Convective cell recognition method, device, equipment and storage medium

Abstract

This invention discloses a method, apparatus, device, and storage medium for identifying convective cells. The identification method includes extracting the maximum value from radar 3D network data along the height axis to obtain a two-dimensional combined echo; performing multi-level progressive threshold segmentation on the two-dimensional combined echo to obtain echo regions at each level; merging and optimizing each echo region; calculating the centroid of the convective cell based on the merged and optimized echo regions at each level; and extracting the convective cell contour from the first-level echo region based on the convective cell centroid to achieve convective cell identification. This invention effectively avoids the over-segmentation problem caused by noise points or isolated regions, making the convective cell identification results closer to the actual structure, improving the ability to distinguish multi-cell storms, and significantly improving the identification accuracy.

Description

Technical Field

[0001] This invention belongs to the field of weather recognition technology, and in particular relates to a method, device, equipment and storage medium for identifying convective cells based on radar network data. Background Technology

[0002] In the field of meteorological monitoring, accurate identification of convective cells is a crucial step in predicting severe convective weather and issuing disaster warnings. Traditional methods for identifying convective cells based on local maxima have the following significant drawbacks in practical applications:

[0003] (1) The problem of oversegmentation is serious. Due to the complex and variable nature of meteorological radar echo data and the large amount of noise interference, this method is easily affected by noise, which leads to the oversegmentation of convective cells. A large number of small and discrete reflectivity nuclei are identified, which makes the structure of convective cells fragmented and unable to fully present their actual shape and dynamic evolution process. This greatly increases the complexity of data processing and seriously interferes with meteorological analysts' judgment of the overall trend of the convective system.

[0004] (2) Inaccurate boundary judgment. This method also has the problem of identifying structures that are too large and not detailed enough in identifying multi-unit structures. When faced with multiple closely adjacent convective units, this method has difficulty in accurately distinguishing the boundaries of different units, and often mistakenly identifies multiple convective units as a large convective structure, resulting in overly coarse segmentation.

[0005] (3) It cannot express the structure of storm intensity. Due to the lack of an effective mechanism for judging the relationship between echo regions of different reflectivity factors, it cannot accurately distinguish the hierarchical relationship of echo regions of different reflectivity factors in the same convective cell, and it is also difficult to accurately define the outline boundary of the convective cell. This makes the convective cell identification results not intuitive and difficult to meet the needs of meteorological operations for accurate identification and rapid analysis of convective cells, which greatly limits the ability to provide efficient early warning and prevention of severe convective weather.

[0006] Therefore, a new method for identifying convective monomers is urgently needed to improve the accuracy and efficiency of identification. Summary of the Invention

[0007] The purpose of this invention is to provide a method, apparatus, device, and storage medium for identifying convective cells, in order to solve the problems of over-segmentation of convective cells, incomplete structure, lack of judgment on echo region relationship, and inaccurate contour definition in traditional methods.

[0008] This invention solves the above-mentioned technical problems through the following technical solution: a method for identifying convective monomers, comprising:

[0009] The maximum value is extracted from the three-dimensional radar network data along the height axis to obtain the two-dimensional combined echo;

[0010] The two-dimensional combined echo is subjected to multi-level progressive threshold segmentation to obtain echo regions at each level;

[0011] The echo regions at each level are merged and optimized separately;

[0012] Calculate the centroid of the convective unit based on the merged and optimized echo regions at each level;

[0013] Based on the centroid of the convective cell, the contour of the convective cell is extracted from the first-level echo region to achieve convective cell identification.

[0014] Furthermore, the two-dimensional combined echo is subjected to a three-level progressive threshold segmentation, the specific process of which includes:

[0015] Extract the portion of the two-dimensional combined echo with a reflectivity factor greater than or equal to the first-level threshold, and remove isolated points or invalid regions of this portion through morphological operations to obtain the first-level echo region;

[0016] Extract the portion of the first-level echo region whose reflectivity factor is greater than or equal to the second-level threshold, and remove isolated points or invalid regions of this portion through morphological operations to obtain the second-level echo region;

[0017] Extract the portion of the second-level echo region whose reflectivity factor is greater than or equal to the third-level threshold, and remove isolated points or invalid regions of this portion through morphological operations to obtain the third-level echo region;

[0018] Among them, the first-level threshold is less than the second-level threshold, which is less than the third-level threshold.

