be considered.

[0100] For node facility fencing, typically corresponding to key nodes such as pumping stations and underpasses, since the cause of water accumulation within a certain range centered on the facility is closely related to the facility's operational status (e.g., a pumping station malfunction causing surrounding water accumulation), all water accumulation clue data around the facility are consolidated for centralized assessment of the facility's status. Correspondingly, a circular or polygonal buffer zone is generated with a preset radius centered on the facility's coordinates. If the spatial coordinates of water accumulation clue data fall within this buffer zone, the water accumulation events associated with that facility are merged into a single water accumulation event.

[0101] The above-mentioned method calculates the spatial intersection of the water accumulation clue data set after four-dimensional verification (space, time, elevation, and topology) with the pre-set spatial fences of three types of business semantic entities: area, line, and node. It also forces the water accumulation clue data within the same fence to be merged into the same water accumulation event. This ensures that the merging result is fully aligned with the actual flood control and response unit boundary, solving the problem of not being able to identify irregular entity boundaries and generating redundant alarms. This provides accurate water accumulation event information for flood control command.

[0102] S6. If there are no active water accumulation events, create a new water accumulation event based on the current water accumulation clue data, and store the new water accumulation event in the current grid.

[0103] If the current water accumulation clue data does not find any existing water accumulation events that are in progress or receding within the spatial neighborhood, it usually corresponds to two possibilities: one is that it corresponds to a completely new water accumulation process, where water accumulation has never occurred in the area before or the previous round of water accumulation has completely receded and is not within the same meteorological cycle; the other is that although there are historical water accumulation events in the area that have receded, they should not be reused because they are not within the same meteorological cycle.

[0104] In this embodiment, when it cannot be determined whether the water accumulation clue data belongs to an existing active water accumulation event, it tends to create a new independent water accumulation event to avoid incorrectly classifying a new round of water accumulation into a completed historical water accumulation event. Once the newly created water accumulation event is associated and stored in the grid index, it gains the same status as all existing water accumulation events and can be retrieved and merged by subsequent water accumulation clue data. This enables the system to dynamically build a water accumulation event system from scratch: when the system first starts, there are no water accumulation events in the grid index, and all initial water accumulation clue data will trigger the creation of a new water accumulation event; as operation continues, newly created water accumulation events are continuously accumulated and written into the index, and subsequent water accumulation clue data can be retrieved and merged to form a positive cycle.

[0105] Specifically, when it is confirmed that no active water accumulation events exist in the query results, a new water accumulation event record is created for the current water accumulation clue data, and a unique water accumulation event identifier is assigned. The spatial coordinates of the current water accumulation clue data are set as the spatial coordinates of the water accumulation event center point of the new water accumulation event, and the standard timestamp of the current water accumulation clue data is set as the timestamp of the water accumulation event center point of the new water accumulation event. The status of the new water accumulation event is initialized to "New Discovery". The spatial coordinates of the current water accumulation clue data are mapped using a spatial grid encoding algorithm to obtain its corresponding grid code. The water accumulation event identifier, water accumulation event center point timestamp, and spatial coordinates of the new water accumulation event are written into the storage space corresponding to the grid code in the global spatial grid index.

[0106] As described above, by creating new water accumulation events when there are no active water accumulation events in the spatial neighborhood and storing them in the grid index, the system has the ability to dynamically build a water accumulation event system from scratch, realizing a complete tracking loop for the entire life cycle of water accumulation events from creation, development to disappearance.

[0107] In one specific implementation, the water accumulation event merging method further includes: based on the meteorological cycle identifier corresponding to the current rainfall event, retrieving historical water accumulation events that have receded within the same meteorological cycle and fall within the same business semantic entity spatial fence; if a historical water accumulation event that has receded is retrieved, its status is updated back to under development, and the number of times it has repeatedly accumulated water is added.

[0108] The meteorological cycle identifier is a unique identifier assigned by the system to each round of rainfall based on meteorological warning information or rainfall event segmentation rules. A typhoon, a continuous rainstorm, or a round of severe convective weather can all correspond to a meteorological cycle identifier, which is used to group water accumulation events occurring at different times according to rainfall events, supporting the association retrieval and status rollback of water accumulation events within the same rainfall cycle across time.

[0109] During prolonged rainfall cycles, water accumulation at the same location may repeatedly occur, recede, and re-accumulate. Historical receding water accumulation events refer to water accumulation events that were created before the current system time and whose current status is "receded." A historical receding water accumulation event indicates that the previously occurring water accumulation has completely subsided and the event has ended. However, if water accumulation reappears at the same location within the same meteorological cycle, it should not be considered a completely new and independent water accumulation event, but rather a repeated water accumulation process of the same event. The status of the receding water accumulation event should be changed from "receded" to "developing" through status rollback.

[0110] Repeated flooding frequency refers to the cumulative number of times the same flooding event experiences its generation, receding, and re-accumulation within the same meteorological cycle. The first flooding event is counted as one instance, and one instance is added each time the status is adjusted back. This is used to provide flood control management departments with an indicator of the degree of repeated flooding at flood-prone areas, and to help identify systemic problems such as backflow in drainage networks and insufficient pumping station capacity.

