A cloud type identification method and device based on ground radar, medium and product
Patent Information
- Application Number
- CN202511875099.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-12
- Publication Date
- 2026-10-16
AI Technical Summary
飞机采样是最为准确的云微物理特征量的观测手段，但观测成本较高，无法开展长期连续性观测
本申请公开了一种基于地面雷达的云类型识别方法、装置、介质及产品，首先，获取地面雷达的当前时段的多个时刻的雷达廓线；雷达廓线包括多个雷达观测点的雷达观测数据，雷达观测数据包括反射率因子、云特征标识、高度和温度；其次，对当前时段的多个时刻的雷达廓线进行处理，得到当前时段的多个时刻的处理后的雷达廓线；然后，基于当前时段的多个时刻的处理后的雷达廓线，构建云雷达数据矩阵；再次，基于云雷达数据矩阵，确定当前时段中的云的云物理特征；云物理特征包括：云顶高度、云底高度、云顶温度、云底温度、云层厚度和反射率因子的统计特征；统计特征包括均值、众数、中位数和95%区间范围；最后，将当前时段中的云的云物理特征输入至云类型识别模型中，得到当前时段中的云类型的识别值；云类型识别模型是对卷积神经网络进行训练得到的；云类型包括：卷云、积云、层积云、雨层云、高层云、高积云和深对流云。本申请利用对卷积神经网络进行训练得到的云类型识别模型和雷达廓线对云类型进行识别，实现了利用单一雷达观测数据，无需其他观测资料支撑的云类型识别。
Smart Images
Figure CN122925995A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of cloud type identification technology based on ground radar, and in particular to a cloud type identification method, device, medium and product based on ground radar. Background Technology
[0002] Clouds are one of the most common and important factors in the Earth-atmosphere system, covering approximately 70% of the Earth's surface. Clouds not only play a major role in regulating regional and global radiative energy balances but also significantly influence the water cycle within the Earth-atmosphere system, having a significant impact on global atmospheric circulation and climate change. However, due to the dramatic spatiotemporal variations of clouds and the complexity of their response to climate change, even small changes in cloud physical parameters can alter the radiative energy balance, generating significant feedback effects on regional and global climate change. This makes clouds one of the most uncertain factors in climate simulation and prediction.
[0003] Cloud observation and research primarily utilizes active and passive remote sensing via ground-based, aircraft, and satellite methods. Satellites offer unparalleled advantages in cloud observation, being the only means to provide long-term, continuous spatiotemporal distribution data of clouds over global or large-scale regions. Aircraft sampling is the most accurate method for observing cloud microphysical characteristics, but its high cost prevents long-term continuous observation. Compared to satellite and aircraft observations, ground-based radar offers high temporal resolution, particularly millimeter-wave cloud radar with excellent cloud penetration, providing vertical cloud profile information. However, it lacks corresponding cloud types and cannot identify cloud types. Therefore, a ground-based radar-based cloud type identification method is needed. Summary of the Invention
[0004] The purpose of this application is to provide a method, device, medium, and product for cloud type identification based on ground radar, so as to realize cloud type identification based on ground radar.
[0005] To achieve the above objectives, this application provides the following solution.
[0006] Firstly, this application provides a cloud type identification method based on ground radar, including: The radar profiles of the ground radar at multiple times during the current period are obtained; the radar profiles include radar observation data from multiple radar observation points, and the radar observation data includes reflectivity factor, cloud feature identifiers, altitude, and temperature; The radar profiles at multiple moments in the current time period are processed to obtain the processed radar profiles at multiple moments in the current time period. A cloud radar data matrix is constructed based on the processed radar profiles from multiple moments in the current time period. Based on the cloud radar data matrix, the cloud physical characteristics of the clouds in the current time period are determined; the cloud physical characteristics include: statistical characteristics of cloud top height, cloud bottom height, cloud top temperature, cloud bottom temperature, cloud thickness, and reflectivity factor; the statistical characteristics include mean, mode, median, and 95% interval range. The cloud physical characteristics of the clouds in the current time period are input into the cloud type identification model to obtain the identification value of the cloud type in the current time period; the cloud type identification model is obtained by training a convolutional neural network; the cloud types include: cirrus clouds, cumulus clouds, stratocumulus clouds, nimbostratus clouds, altostratus clouds, altocumulus clouds and deep convection clouds.