[0019] Furthermore, the echo regions at each level are merged and optimized separately, including:

[0020] For each echo region, calculate the minimum Euclidean distance between the contours of each connected domain within it;

[0021] If the minimum Euclidean distance is less than the set distance threshold, then morphological operations are used to merge the corresponding connected components.

[0022] Furthermore, the centroid of the convective unit is calculated based on the merged and optimized echo regions at each level, including:

[0023] Analyze the spatial inclusion relationships of echo regions at each level after merging and optimization;

[0024] The centroid of the last-order echo region in the computational spatial containment relation is used as the centroid of the convective unit.

[0025] Furthermore, the formula for calculating the centroid of the last-order echo region in the spatial inclusion relationship is as follows:

[0026] , ;

[0027] ;

[0028] in, This represents the centroid of the last-order echo region in a spatial containment relation; This represents the coordinates of the i-th pixel in the last-level echo region within the spatial containment relationship. This represents the weight of the i-th pixel in the last-level echo region of the spatial containment relation; represents the reflectivity factor of the i-th pixel in the last-level echo region of the spatial containment relation; n represents the number of pixels in the last-level echo region of the spatial containment relation; This represents the weighting function.

[0029] Furthermore, based on the centroid of the convective cell, the watershed algorithm is used to extract the contour of the convective cell from the first-order echo region. The specific implementation process is as follows:

[0030] The centroid of the convective unit is used as the seed point for the watershed algorithm;

[0031] Within the first-level echo region, the seed point is used as the starting point for automatic outward diffusion and growth until it reaches the boundary position, thus obtaining the convective monomer profile.

[0032] Furthermore, the identification method also includes defining a vertical projection area in the radar three-dimensional network data based on the contour of the convective unit;

[0033] Feature parameters are extracted from the vertical projection region, including storm top height, maximum reflectivity value, and effective volume.

[0034] Based on the same concept, the present invention also provides a convection monomer identification device, comprising:

[0035] The extraction unit is used to extract the maximum value from the radar three-dimensional network data along the height axis to obtain the two-dimensional combined echo;

[0036] The segmentation unit is used to perform multi-level progressive threshold segmentation on the two-dimensional combined echo to obtain echo regions at each level;

[0037] The merging unit is used to merge and optimize echo regions at each level separately.

[0038] The calculation unit is used to calculate the centroid of the convective unit based on the merged and optimized echo regions at all levels;

[0039] The identification unit is used to extract the contour of the convective cell from the first-level echo region based on the centroid of the convective cell, thereby realizing the identification of the convective cell.

[0040] Based on the same concept, the present invention also provides an electronic device, including a memory, a processor, and a computer program / instructions stored in the memory, wherein the processor executes the computer program / instructions to implement the convection single-unit identification method as described above.

[0041] Based on the same concept, the present invention also provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the convection single-unit identification method as described above.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] This invention employs a multi-level progressive threshold segmentation strategy, and merges and optimizes the echo regions at each level. Then, it calculates the centroid of the convective cell based on the merged and optimized echo regions at each level. This effectively avoids the over-segmentation problem caused by noise points or isolated regions, making the convective cell identification results closer to the actual structure, improving the ability to distinguish multi-cell storms, and significantly improving the identification accuracy. Attached Figure Description

[0044] To more clearly illustrate the technical solution 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 one embodiment of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a flowchart of the convective monomer identification method in an embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of the convective monomer identification results in an embodiment of the present invention. Detailed Implementation

[0047] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. 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.

[0048] 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.

[0049] Example 1

[0050] like Figure 1As shown, the convective monomer identification method provided in this embodiment of the invention includes the following steps:

[0051] Step 1: Extract the maximum value from the radar 3D network data along the height axis to obtain the 2D combined echo.

[0052] Multiple X-band phased array radars are deployed in the meteorological monitoring area. Meteorological echo data is collected in real time through these radars at set time intervals. The meteorological echo data from the polar coordinate system is converted to the Cartesian coordinate system, and the Cartesian coordinate system meteorological echo data from each X-band phased array radar is networked to obtain standardized three-dimensional radar network data. The three dimensions of the radar three-dimensional network data are altitude, latitude, and longitude. This networking achieves blind zone complementarity among the detection ranges of each X-band phased array radar, resulting in more comprehensive and detailed radar echo data.