[0111] Specifically, the system retrieves the meteorological cycle identifier corresponding to the current rainfall event and the fence identifier of the business semantic entity space fence to which the current waterlogging clue data falls. In the historical waterlogging event database, using "same meteorological cycle identifier" and "same fence identifier" as joint query conditions, it retrieves historical waterlogging events with a status of "water receded". If a waterlogging event that meets the conditions is found, its status is updated from "water receded" to "developing". The repeated waterlogging counter corresponding to the waterlogging event is incremented by 1, and a corresponding repeated waterlogging tag is generated, such as "second repeated waterlogging". The current waterlogging clue data is then included in the waterlogging event after the callback. If no waterlogging event that meets the conditions is found, a new waterlogging event is created.

[0112] The above-mentioned method, by retrieving and reusing historical receding water accumulation events that fall within the same business semantic fence within the same meteorological cycle, and by reverting their execution status and accumulating the number of repeated water accumulations, has a complete lifecycle tracking capability for water accumulation events, providing a more accurate data foundation for flood control command and flood-prone area management.

[0113] As described above, by mapping water accumulation clue data to a global spatial grid index and querying only existing water accumulation events within the current grid and adjacent grids, the retrieval scope is reduced from global traversal to local neighborhood, improving the merging efficiency of water accumulation events. By calculating the spatiotemporal distance between the current water accumulation clue data and candidate water accumulation events, and grouping water accumulation clue data that meet the threshold condition into the corresponding water accumulation clue dataset to form candidate merging clusters, both spatial proximity and temporal similarity are taken into account, avoiding the mismerging of water accumulation events that are far apart in time by relying solely on spatial distance, thus improving the accuracy of merging. By removing water accumulation events with elevation differences exceeding the threshold from the candidate merging clusters, the retrieval scope is reduced from global traversal to local neighborhood, improving the merging accuracy. The system correctly distinguishes independent water accumulation events at different three-dimensional levels and with physical isolation from water accumulation clues that are unreachable by value or topology, thus solving the problem of mismerging in three-dimensional scenes and complex road network scenarios. By merging water accumulation clues that fall within the same fence into the same water accumulation event, the merged result is consistent with the boundary of the actual flood control and disposal unit, further improving the accuracy and reliability of merging. By creating new water accumulation events when no active water accumulation events exist and storing them in the corresponding grid cells, the system has the ability to dynamically construct water accumulation events from scratch and to fully track the entire life cycle of water accumulation events from generation to dissipation.

[0114] Example 2 Embodiment 2 of the present invention provides a non-transitory computer-readable storage medium, which can be disposed in an electronic device to store at least one instruction or at least one program related to implementing a method in the method embodiment. The at least one instruction or at least one program is loaded and executed by the processor to implement the water accumulation event merging method provided in the above embodiment.

[0115] Example 3 Embodiment 3 of the present invention provides an electronic device, which includes a processor and the non-transitory computer-readable storage medium of Embodiment 2 of the present invention.

[0116] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for merging water accumulation events, characterized in that, The method includes the following steps: S1, for any acquired water accumulation clue data, determine the current grid of the current water accumulation clue data in the global spatial grid index according to the spatial coordinates, and query whether there are active water accumulation events in the current grid and the adjacent grids of the current grid, wherein the active water accumulation event is a water accumulation event that is in progress and has not yet ended; S2, if the active water accumulation event exists, the active water accumulation event is determined as the candidate water accumulation event corresponding to the current water accumulation clue data, and for any candidate water accumulation event, the spatiotemporal distance between the current water accumulation clue data and the current candidate water accumulation event is calculated; S3, if the spatiotemporal distance is less than the preset initial screening threshold, the current water accumulation clue data is assigned to the preset clue set corresponding to the current candidate water accumulation event, and all water accumulation clue data are traversed, and a candidate merging cluster corresponding to the current candidate water accumulation event is formed based on all water accumulation clue data in the preset clue set. S4. Based on the preset elevation data and road network topology, remove water accumulation clue data from each candidate merge cluster that have an elevation difference greater than the preset three-dimensional layering threshold or are inaccessible by the road network topology to obtain the corresponding spatiotemporal aggregated water accumulation clue dataset. S5, calculate the spatial intersection of the spatial coordinates of each water accumulation clue data in each spatiotemporal aggregated water accumulation clue data set with the business semantic entity spatial fence, and merge all water accumulation clue data falling within the same business semantic entity spatial fence into the same water accumulation event. S6. If no active water accumulation event exists, a new water accumulation event is created based on the current water accumulation clue data, and the newly created water accumulation event is associated and stored in the current grid.