[0007] In one embodiment, radar profiles at multiple times within the current time period are processed to obtain processed radar profiles at multiple times within the current time period, including: The radar profiles with a first proportion greater than 35% in the current time period are updated to missing values to obtain processed radar profiles for multiple moments in the current time period; the first proportion is the ratio of the number of radar observation data with a reflectivity factor greater than -10dBZ at temperatures above 0℃ to the total number of radar observation data in the radar profile.
[0008] In one embodiment, each row of the cloud radar data matrix corresponds to a time point, and each column of the cloud radar data matrix corresponds to a processed radar profile; each element of the cloud radar data matrix corresponds to a radar observation data point in the processed radar profile at a given time point. Based on the cloud radar data matrix, the cloud physical characteristics of the clouds in the current time period are determined, including: Based on the cloud radar data matrix, the edge determination results of the cloud in the current time period are determined; the edge determination results include radar observation data corresponding to the vertical edges and radar observation data corresponding to the horizontal edges; Based on the edge determination results of the clouds in the current time period, determine the cloud physical characteristics of the clouds in the current time period.
[0009] In one embodiment, determining the edge determination result of the cloud in the current time period based on the cloud radar data matrix includes: Any radar observation data with a cloud feature identifier of 1 in the cloud radar data matrix is identified as the target radar observation data; When there are more than 3 consecutive radar observation data points in the column element where the target radar observation data is located that can be represented as 0, the target radar observation data is determined to be the radar observation data corresponding to the upper and lower edges of the cloud. When the cloud feature identifier of the processed radar profile at each moment of a consecutive minute in the row element containing the target radar observation data is 0, the target radar observation data is determined to be the radar observation data corresponding to the left and right edges of the cloud.
[0010] In one embodiment, the process of determining the cloud type identification model includes: Construct a training set; the training set includes cloud physical characteristics at multiple historical moments in multiple historical periods and actual values of cloud types in each historical period; Constructing a convolutional neural network; By taking the cloud physical characteristics at multiple moments in each historical period as input and the actual value of the cloud type in the corresponding historical period as output, a convolutional neural network is trained to obtain a cloud type recognition model.
[0011] In one implementation, constructing a training set includes: The system acquires multiple satellite profiles within a 200-kilometer radius centered on the latitude and longitude of the ground radar location. These satellite profiles include: satellite observation time, melt layer height, and satellite observation data from multiple satellite observation points along the satellite profile. The satellite observation data includes reflectivity factor, cloud feature identifier, altitude, gas attenuation correction data, and cloud type. Define any satellite scan trajectory as the current satellite scan trajectory; Each satellite profile in the current satellite scan trajectory is preprocessed to obtain multiple preprocessed satellite profiles corresponding to the current satellite scan trajectory; The mean time corresponding to the current satellite scan trajectory is obtained by averaging the satellite observation times of all satellite profiles in the current satellite scan trajectory. The current historical period is defined as one hour before and after the mean time corresponding to the current satellite scanning trajectory. The radar profiles of each historical time in the current historical period are obtained, and the radar profile of any historical time is defined as the current radar profile. Update the radar observation data with cloud feature identifier of 0 in the current radar profile to the missing value to obtain the current radar profile after one processing step. The average value of the gas attenuation correction data in all preprocessed satellite profiles corresponding to the current satellite scan trajectory is calculated, and the difference is calculated based on the number of radar observation points in the current radar profile to obtain the gas attenuation correction data profile corresponding to the current satellite scan trajectory. Based on the gas attenuation correction data profile corresponding to the current satellite scan trajectory, the reflectivity factors in the current radar profile after the first processing are corrected to obtain the current radar profile after the second processing. The first judgment result is obtained by determining whether the second proportion of the current radar profile after secondary processing is greater than 35%. The second proportion is the ratio of the number of radar observation data with reflectivity factor greater than -10dBZ whose height is below the average of the melting layer height of all preprocessed satellite profiles corresponding to the current satellite scanning trajectory to the number of radar observation data with cloud feature identifier of 1 in the current radar profile after secondary processing. If the first judgment result is yes, then the current radar profile after the second processing is updated to the missing value, and the current radar profile after the third processing is obtained. If the first judgment result is negative, then the current radar profile after the second processing is determined as the current radar profile after the third processing. Linearly interpolate the cloud types in all preprocessed satellite profiles corresponding to the current satellite scan trajectory to the current radar profile after three-stage processing to obtain the current radar profile with cloud types. Based on all radar profiles with cloud types corresponding to all satellite scan trajectories, a cloud type matrix is constructed; each row of the cloud type matrix corresponds to a historical moment, and each column of the cloud type matrix corresponds to the current radar profile with cloud type; each element of the cloud type matrix corresponds to a radar observation data and cloud type of the current radar profile with cloud type at a historical moment. Based on the cloud type matrix, the cloud physical characteristics at multiple historical moments in multiple historical periods and the actual values of cloud types in each historical period are determined, thereby obtaining the training set.
