Convection monomer identification method, apparatus and device, and storage medium

Abstract

The invention discloses a convection monomer identification method, device and equipment and a storage medium, and the identification method comprises the steps: extracting a maximum value from radar three-dimensional networking data along a height axis, and obtaining a two-dimensional combined echo; performing multi-stage progressive threshold segmentation on the two-dimensional combined echo to obtain each stage of echo region; performing combination optimization on each level of echo region; calculating the mass center of a convection single body according to each level of echo region after combination and optimization; and based on the mass center of the convective single body, extracting the contour of the convective single body from the first-stage echo region to realize convective single body identification. According to the method, the over-segmentation problem caused by noise points or isolated areas is effectively avoided, so that the convection monomer identification result is closer to an actual structure, the capability of distinguishing the multi-monomer storm is improved, and the identification accuracy is remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the field of weather recognition technology, and in particular relates to a method, device, equipment and storage medium for convective cell recognition based on radar networking data. Background Art

[0002] In the field of meteorological monitoring, accurate identification of convective cells is a key step in predicting severe convective weather and issuing disaster warnings. The traditional method of obtaining convective cells based on local maxima has the following obvious defects in practical applications:

[0003] (1) The over-segmentation problem is serious. Since meteorological radar echo data is complex and variable and has strong noise interference, this method is easily affected by noise, resulting in over-segmentation of convective cells. A large number of small and discrete reflectivity nuclei are identified, which makes the convective cell structure fragmented and unable to fully present its actual form 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 being detailed in identifying multi-unit structures. When faced with multiple closely adjacent convective units, this method has difficulty 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) Unable to express the storm intensity structure. Due to the lack of an effective judgment mechanism for the relationship between echo areas with different reflectivity factors, it is impossible to accurately distinguish the hierarchical relationship of echo areas with different reflectivity factors in the same convective cell, and it is also difficult to accurately define the contour boundary of the convective cell. This makes the convective cell identification results unintuitive and difficult to meet the needs of meteorological services for accurate identification and rapid analysis of convective cells, which greatly limits the efficient warning and prevention capabilities for 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 the present invention is to provide a method, device, equipment and storage medium for identifying convective monomers to solve the problems of over-segmentation of convective monomers, incomplete structure, lack of echo area relationship judgment and inaccurate contour definition in traditional methods.

[0008] The present invention solves the above technical problems through the following technical solutions: a convection monomer identification method, comprising:

[0009] Along the height axis, the maximum value is extracted from the radar 3D network data to obtain a 2D combined echo;

[0010] Performing multi-level progressive threshold segmentation on the two-dimensional combined echo to obtain echo regions at each level;

[0011] Merge and optimize echo areas at all levels respectively;

[0012] Calculate the centroid of the convective cell based on the merged and optimized echo areas at all levels;

[0013] Based on the centroid of the convective cell, the contour of the convective cell is extracted from the first-order echo area to realize convective cell identification.

[0014] Furthermore, a three-level progressive threshold segmentation is performed on the two-dimensional combined echo, and the specific process includes:

[0015] Extracting a portion of the two-dimensional combined echo having a reflectivity factor greater than or equal to a first-level threshold value, and removing isolated points or invalid areas of the portion through morphological operations to obtain a first-level echo region;

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

[0017] Extracting a portion of the second-level echo region whose reflectivity factor is greater than or equal to the third-level threshold value, and removing isolated points or invalid areas of the portion through morphological operations to obtain a third-level echo region;

[0018] Among them, the first level threshold < the second level threshold < the third level threshold.

[0019] Furthermore, each level of echo area is merged and optimized separately, including:

[0020] For each level of echo area, 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, the corresponding connected domains are merged using morphological operations.

[0022] Furthermore, the centroid of the convective cell is calculated based on the merged and optimized echo areas at all levels, including:

[0023] Analyze the spatial inclusion relationship of the echo areas at all levels after merging and optimization;

[0024] Calculate the centroid of the last-level echo area in the spatial inclusion relationship and use this centroid as the centroid of the convection cell.