[0053] X-band phased array radar possesses high temporal resolution and high spatial accuracy. Through refined scanning, it acquires denser and more continuous meteorological echo data, significantly enhancing its ability to capture the structure and evolution of convective cells. Utilizing the high-precision input data provided by X-band phased array radar, it is possible to identify storm development signs earlier and promptly capture minute-scale boundary changes, improving response speed and identification sensitivity during severe weather evolution and providing a more reliable data foundation for short-term warnings.

[0054] The maximum echo data is extracted along the height axis from the radar three-dimensional network data to highlight the characteristics of the convective cell on the horizontal plane, thus obtaining the two-dimensional combined echo.

[0055] Step 2: Perform multi-level progressive threshold segmentation on the two-dimensional combined echo to obtain echo regions at each level.

[0056] In a specific embodiment of the present invention, a three-level progressive threshold segmentation is performed on the two-dimensional combined echo, the specific process of which includes:

[0057] Step 2.1: Extract the portion of the two-dimensional combined echo with a reflectivity factor greater than or equal to the first-level threshold, and remove isolated points or invalid regions of this portion through morphological operations to obtain the first-level echo region;

[0058] Step 2.2: If there is a part in the first echo region with a reflectivity factor greater than or equal to the second threshold, then extract the part with a reflectivity factor greater than or equal to the second threshold from the first echo region, and remove isolated points or invalid regions of this part through morphological operations to obtain the second echo region;

[0059] Step 2.3: If there is a portion in the second echo region with a reflectivity factor greater than or equal to the third threshold, extract the portion with a reflectivity factor greater than or equal to the third threshold from the second echo region, and remove isolated points or invalid regions of this portion through morphological operations to obtain the third echo region.

[0060] The morphological operations in this embodiment include erosion and dilation operations. After each level of threshold segmentation, this invention uses morphological operations to remove areas smaller than a set area threshold (e.g., 15 km²). 2 Invalid areas (areas that are too small to develop into storms) are excluded, while valid echo areas are retained to provide an accurate data basis for subsequent analysis.

[0061] In this embodiment, the first-level threshold is set to TH, the second-level threshold is TH+10, and the third-level threshold is TH+15. The color mark interval in radar echo maps is typically 5 dBZ. It is generally believed in the industry that storms will occur above 40 dBZ, and some methods even filter echoes above 35 dBZ. This embodiment considers that echoes at 45 dBZ are more likely to develop into storms, thus determining the storm's periphery. Therefore, TH is set to 45 dBZ, the first-level threshold is 45 dBZ, the second-level threshold is 55 dBZ, and the third-level threshold is 60 dBZ. The 10 dBZ gradient between the second-level and first-level thresholds can locate the strong convection core. Only under the strong convection core can the storm be further segmented; the third-level threshold is used to reconfirm the strong convection core region.

[0062] By using multi-level progressive threshold segmentation to finely divide the individual structures, it is more in line with the judgment of weather forecasts and also facilitates the subsequent analysis of spatial inclusion relationships.

[0063] Step 3: Merge and optimize the echo regions at each level.

[0064] In a specific embodiment of the present invention, the merging and optimization of echo regions at each level includes:

[0065] Step 3.1: For each echo region, calculate the minimum Euclidean distance between the contours of each connected region within it;

[0066] Step 3.2: If the minimum Euclidean distance is less than the set distance threshold, then morphological operations are used to merge the corresponding connected components.

[0067] For each echo region, the spatial proximity of two connected regions is determined by the minimum Euclidean distance between their contours. When the minimum Euclidean distance is less than a set distance threshold (e.g., 14 pixels), it indicates that the two connected regions are spatially close and may belong to the same convective cell, or are in the early stages of structural separation and not yet completely independent. To maintain the continuity and integrity of structural identification, morphological opening and closing operations are used to process the two connected regions, connecting them to achieve merging optimization. This eliminates redundant boundaries caused by over-segmentation and ensures the integrity of the convective cell structure.

[0068] This invention considers the possibility of storm splitting during development. If there are two core regions that are very close together, the storm may be in a state of "potential splitting." To ensure the continuity of storm identification, if a storm initially shows signs of splitting but has not yet reached the required distance, it is not considered "divisible." Therefore, two connected regions whose distance has not reached a set threshold are merged for optimization.