[0012] In one embodiment, each satellite profile in the current satellite scan trajectory is preprocessed to obtain multiple preprocessed satellite profiles corresponding to the current satellite scan trajectory, including: Define any satellite profile in the current satellite scan trajectory as the current satellite profile; Update the satellite observation data with cloud feature identifier of 0 in the current satellite profile to the missing value to obtain the current satellite profile after one processing step; The reflectivity factor of the corresponding satellite observation point is corrected by using the gas attenuation correction data of each satellite observation point in the current satellite profile after the first processing, and the current satellite profile after the second processing is obtained. The second judgment result is obtained by determining whether the third proportion of the current satellite profile after secondary processing is greater than 35%. The third proportion is the ratio of the number of satellite observation data with a reflectivity factor greater than -10dBZ that are below the melting layer height of the current satellite profile to the number of satellite observation data with a cloud feature identifier of 1 in the current satellite profile after secondary processing. If the second judgment result is yes, then the current satellite profile after secondary processing is updated to the missing value to obtain the preprocessed current satellite profile; If the second judgment result is negative, then the current satellite profile after secondary processing will be determined as the current satellite profile after preprocessing.
[0013] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for identifying cloud types based on ground radar.
[0014] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned cloud type identification method based on ground radar.
[0015] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned cloud type identification method based on ground radar.
[0016] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application discloses a method, device, medium, and product for cloud type identification based on ground radar. First, it acquires radar profiles at multiple times within the current time period from the ground radar. The radar profiles include radar observation data from multiple radar observation points, and the radar observation data includes reflectivity factor, cloud feature identifiers, altitude, and temperature. Second, it processes the radar profiles at multiple times within the current time period to obtain processed radar profiles at multiple times within the current time period. Then, based on the processed radar profiles at multiple times within the current time period, it constructs a cloud radar data matrix. Third, it uses cloud radar... A data matrix is used to determine the cloud physical characteristics of clouds in the current time period. These characteristics include statistical features of cloud top height, cloud base height, cloud top temperature, cloud base temperature, cloud thickness, and reflectivity factor. The statistical features include the mean, mode, median, and 95% percentile range. Finally, the cloud physical characteristics of clouds in the current time period are input into a cloud type identification model to obtain the cloud type identification value for that period. The cloud type identification model is trained using a convolutional neural network. Cloud types include cirrus, cumulus, stratocumulus, nimbostratus, altostratus, altocumulus, and deep convective clouds. This application utilizes a cloud type identification model trained on a convolutional neural network and radar profiles to identify cloud types, achieving cloud type identification using only single radar observation data without the need for other observational data. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 A schematic flowchart of a cloud type identification method based on ground radar provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation
[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0020] The purpose of this application is to provide a method, device, medium, and product for cloud type identification based on ground radar, aiming to achieve cloud type identification based on ground radar.
[0021] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0022] In one exemplary embodiment, such as Figure 1 As shown, a cloud type identification method based on ground radar is provided, which includes the following steps.