[0025] Furthermore, the calculation formula for the centroid of the last-level echo area in the spatial inclusion relationship is:

[0026] , ;

[0027] ;

[0028] in, Represents the centroid of the last level echo area in the spatial inclusion relationship; Represents the coordinates of the i-th pixel point in the last level echo area in the spatial inclusion relationship; Represents the weight of the i-th pixel point in the last level echo area in the spatial inclusion relationship; It represents the reflectivity factor of the i-th pixel point in the last level echo area in the spatial inclusion relationship; n represents the number of pixels in the last level echo area in the spatial inclusion relationship; represents the weight function.

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

[0030] Using the centroid of the convective cell as a seed point for the watershed algorithm;

[0031] In the first-order echo region, the convection cell outline is obtained by automatically diffusing outwards with the seed point as the starting point until it reaches the boundary position.

[0032] Furthermore, the identification method further includes defining a vertical projection area in the radar three-dimensional networking data based on the convection cell profile;

[0033] Characteristic parameters are extracted from the vertical projection area, wherein the characteristic parameters include 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] An 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;

[0036] a segmentation unit, configured 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 the echo areas at each level;

[0038] A calculation unit, used for calculating the centroid of the convective cell based on the merged and optimized echo areas of each level;

[0039] The identification unit is used to extract the convective cell outline from the first-order echo area based on the convective cell centroid to realize convective cell identification.

[0040] Based on the same concept, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the above-mentioned convection cell identification method.

[0041] Based on the same concept, the present invention further provides a computer-readable storage medium having a computer program / instruction stored thereon, which implements the above-mentioned convection cell identification method when executed by a processor.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention adopts a multi-level progressive threshold segmentation strategy, merges and optimizes the echo areas at each level respectively, and then calculates the centroid of the convective cell based on the merged and optimized echo areas at each level, effectively avoiding the over-segmentation problem caused by noise points or isolated areas, making the convective cell identification results closer to the actual structure, improving the ability to distinguish multi-cell storms, and significantly improving the recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only one embodiment of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 This is a flow chart of a method for identifying a convection cell according to an embodiment of the present invention;

[0046] Figure 2 Schematic diagram of the convective monomer identification result in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following is a clear and complete description of the technical solutions of the present invention in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts are within the scope of protection of the present invention.

[0048] The following specific embodiments are used to describe the technical solution of the present invention in detail. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0049] Example 1

[0050] like Figure 1As shown, the convection monomer identification method provided by the embodiment of the present invention includes the following steps:

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

[0052] Multiple X-band phased array radars are deployed in the meteorological monitoring area. They collect weather echo data in real time at set intervals. This polar coordinate data is converted to a Cartesian coordinate system. The Cartesian echo data from each X-band phased array radar is then networked to generate standardized three-dimensional radar network data. The three dimensions of this data are altitude, latitude, and longitude. This networking allows for complementary blind spots within the detection ranges of each X-band phased array radar, resulting in more comprehensive and detailed radar echo data.

[0053] X-band phased array radars offer high temporal resolution and spatial precision. Through refined scanning, they acquire denser and more continuous meteorological echo data, significantly enhancing the ability to capture the structure and evolution of convective cells. Leveraging the high-precision input data provided by X-band phased array radars, early signs of storm development can be identified, enabling timely capture of micro-scale boundary changes. This improves response speed and recognition sensitivity during severe weather development, providing a more reliable data foundation for short-term warnings.

[0054] The maximum echo data is extracted along the height axis from the three-dimensional radar network data to highlight the characteristics of the convection cell on the horizontal plane, thus obtaining a 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, and the specific process includes:

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

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

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

[0060] The morphological operations in this embodiment include erosion and dilation operations. After each level of threshold segmentation, the present invention uses morphological operations to remove areas smaller than the set area threshold (e.g., 15km 2 , areas that are too small will not develop into storms) and retain valid echo areas to provide an accurate data basis for subsequent analysis.