[0069] This invention breaks through the limitations of traditional single-threshold segmentation. By combining multi-level progressive threshold segmentation with merging optimization, it can finely divide the echo regions of different reflectivity factors. With the help of contour distance calculation and morphological operations, it can accurately define the boundaries of closely adjacent convective cells, clarify the hierarchical relationship of echo regions of different reflectivity factors, and achieve fine identification and accurate segmentation of convective cells.

[0070] Step 4: Calculate the centroid of the convective unit based on the merged and optimized echo regions at each level.

[0071] In a specific embodiment of the present invention, the centroid of a convective unit is calculated based on the merged and optimized echo regions at each level, including:

[0072] Step 4.1: Analyze the spatial inclusion relationship of the echo regions at each level after merging and optimization.

[0073] Radar echoes are typically layered, with one layer containing another. For example, a 60 dBZ echo surrounds a 55-60 dBZ echo, which in turn surrounds a 50-55 dBZ echo, and so on. Therefore, by analyzing the spatial inclusion relationships of the merged and optimized echo regions at each level, we can determine the parent and child echo regions (or parent and child storm structures) and clarify the hierarchical relationships between echo regions with different reflectivity factors.

[0074] For example, if a first-level echo region completely contains a second-level echo region, then the first-level echo region is the parent echo region of the second-level echo region, and the second-level echo region is the child echo region of the first-level echo region; if a second-level echo region completely contains a third-level echo region, then the second-level echo region is the parent echo region of the third-level echo region, and the third-level echo region is the child echo region of the second-level echo region.

[0075] Step 4.2: Calculate the centroid of the last-level echo region in the spatial containment relationship, and use this centroid as the centroid of the convective unit.

[0076] Using the reflectivity factor of the pixels in the last echo region in the spatial inclusion relationship as the weight, the centroid of the last echo region is calculated by weighted average method, and it is used as the centroid of the corresponding convective cell. The centroid of the convective cell can accurately reflect the core position of the convective cell.

[0077] For example, if there exists a first-level echo region that completely contains the second-level echo region, and there does not exist a second-level echo region that completely contains the third-level echo region, then the second-level echo region is a sub-echo region of the first-level echo region. The second-level echo region is the last-level echo region in this spatial inclusion relationship, with the centroid of the second-level echo region as the centroid of the corresponding convective unit.

[0078] If there exists a first-level echo region that completely contains the second-level echo region, and a second-level echo region that completely contains the third-level echo region, then the third-level echo region is the last echo region in this spatial inclusion relationship, with the centroid of the third-level echo region as the centroid of the corresponding convective unit.

[0079] If there is no first-level echo region that completely contains the second-level echo region, and no second-level echo region that completely contains the third-level echo region, then the first-level echo region is the last-level echo region in this spatial inclusion relationship, with the centroid of the first-level echo region as the centroid of the corresponding convective unit.

[0080] In this embodiment, the formula for calculating the centroid of the last-level echo region in the spatial inclusion relationship is as follows:

[0081] , (1)

[0082] (2)

[0083] (3)

[0084] in, This represents the centroid of the last-order echo region in a spatial containment relation; This represents the coordinates of the i-th pixel in the last-level echo region within the spatial containment relationship. This represents the weight of the i-th pixel in the last-level echo region of the spatial containment relation; represents the reflectivity factor of the i-th pixel in the last-level echo region of the spatial containment relation; n represents the number of pixels in the last-level echo region of the spatial containment relation; The weighting function (3) is an empirical formula that can be adjusted according to the reflectivity factor.

[0085] Using the centroid of the last echo region in the spatial containment relationship as the centroid of the convective unit yields a more stable storm center. Compared to using the maximum value as the storm center, the calculation method of the convective unit centroid in this invention avoids the influence of outliers; as time goes by, the core region (i.e., the last echo region) changes relatively little and can more stably represent the storm center.

[0086] Step 5: Based on the centroid of the convective cell, extract the contour of the convective cell from the first-level echo region to achieve convective cell identification.

[0087] In a specific embodiment of this invention, a watershed algorithm is used to extract the convective cell profile from the first-level echo region based on the centroid of the convective cell. Specifically, the centroid of the convective cell is used as the seed point of the watershed algorithm. A mask is generated based on the two-dimensional combined echo and the first-level echo region. Within the effective area defined by the mask, the convective cell automatically diffuses outward from the seed point. As the diffusion process progresses, the watershed algorithm automatically identifies and captures the boundary formed by the change in reflectivity factor based on the two-dimensional combined echo. When the diffusion reaches the boundary position (e.g., 45 dBZ), the diffusion growth automatically stops, accurately determining the convective cell profile (i.e., the storm profile). This invention, through continuous iterative diffusion, completely outlines the convective cell profile, achieving accurate identification and boundary extraction of convective cells.