[0023] Step 1: Obtain radar profiles at multiple times during the current period from the ground radar.
[0024] The radar profile includes radar observation data from multiple radar observation points, which includes reflectivity factor, cloud feature identifiers, altitude, and temperature.
[0025] Step 2: Process the radar profiles at multiple times during the current time period to obtain the processed radar profiles at multiple times during the current time period.
[0026] As an optional implementation, step 2 includes: The radar profiles with a first proportion greater than 35% in the current time period are updated to missing values to obtain processed radar profiles for multiple moments in the current time period; the first proportion is the ratio of the number of radar observation data with a reflectivity factor greater than -10dBZ at temperatures above 0℃ to the total number of radar observation data in the radar profile.
[0027] Step 3: Construct a cloud radar data matrix based on the processed radar profiles from multiple moments in the current time period.
[0028] Step 4: Based on the cloud radar data matrix, determine the cloud physical characteristics of the clouds in the current time period.
[0029] Among them, cloud physical characteristics include: statistical characteristics of cloud top height, cloud bottom height, cloud top temperature, cloud bottom temperature, cloud thickness, and reflectivity factor; statistical characteristics include mean, mode, median, and 95% interval range.
[0030] As an optional implementation, each row of the cloud radar data matrix corresponds to a time point, and each column of the cloud radar data matrix corresponds to a processed radar profile; each element of the cloud radar data matrix corresponds to a radar observation data point in the processed radar profile at a given time point. Step 4 includes: Step 41: Based on the cloud radar data matrix, determine the edge determination results of the cloud in the current time period; the edge determination results include radar observation data corresponding to the vertical edges and radar observation data corresponding to the horizontal edges; Step 42: Determine the cloud physical characteristics of the clouds in the current time period based on the edge determination results of the clouds in the current time period.
[0031] As an optional implementation, step 41 includes: Step 411: Identify any radar observation data in the cloud radar data matrix whose cloud feature is identified as 1 as the target radar observation data; Step 412: When there are more than 3 consecutive radar observation data points in the column element where the target radar observation data is located that can be represented as 0, the target radar observation data is determined to be the radar observation data corresponding to the upper and lower edges of the cloud. Step 413: When the cloud feature identifier of the processed radar profile at each moment of a consecutive minute in the row element where the target radar observation data is located is 0, the target radar observation data is determined to be the radar observation data corresponding to the left and right edges of the cloud.
[0032] Step 5: Input the cloud physical characteristics of the clouds in the current time period into the cloud type recognition model to obtain the cloud type recognition value in the current time period.
[0033] The cloud type identification model is obtained by training a convolutional neural network; cloud types include: cirrus, cumulus, stratocumulus, nimbostratus, altostratus, altocumulus, and deep convection clouds.
[0034] As an optional implementation, step 5, the process of determining the cloud type identification model, includes the following steps.
[0035] Step 51: Construct a training set; the training set includes cloud physical characteristics at multiple historical moments in multiple historical periods and actual values of cloud types in each historical period.
[0036] As an optional implementation, step 51 includes the following steps.
[0037] Step 5101: Acquire multiple satellite profiles in multiple satellite scan trajectories within a 200-kilometer radius centered on the latitude and longitude of the ground radar location; the satellite profiles include: satellite observation time, melting layer height, and satellite observation data of multiple satellite observation points on the satellite profiles, including reflectivity factor, cloud feature identifier, altitude, gas attenuation correction data, and cloud type.
[0038] Step 5102: Determine any satellite scan trajectory as the current satellite scan trajectory.
[0039] Step 5103: Preprocess each satellite profile in the current satellite scan trajectory to obtain multiple preprocessed satellite profiles corresponding to the current satellite scan trajectory.