[0061] In this embodiment, TH is set as the first-level threshold, then the second-level threshold is TH+10, and the third-level threshold is TH+15. The color scale interval in the radar echo map is usually 5dBZ. The industry generally believes that storms will appear above 40dBZ, and some methods even filter echoes above 35dBZ. This embodiment takes into account that echoes of 45dBZ are more likely to develop into storms, thereby determining the periphery of the storm. Therefore, TH is set to 45dBZ, then the first-level threshold is 45dBZ, the second-level threshold is 55dBZ, and the third-level threshold is 60dBZ. There is a gradient change of 10dBZ between the second-level threshold and the first-level threshold, which can locate the strong convective core. Only under the strong convective core can the storm be divided twice; the strong convective core area is reconfirmed by the third-level threshold.

[0062] The monomer structure is finely divided through multi-level progressive threshold segmentation, which is more consistent with the judgment of weather forecast and 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, merging and optimizing the echo regions at each level includes:

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

[0066] Step 3.2: If the minimum Euclidean distance is less than the set distance threshold, the corresponding connected domains are merged using morphological operations.

[0067] For each echo level, the spatial proximity of two connected domains 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), the two connected domains are spatially close and may belong to the same convection cell or be in the early stages of structural separation and not yet fully independent. To maintain the continuity and integrity of structural identification, the two connected domains are processed through morphological opening and closing operations to connect them, thus achieving merge optimization, eliminating redundant boundaries caused by over-segmentation, and ensuring the integrity of the convection cell structure.

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

[0069] The present invention breaks through the limitations of traditional single-threshold segmentation. Through the combination of multi-level progressive threshold segmentation and merge optimization, it can finely divide the echo areas 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 areas of different reflectivity factors, and realize the fine identification and accurate segmentation of convective cells.

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

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

[0072] Step 4.1: Analyze the spatial inclusion relationship of the merged and optimized echo regions at all levels.

[0073] Radar echoes typically appear layered within layer. For example, a 60dBZ echo surrounds a 55-60dBZ echo, a 55-60dBZ echo surrounds a 50-55dBZ echo, and so on. Therefore, by analyzing the spatial inclusion relationships among the merged and optimized echo regions, 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] Exemplarily, if there is a first-level echo region that 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 there is a second-level echo region that 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 area in the spatial inclusion relationship and use this centroid as the centroid of the convection cell.

[0076] The reflectivity factor of the pixel points in the last-level echo area in the spatial inclusion relationship is used as the weight, and the weighted average method is used to calculate the centroid of the last-level echo area, 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 is a first-level echo region that completely contains the second-level echo region, and there is no second-level echo region that completely contains the third-level echo region, then the second-level echo region is a sub-level echo region of the first-level echo region, and the second-level echo region is the last-level echo region in this spatial inclusion relationship, and the centroid of the second-level echo region is the centroid of the corresponding coherent flow monomer;

[0078] If there is a first-level echo region that completely contains the second-level echo region, and the second-level echo region completely contains the third-level echo region, then the third-level echo region is the last-level echo region in this spatial inclusion relationship, and the centroid of the third-level echo region is the centroid of the corresponding flow monomer;

[0079] If there is no first-order echo region that completely contains the second-order echo region, and there is no second-order echo region that completely contains the third-order echo region, then the first-order echo region is the last-order echo region in this spatial inclusion relationship, and the centroid of the first-order echo region is the centroid of the corresponding flow monomer.

[0080] In this embodiment, the calculation formula for the centroid of the last-level echo area in the spatial inclusion relationship is:

[0081] , (1)

[0082] (2)

[0083] (3)

[0084] in, Represents the centroid of the last level echo area in the spatial inclusion relationship; Represents the coordinates of the i-th pixel point in the last level echo area in the spatial inclusion relationship; Represents the weight of the i-th pixel point in the last level echo area in the spatial inclusion relationship; It represents the reflectivity factor of the i-th pixel point in the last level echo area in the spatial inclusion relationship; n represents the number of pixels in the last level echo area in the spatial inclusion relationship; Represents the weight function. The weight function (3) is an empirical formula that can adjust the corresponding weight value according to the reflectivity factor.