[0088] Figure 2 The results of the convective monomer identification are shown. Figure 2 In the diagram, the box portion has two distinct strong centroids. This invention accurately identifies this portion as two storm structures, demonstrating the effectiveness and accuracy of the method in identifying convective cells. In contrast, the single threshold segmentation rule identifies the box portion as a single storm structure.

[0089] Step 6: Feature calculation.

[0090] After completing the contour recognition of convective cells, this invention also defines a vertical projection area in the radar 3D network data based on the contour of the convective cells; and extracts feature parameters from the vertical projection area to support subsequent operational applications such as weather classification, evolution trend analysis and disaster level early warning of convective cells.

[0091] In this embodiment, the characteristic parameters include storm top height, maximum reflectivity value, high reflectivity core thickness, and effective volume.

[0092] Traditional methods, which segment storms using only the maximum value, fail to consider the rationality of multiple maximum values. This forced segmentation leads to inaccurate boundaries. This invention achieves more rigorous storm segmentation through multi-level progressive threshold segmentation, avoiding over-segmentation. Based on the results of multi-level progressive threshold segmentation, spatial inclusion relationships are analyzed, more clearly expressing the storm intensity structure. For example, secondary segmentation is only performed when the reflectivity factor difference reaches 10 dBZ or more. This avoids over-segmentation and also indicates the presence of a strong storm core, thus more clearly expressing the storm structure and better aligning with weather forecasters' judgments.

[0093] The convective cell identification results of this invention are visually closer to the subjective perception of manually delineating storm areas, which aligns with the cognitive preferences of operational forecasters for the "reasonable outline" of storms in their experience-based judgments. It has strong operational interpretability and practical value, significantly improving the intuitive acceptability and meteorological application adaptability of the algorithm results, and can meet the needs of meteorological operations for accurate identification and rapid analysis of convective cells.

[0094] Example 2

[0095] The convective cell identification device provided in this embodiment of the invention includes an extraction unit, a segmentation unit, a merging unit, a calculation unit, and an identification unit. The extraction unit is used to extract the maximum value from the radar three-dimensional network data along the height axis to obtain a two-dimensional combined echo. The segmentation unit is used to perform multi-level progressive threshold segmentation on the two-dimensional combined echo to obtain echo regions at each level. The merging unit is used to perform merging optimization on the echo regions at each level respectively. The calculation unit is used to calculate the centroid of the convective cell based on the merged and optimized echo regions at each level. The identification unit is used to extract the contour of the convective cell from the first-level echo region based on the centroid of the convective cell to realize convective cell identification.

[0096] In some specific embodiments of the present invention, the convective monomer identification device may incorporate the features of the convective monomer identification method in Embodiment 1 of the present invention, and vice versa.

[0097] Example 3

[0098] This invention also provides an electronic device, which includes a memory, a processor, and a computer program / instructions stored in the memory. The processor executes the computer program / instructions to implement the convection single-unit identification method in this invention.

[0099] Although not shown, the electronic device includes a processor that can perform various appropriate operations and processes based on programs and / or data stored in read-only memory (ROM) or loaded from a storage portion into random access memory (RAM). The processor can be a multi-core processor or may contain multiple processors. In some embodiments, the processor may include a general-purpose main processor and one or more specialized coprocessors, such as a central processing unit, graphics processing unit (GPU), neural network processor (NPU), digital signal processor (DSP), etc. Various programs and data required for device operation are also stored in RAM. The processor, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0100] The processor and memory described above are used together to execute programs / instructions stored in the memory. When the program / instructions are executed by the computer, they can implement the methods, steps, or functions described in the above embodiments.

[0101] Although not shown, embodiments of the present invention also provide a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the convection single-unit identification method of the present invention.

[0102] Readable storage media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient media, such as modulated data signals and carrier waves.