[0040] As an optional implementation, step 5103 includes: Step 51031: Determine any satellite profile in the current satellite scan trajectory as the current satellite profile; Step 51032: Update the satellite observation data with cloud feature identifier of 0 in the current satellite profile to the missing value to obtain the current satellite profile after one processing step; Step 51033: Correct the reflectivity factor of the corresponding satellite observation point using the gas attenuation correction data of each satellite observation point in the current satellite profile after the first processing, and obtain the current satellite profile after the second processing; Step 51034: Determine whether the third proportion of the current satellite profile after secondary processing is greater than 35% to obtain the second judgment result; the third proportion is the ratio of the number of satellite observation data with a reflectivity factor greater than -10dBZ that is below the melting layer height of the current satellite profile to the number of satellite observation data with a cloud feature identifier of 1 in the current satellite profile after secondary processing. Step 51035: If the second judgment result is yes, then update the current satellite profile after secondary processing to the missing value to obtain the preprocessed current satellite profile; Step 51036: If the second judgment result is negative, then the current satellite profile after secondary processing is determined as the current satellite profile after preprocessing.
[0041] Step 5104: Calculate the average satellite observation time for all satellite profiles in the current satellite scan trajectory to obtain the average time corresponding to the current satellite scan trajectory.
[0042] Step 5105: Determine the current historical time period as one hour before and after the mean time corresponding to the current satellite scanning trajectory, obtain the radar profile of each historical time in the current historical time period, and determine the radar profile of any historical time as the current radar profile.
[0043] Step 5106: Update the radar observation data with cloud feature identifier of 0 in the current radar profile to the missing value to obtain the current radar profile after one processing step.
[0044] Step 5107: Calculate the mean of the gas attenuation correction data in all preprocessed satellite profiles corresponding to the current satellite scan trajectory, and calculate the difference based on the number of radar observation points in the current radar profile to obtain the gas attenuation correction data profile corresponding to the current satellite scan trajectory.
[0045] Step 5108: Based on the gas attenuation correction data profile corresponding to the current satellite scan trajectory, correct each reflectivity factor in the current radar profile after the first processing to obtain the current radar profile after the second processing.
[0046] Specifically, since the satellite's transmitted echo and the radar's transmitted echo are in opposite directions, the gas attenuation correction data profile needs to be converted using the following formula during correction: the satellite's transmitted echo frequency is 94 GHz, and the radar's transmitted echo frequency is 35 GHz. ; in, Correct the gas attenuation data profile after conversion; For gas attenuation correction data profile; This is a constant related to the geographical location of the cloud radar, with a value of 3.36. Step 5109: Determine whether the second proportion of the current radar profile after secondary processing is greater than 35%, and obtain the first judgment result; the second proportion is the ratio of the number of radar observation data with a reflectivity factor greater than -10dBZ whose height is below the average of the melting layer height of all preprocessed satellite profiles corresponding to the current satellite scanning trajectory to the number of radar observation data with cloud feature identifier of 1 in the current radar profile after secondary processing.
[0047] Step 5110: If the first judgment result is yes, then update the current radar profile after the second processing to the missing value to obtain the current radar profile after the third processing.
[0048] Step 5111: If the first judgment result is negative, then the current radar profile after the second processing is determined as the current radar profile after the third processing.
[0049] Step 5112: Linearly interpolate the cloud types in all preprocessed satellite profiles corresponding to the current satellite scan trajectory to the current radar profile after three processing steps to obtain the current radar profile with cloud types.
[0050] Step 5113: Construct a cloud type matrix based on all radar profiles with cloud types corresponding to all satellite scan trajectories; each row of the cloud type matrix corresponds to a historical moment, and each column of the cloud type matrix corresponds to the current radar profile with cloud type; each element of the cloud type matrix corresponds to a radar observation data and cloud type in the current radar profile with cloud type at a historical moment.
[0051] Step 5114: Based on the cloud type matrix, determine the cloud physical characteristics at multiple historical moments in multiple historical periods and the actual values of cloud types in each historical period, thereby obtaining the training set.
[0052] Specifically, to prevent incorrect cloud type assignments at the edges of satellite profiles, the cloud type of each cloud cluster in the cloud type matrix is standardized.
[0053] Step 52: Construct a convolutional neural network.
[0054] Step 53: Using the cloud physical features at multiple moments in each historical period as input and the actual values of cloud types in the corresponding historical periods as output, train the convolutional neural network to obtain the cloud type recognition model.
[0055] In one exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement a ground-based radar-based cloud type identification method.