[0085] Using the centroid of the last-order echo region in the spatial inclusion relationship as the centroid of the convective cell yields a more stable storm center. Compared to using the maximum value as the storm center, this method avoids the influence of outliers. Over time, the core region (i.e., the last-order echo region) changes relatively little, providing a more stable representation of the storm center.

[0086] Step 5: Based on the centroid of the convective cell, the convective cell outline is extracted from the first-order echo area to achieve convective cell identification.

[0087] In a specific embodiment of the present invention, a watershed algorithm is used to extract the contour of the convective cell from the first-order echo area 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, and a mask is generated according to the first-order echo area based on the two-dimensional combined echo. Within the effective area defined by the mask, the algorithm automatically diffuses outward with the seed point as the starting 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 it expands to the boundary position (for example, 45dBZ), the diffusion growth is automatically stopped, and the contour of the convective cell (i.e., the storm contour) is accurately determined. The present invention completely outlines the contour of the convective cell through continuous iterative diffusion, thereby achieving accurate identification and boundary extraction of the convective cell.

[0088] Figure 2 The results of convective cell identification are shown. Figure 2 In the figure, the box part has two obvious strong centroids. The present invention accurately identifies this part as two storm structures, which shows the effectiveness and accuracy of the method of the present invention in identifying convective cells; while the single threshold segmentation rule will identify the box part as one storm structure.

[0089] Step 6: Feature calculation.

[0090] After completing the identification of the convective cell outline, the present invention also defines the vertical projection area in the radar three-dimensional networking data based on the convective cell outline; extracts characteristic parameters from the vertical projection area to support subsequent convective cell weather classification, evolution trend analysis, disaster level warning and other business applications.

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

[0092] Traditional methods use only the maximum value for segmentation, without considering the rationality of multiple maxima. This forced segmentation leads to inaccurate boundaries. This invention achieves stricter storm segmentation through multi-level progressive threshold segmentation, avoiding the problem of over-segmentation. Based on the results of multi-level progressive threshold segmentation, spatial inclusion relationships are analyzed to more clearly express the storm intensity structure. For example, secondary segmentation is only performed when the reflectivity factor reaches a difference of 10dBZ or more. This not only avoids over-segmentation but also indicates the presence of a strong storm core, allowing for a clearer representation of the storm structure and better aligning with the weather forecaster's judgment.

[0093] The visual presentation of the convective cell identification results of the present invention is closer to the subjective perception of the artificially delineated storm area, which is in line with the cognitive preference of business forecasters for the "reasonable outline" of storms in their empirical judgment. It has strong business interpretability and practical value, significantly improves the intuitive acceptability and meteorological application adaptability of the algorithm results, and can meet the needs of meteorological business for accurate identification and rapid analysis of convective cells.

[0094] Example 2

[0095] The convective cell identification device provided in an embodiment of the present 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 networking 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 areas at each level; the merging unit is used to merge and optimize the echo areas at each level respectively; the calculation unit is used to calculate the convective cell centroid based on the merged and optimized echo areas at each level; the identification unit is used to extract the convective cell contour from the first-level echo area based on the convective cell centroid to realize convective cell identification.

[0096] In some specific embodiments of the present invention, the convective cell identification device may be combined with the features of the convective cell identification method in the first embodiment of the present invention, and vice versa.

[0097] Example 3

[0098] An embodiment of the present invention further provides an electronic device, comprising: a memory, a processor, and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the convection cell identification method in the embodiment of the present 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 a read-only memory (ROM) or programs and / or data loaded from a storage portion into a random access memory (RAM). The processor can be a multi-core processor or can include multiple processors. In some embodiments, the processor can include a general-purpose main processor and one or more special coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), etc. Various programs and data required for device operation are also stored in the RAM. The processor, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

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

[0101] Although not shown, an embodiment of the present invention further provides a computer-readable storage medium having a computer program / instruction stored thereon. When the computer program / instruction is executed by a processor, the convection cell identification method in the embodiment of the present invention is implemented.