[0103] The above description only discloses specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or modifications that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for identifying convective monomers, characterized in that, The identification method includes: The maximum value is extracted from the three-dimensional radar network data along the height axis to obtain the two-dimensional combined echo; The two-dimensional combined echo is subjected to multi-level progressive threshold segmentation to obtain echo regions at each level; The echo regions at each level are merged and optimized separately; Calculate the centroid of the convective unit based on the last stage echo region after merging and optimization; Based on the centroid of the convective cell, the contour of the convective cell is extracted from the first-level echo region to achieve convective cell identification. The first-level echo region is obtained by extracting the portion of the two-dimensional combined echo with a reflectivity factor greater than or equal to the minimum threshold, and then removing isolated points or invalid regions of this portion through morphological operations.

2. The method for identifying convective monomers according to claim 1, characterized in that, The three-level progressive threshold segmentation of the two-dimensional combined echo includes the following specific steps: Extract the portion of the two-dimensional combined echo with a reflectivity factor greater than or equal to the first-level threshold, and remove isolated points or invalid regions of this portion through morphological operations to obtain the first-level echo region; Extract the portion of the first-level echo region whose reflectivity factor is greater than or equal to the second-level threshold, and remove isolated points or invalid regions of this portion through morphological operations to obtain the second-level echo region; Extract the portion of the second-level echo region whose reflectivity factor is greater than or equal to the third-level threshold, and remove isolated points or invalid regions of this portion through morphological operations to obtain the third-level echo region; Among them, the first-level threshold is less than the second-level threshold, which is less than the third-level threshold.

3. The method for identifying convective monomers according to claim 1, characterized in that, The echo regions at each level are merged and optimized separately, including: For each echo region, calculate the minimum Euclidean distance between the contours of each connected domain within it; If the minimum Euclidean distance is less than the set distance threshold, then morphological operations are used to merge the corresponding connected components.

4. The method for identifying convective monomers according to claim 1, characterized in that, The centroid of the convective unit is calculated based on the merged and optimized echo regions at each level, including: Analyze the spatial inclusion relationships of echo regions at each level after merging and optimization; The centroid of the last-order echo region in the computational spatial containment relation is used as the centroid of the convective unit.

5. The method for identifying convective monomers according to claim 4, characterized in that, The formula for calculating the centroid of the last-level echo region in the spatial containment relationship is as follows: ， ； ； in, This represents the centroid of the last-order echo region in a spatial containment relation; This represents the coordinates of the i-th pixel in the last-level echo region within the spatial containment relationship. This represents the weight of the i-th pixel in the last-level echo region of the spatial containment relation; represents the reflectivity factor of the i-th pixel in the last-level echo region of the spatial containment relation; n represents the number of pixels in the last-level echo region of the spatial containment relation; This represents the weighting function.

6. The method for identifying convective monomers according to claim 1, characterized in that, Based on the centroid of the convective cell, the watershed algorithm is used to extract the contour of the convective cell from the first-order echo region. The specific implementation process is as follows: The centroid of the convective unit is used as the seed point for the watershed algorithm; Within the first-level echo region, the seed point is used as the starting point for automatic outward diffusion and growth until it reaches the boundary position, thus obtaining the convective monomer profile.

7. The method for identifying convective monomers according to any one of claims 1 to 6, characterized in that, The identification method further includes defining a vertical projection area in the radar three-dimensional network data based on the convective single-unit profile; Feature parameters are extracted from the vertical projection region, including storm top height, maximum reflectivity value, and effective volume.

8. A convection monomer identification device, characterized in that, The identification device includes: The extraction unit is used to extract the maximum value from the radar three-dimensional network data along the height axis to obtain the two-dimensional combined echo; The segmentation unit is used to perform multi-level progressive threshold segmentation on the two-dimensional combined echo to obtain echo regions at each level; The merging unit is used to merge and optimize echo regions at each level separately. The calculation unit is used to calculate the centroid of the convective unit based on the last-stage echo region after merging and optimization. The identification unit is used to extract the contour of the convective cell from the first-level echo region based on the centroid of the convective cell, thereby realizing the identification of the convective cell; wherein, the first-level echo region is obtained by extracting the part with a reflectivity factor greater than or equal to the minimum threshold from the two-dimensional combined echo, and removing isolated points or invalid regions of this part through morphological operations.

9. An electronic device comprising a memory, a processor, and a computer program / instructions stored in the memory, characterized in that, The processor executes the computer program / instructions to implement the convection monomer identification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instruction is executed by the processor, it implements the convection single-unit identification method as described in any one of claims 1 to 7.

Images