[0056] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements a ground-based radar-based cloud type identification method.
[0057] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements a ground-based radar-based cloud type identification method.
[0058] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 2 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and databases. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a cloud type identification method based on ground radar.
[0059] Those skilled in the art will understand that Figure 2 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0060] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).
[0061] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.
[0062] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0063] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0064] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A cloud type identification method based on ground radar, characterized in that, The cloud type identification method based on ground radar includes: The radar profiles of the ground radar at multiple times during the current period are obtained; the radar profiles include radar observation data from multiple radar observation points, and the radar observation data includes reflectivity factor, cloud feature identifiers, altitude, and temperature; The radar profiles at multiple moments in the current time period are processed to obtain the processed radar profiles at multiple moments in the current time period. A cloud radar data matrix is constructed based on the processed radar profiles from multiple moments in the current time period. Based on the cloud radar data matrix, the cloud physical characteristics of the clouds in the current time period are determined; the cloud physical characteristics include: statistical characteristics of cloud top height, cloud bottom height, cloud top temperature, cloud bottom temperature, cloud thickness, and reflectivity factor; the statistical characteristics include mean, mode, median, and 95% interval range. The cloud physical characteristics of the clouds in the current time period are input into the cloud type identification model to obtain the identification value of the cloud type in the current time period; the cloud type identification model is obtained by training a convolutional neural network; the cloud types include: cirrus clouds, cumulus clouds, stratocumulus clouds, nimbostratus clouds, altostratus clouds, altocumulus clouds and deep convection clouds.
2. The cloud type identification method based on ground radar according to claim 1, characterized in that, The radar profiles at multiple moments within the current time period are processed to obtain the processed radar profiles at those multiple moments, including: The radar profiles with a first proportion greater than 35% in the current time period are updated to missing values to obtain processed radar profiles for multiple moments in the current time period; the first proportion is the ratio of the number of radar observation data with a reflectivity factor greater than -10dBZ at temperatures above 0℃ to the total number of radar observation data in the radar profile.
3. The cloud type identification method based on ground radar according to claim 1, characterized in that, Each row of the cloud radar data matrix corresponds to a time point, and each column of the cloud radar data matrix corresponds to a processed radar profile. Each element of the cloud radar data matrix corresponds to a radar observation data point in the processed radar profile at a given time point. Based on the cloud radar data matrix, the cloud physical characteristics of the clouds in the current time period are determined, including: Based on the cloud radar data matrix, the edge determination results of the cloud in the current time period are determined; the edge determination results include radar observation data corresponding to the vertical edges and radar observation data corresponding to the horizontal edges; Based on the edge determination results of the clouds in the current time period, determine the cloud physical characteristics of the clouds in the current time period.
4. The cloud type identification method based on ground radar according to claim 3, characterized in that, Based on the cloud radar data matrix, the edge determination result of the cloud in the current time period is determined, including: Any radar observation data with a cloud feature identifier of 1 in the cloud radar data matrix is identified as the target radar observation data; When there are more than 3 consecutive radar observation data points in the column element where the target radar observation data is located that can be represented as 0, the target radar observation data is determined to be the radar observation data corresponding to the upper and lower edges of the cloud. When the cloud feature identifier of the processed radar profile at each moment of a consecutive minute in the row element containing the target radar observation data is 0, the target radar observation data is determined to be the radar observation data corresponding to the left and right edges of the cloud.
5. The cloud type identification method based on ground radar according to claim 2, characterized in that, The process of determining the cloud type identification model includes: Construct a training set; the training set includes cloud physical characteristics at multiple historical moments in multiple historical periods and actual values of cloud types in each historical period; Constructing a convolutional neural network; By taking the cloud physical characteristics at multiple moments in each historical period as input and the actual value of the cloud type in the corresponding historical period as output, a convolutional neural network is trained to obtain a cloud type recognition model.