[0102] Computer-readable storage media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The 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 technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0103] The above disclosure is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or modifications within the technical scope disclosed in the present invention, and they should all be covered by the scope of protection of the present invention.

Claims

1. A method for identifying convection monomers, characterized in that: The identification method comprises: Along the height axis, the maximum value is extracted from the radar 3D network data to obtain a 2D combined echo; Performing multi-level progressive threshold segmentation on the two-dimensional combined echo to obtain echo regions at each level; Merge and optimize echo areas at all levels respectively; Calculate the centroid of the convective cell based on the merged and optimized echo areas at all levels; Based on the centroid of the convective cell, the contour of the convective cell is extracted from the first-order echo area to realize convective cell identification.

2. The method for identifying convection cells according to claim 1, wherein: The three-level progressive threshold segmentation is performed on the two-dimensional combined echo, and the specific process includes: Extracting a portion of the two-dimensional combined echo having a reflectivity factor greater than or equal to a first-level threshold value, and removing isolated points or invalid areas of the portion through morphological operations to obtain a first-level echo region; Extracting a portion of the first-level echo region whose reflectivity factor is greater than or equal to the second-level threshold from the first-level echo region, and removing isolated points or invalid regions of the portion through morphological operations to obtain a second-level echo region; Extracting a portion of the second-level echo region whose reflectivity factor is greater than or equal to the third-level threshold value, and removing isolated points or invalid areas of the portion through morphological operations to obtain a third-level echo region; Among them, the first level threshold < the second level threshold < the third level threshold.

3. The method for identifying convection cells according to claim 1, wherein: Merge and optimize the echo areas at each level separately, including: For each level of echo area, 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, the corresponding connected domains are merged using morphological operations.

4. The method for identifying convection cells according to claim 1, wherein: The centroid of the convective cell is calculated based on the merged and optimized echo areas at all levels, including: Analyze the spatial inclusion relationship of the echo areas at all levels after merging and optimization; Calculate the centroid of the last-level echo area in the spatial inclusion relationship and use this centroid as the centroid of the convection cell.

5. The method for identifying convection cells according to claim 4, characterized in that: The calculation formula for the centroid of the last-level echo area in the spatial inclusion relationship is: ， ； ； in, Represents the centroid of the last level echo area in the spatial inclusion relationship; Represents the coordinates of the i-th pixel point in the last level echo area in the spatial inclusion relationship; Represents the weight of the i-th pixel point in the last level echo area in the spatial inclusion relationship; It represents the reflectivity factor of the i-th pixel point in the last level echo area in the spatial inclusion relationship; n represents the number of pixels in the last level echo area in the spatial inclusion relationship; represents the weight function.

6. The method for identifying convection cells according to claim 1, wherein: Based on the centroid of the convective cell, the watershed algorithm is used to extract the convective cell outline from the first-order echo area. The specific implementation process is as follows: Using the centroid of the convective cell as a seed point for the watershed algorithm; In the first-order echo region, the convection cell outline is obtained by automatically diffusing outwards with the seed point as the starting point until it reaches the boundary position.

7. The method for identifying convection 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 networking data based on the convective cell profile; Characteristic parameters are extracted from the vertical projection area, wherein the characteristic parameters include storm top height, maximum reflectivity value, and effective volume.

8. A convection monomer identification device, characterized in that: The identification device comprises: An 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; a segmentation unit, configured 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 the echo areas at each level; A calculation unit, used for calculating the centroid of the convective cell based on the merged and optimized echo areas of each level; The identification unit is used to extract the convective cell outline from the first-order echo area based on the convective cell centroid to realize convective cell identification.

9. An electronic device comprising a memory, a processor, and a computer program / instruction stored in the memory, characterized in that: The processor executes the computer program / instruction to implement the convective cell identification method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the convection cell identification method according to any one of claims 1 to 7 is implemented.

Images