2. The method for merging waterlogging events according to claim 1, characterized in that, The construction of the global spatial grid index includes the following steps: S101, Obtain the preset spatial grid encoding algorithm and preset grid level parameters; S102, Based on the deployment range of several front-end sensing devices and the grid level parameters, determine the geographic spatial range covered by the global spatial grid index; S103, using the spatial grid coding algorithm, the geographic spatial range is divided into grids to generate several grid units, wherein each grid unit has a unique grid code; S104, establish a corresponding storage space for each grid code, wherein the storage space is used to associate and store the water accumulation event identifier, the center timestamp of the center point of the water accumulation event and the spatial coordinates of the newly created water accumulation event when the spatial coordinates of the newly created water accumulation event fall within the range of the grid code after being mapped by the spatial grid coding algorithm, and the water accumulation event identifier set associated with each storage space is an empty set in the initial state; S105, Based on the neighborhood rules corresponding to the spatial grid coding algorithm, determine and store the topological adjacency relationship between each grid code and its adjacent grid codes to obtain the global spatial grid index.

3. The method for merging waterlogging events according to claim 2, characterized in that, S1 includes the following steps: S111, receive several water accumulation identification results pushed by each front-end sensing device, wherein the water accumulation identification results include at least device identifier, identification timestamp, collection location coordinates and water accumulation attribute information; S112, standardize each water accumulation identification result to obtain corresponding water accumulation clue data. The water accumulation clue data includes at least a standard timestamp, spatial coordinates, and water accumulation feature evaluation values. The standard timestamp is obtained by formatting the identification timestamp into a preset standard time format. The spatial coordinates are obtained by converting the collection location coordinates into a unified spatial coordinate system. The water accumulation feature evaluation values ​​include the water accumulation depth level and water accumulation area percentage extracted from the water accumulation attribute information.

4. The method for merging water accumulation events according to claim 2, characterized in that, S1 also includes the following steps: S121, transform the spatial coordinates of the current water accumulation clue data to the target spatial coordinate system corresponding to the grid level parameters to obtain the target coordinates; S122, The target coordinates are encoded and calculated according to the grid coding algorithm to obtain the grid code corresponding to the current water accumulation clue data; S123, the grid cell corresponding to the grid code is determined as the current grid corresponding to the current water accumulation clue data.

5. The method for merging water accumulation events according to claim 2, characterized in that, S1 also includes the following steps: S131, based on the grid code of the current grid and the grid code of the adjacent grid, query the water accumulation event identifier associated with each grid code in the global spatial grid index to obtain all initial water accumulation events falling into the current grid and the adjacent grid; S132, based on the water accumulation event identifier of each initial water accumulation event, query the current status corresponding to each initial water accumulation event recorded in the preset water accumulation event status maintenance module, wherein the current status is newly discovered, developing, receding, or already receded; S133, the initial water accumulation event whose current state is newly discovered, developing or receding is determined as the active water accumulation event.

6. The method for merging water accumulation events according to claim 3, characterized in that, S2 includes the following steps: S21, obtain the spatial coordinates of the center point of the current candidate water accumulation event, and calculate the spatial distance between the spatial coordinates of the current water accumulation clue data and the spatial coordinates of the center point of the water accumulation event. S22, obtain the center timestamp of the center point of the current candidate water accumulation event, and calculate the time difference between the standard timestamp of the current water accumulation clue data and the center timestamp; S23, the spatial distance and the time difference are weighted and summed according to preset weights to obtain the spatiotemporal distance.

7. The method for merging waterlogging events according to claim 1, characterized in that, S4 includes the following steps: S41, for any candidate merge cluster, obtain the elevation value of each water accumulation clue data in the current candidate merge cluster according to the preset elevation data, as well as the elevation value of the water accumulation event center point of the candidate water accumulation corresponding to the current candidate merge cluster. S42, For any water accumulation clue data in the current candidate merge cluster, calculate the elevation difference between the elevation value of the current water accumulation clue data and the elevation value of the corresponding water accumulation event center point; S43, if there is no connected path between the spatial coordinates of the current waterlogging clue data and the corresponding waterlogging event center point within a preset number of topological steps, or if the spatial coordinates of the current waterlogging clue data and the corresponding waterlogging event center point belong to the opposite lane with physical isolation, then the current waterlogging clue data is determined to be unreachable in the road network topology. S44, Remove the corresponding water accumulation clue data that have an elevation difference greater than the preset three-dimensional layering threshold or are inaccessible by the road network topology from the current candidate merge clusters to obtain the corresponding spatiotemporal aggregated water accumulation clue dataset.

8. The method for merging waterlogging events according to claim 1, characterized in that, The types of business semantic entity spatial fences include at least planar area fences, linear facility fences, and node facility fences. The planar area fence is a preset polygonal boundary, the linear facility fence is a preset road segment range, and the node facility fence is a buffer area generated centered on preset core facility coordinates.

9. A non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores at least one instruction or at least one program segment, characterized in that, The at least one instruction or the at least one program segment is loaded and executed by the processor to implement the water accumulation event merging method as described in any one of claims 1-8.

10. An electronic device, characterized in that, Includes a processor and the non-transitory computer-readable storage medium as described in claim 9.

Images