6. The cloud type identification method based on ground radar according to claim 5, characterized in that, Constructing the training set includes: The system acquires multiple satellite profiles within a 200-kilometer radius centered on the latitude and longitude of the ground radar location. These satellite profiles include: satellite observation time, melt layer height, and satellite observation data from multiple satellite observation points along the satellite profile. The satellite observation data includes reflectivity factor, cloud feature identifier, altitude, gas attenuation correction data, and cloud type. Define any satellite scan trajectory as the current satellite scan trajectory; Each satellite profile in the current satellite scan trajectory is preprocessed to obtain multiple preprocessed satellite profiles corresponding to the current satellite scan trajectory; The mean time corresponding to the current satellite scan trajectory is obtained by averaging the satellite observation times of all satellite profiles in the current satellite scan trajectory. The current historical period is defined as one hour before and after the mean time corresponding to the current satellite scanning trajectory. The radar profiles of each historical time in the current historical period are obtained, and the radar profile of any historical time is defined as the current radar profile. Update the radar observation data with cloud feature identifier of 0 in the current radar profile to the missing value to obtain the current radar profile after one processing step. The average value of the gas attenuation correction data in all preprocessed satellite profiles corresponding to the current satellite scan trajectory is calculated, and the difference is calculated based on the number of radar observation points in the current radar profile to obtain the gas attenuation correction data profile corresponding to the current satellite scan trajectory. Based on the gas attenuation correction data profile corresponding to the current satellite scan trajectory, the reflectivity factors in the current radar profile after the first processing are corrected to obtain the current radar profile after the second processing. The first judgment result is obtained by determining whether the second proportion of the current radar profile after secondary processing is greater than 35%. The second proportion is the ratio of the number of radar observation data with reflectivity factor greater than -10dBZ whose height is below the average of the melting layer height of all preprocessed satellite profiles corresponding to the current satellite scanning trajectory to the number of radar observation data with cloud feature identifier of 1 in the current radar profile after secondary processing. If the first judgment result is yes, then the current radar profile after the second processing is updated to the missing value, and the current radar profile after the third processing is obtained. If the first judgment result is negative, then the current radar profile after the second processing is determined as the current radar profile after the third processing. Linearly interpolate the cloud types in all preprocessed satellite profiles corresponding to the current satellite scan trajectory to the current radar profile after three-stage processing to obtain the current radar profile with cloud types. Based on all radar profiles with cloud types corresponding to all satellite scan trajectories, a cloud type matrix is constructed; each row of the cloud type matrix corresponds to a historical moment, and each column of the cloud type matrix corresponds to the current radar profile with cloud type; each element of the cloud type matrix corresponds to a radar observation data and cloud type of the current radar profile with cloud type at a historical moment. Based on the cloud type matrix, the cloud physical characteristics at multiple historical moments in multiple historical periods and the actual values of cloud types in each historical period are determined, thereby obtaining the training set.
7. The cloud type identification method based on ground radar according to claim 6, characterized in that, Each satellite profile in the current satellite scan trajectory is preprocessed to obtain multiple preprocessed satellite profiles corresponding to the current satellite scan trajectory, including: Define any satellite profile in the current satellite scan trajectory as the current satellite profile; Update the satellite observation data with cloud feature identifier of 0 in the current satellite profile to the missing value to obtain the current satellite profile after one processing step; The reflectivity factor of the corresponding satellite observation point is corrected by using the gas attenuation correction data of each satellite observation point in the current satellite profile after the first processing, and the current satellite profile after the second processing is obtained. The second judgment result is obtained by determining whether the third proportion of the current satellite profile after secondary processing is greater than 35%. The third proportion is the ratio of the number of satellite observation data with a reflectivity factor greater than -10dBZ that are below the melting layer height of the current satellite profile to the number of satellite observation data with a cloud feature identifier of 1 in the current satellite profile after secondary processing. If the second judgment result is yes, then the current satellite profile after secondary processing is updated to the missing value to obtain the preprocessed current satellite profile; If the second judgment result is negative, then the current satellite profile after secondary processing will be determined as the current satellite profile after preprocessing.
8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the cloud type identification method based on ground radar according to any one of claims 1-7.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the cloud type identification method based on ground radar as described in any one of claims 1-7.
10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the cloud type identification method based on ground radar as described in any one of claims 1-7.