CN111291903A - Precipitation amount estimation method and device, computer equipment and readable storage medium - Google Patents
Precipitation amount estimation method and device, computer equipment and readable storage medium Download PDFInfo
- Publication number
- CN111291903A CN111291903A CN202010065651.4A CN202010065651A CN111291903A CN 111291903 A CN111291903 A CN 111291903A CN 202010065651 A CN202010065651 A CN 202010065651A CN 111291903 A CN111291903 A CN 111291903A
- Authority
- CN
- China
- Prior art keywords
- data
- precipitation
- data set
- radar detection
- machine learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/95—Radar or analogous systems specially adapted for specific applications for meteorological use
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
- G01S7/4802—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Environmental & Geological Engineering (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Electromagnetism (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Atmospheric Sciences (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Environmental Sciences (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The method comprises the steps of obtaining first radar detection data and first precipitation data, matching the first radar detection data and the first precipitation data to obtain a target matching data set, training an initial machine learning model through the target matching data set to obtain a precipitation estimation model, and estimating precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data; the method can adopt a machine learning model to fit the relation between the radar detection data and the discretization precipitation data, so that the purpose of predicting the precipitation in the large coverage range is achieved, the purpose of predicting the precipitation in the large coverage range is achieved through the machine learning model by adopting a machine learning algorithm, and the accuracy of the precipitation prediction result is improved.
Description
Technical Field
The present application relates to the field of weather service technologies, and in particular, to a precipitation amount estimation method, apparatus, computer device, and readable storage medium.
Background
With the development of scientific technology, more and more advanced technologies are applied to the field of weather prediction, and accurate weather forecast service can help people plan outdoor activities and even can provide early warning of accidents such as flood or traffic. The rainfall amount parameter is also an important parameter in weather forecast, and rainfall amount prediction can provide a very important parameter for flood forecast, so that the method is an important means for enhancing cooperation of meteorological departments and water conservancy departments to serve disaster prevention and disaster resistance and economic construction.
In the traditional technology, according to various weather data in a small coverage range, the whole precipitation data in a large coverage range is estimated by adopting a manual statistical mode and an empirical formula. However, the conventional estimation method results in a low accuracy of the estimation result.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a precipitation amount estimation method, device, computer device and readable storage medium capable of improving accuracy of precipitation amount estimation result.
The embodiment of the application provides a precipitation estimation method, which comprises the following steps:
acquiring first radar detection data and first precipitation data;
matching the first radar detection data and the first precipitation data to obtain a target matching data set;
training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
In one embodiment, the acquiring the first radar detection data and the first precipitation data includes:
acquiring first radar detection data in a preset area in a preset time period;
and acquiring first precipitation data in the preset area in the preset time period.
In one embodiment, the matching the first radar detection data and the first precipitation data to obtain a target matching data set includes:
and performing data conversion processing on the first radar detection data and the first precipitation data in the same sampling time period to obtain the target matching data group comprising the first radar detection data and the first precipitation data.
In one embodiment, the performing data conversion processing on the first radar detection data and the precipitation data within the same sampling time period to obtain the target matching data group including the first radar detection data and the first precipitation data includes:
combining the first radar detection data and the precipitation data in the same sampling time period to obtain an initial matching data set;
and extracting the initial matching data set to obtain the target matching data set.
In one embodiment, the extracting the initial matching data set to obtain the target matching data set includes:
sampling the initial matching data set by adopting a sampling method to obtain an intermediate matching data set;
if the intermediate matching data set does not meet the condition of training processing, performing data preprocessing on the intermediate matching data set to obtain the target matching data set; wherein the condition of the training process characterizes the data in the intermediate matched data set as satisfying the initial machine learning model process.
In one embodiment, the method further comprises: and if the intermediate matching data set meets the condition of training processing, taking the intermediate matching data set as the target matching data set.
In one embodiment, the training an initial machine learning model through the target matching data set to obtain a precipitation prediction model includes:
adjusting an initial machine learning model through the target matching data set to obtain an intermediate machine learning model; wherein the adjustment process characterizes an adjustment to a structure of the initial machine learning model and a hyper-parameter in the initial machine learning model;
and training the intermediate machine learning model through the target matching data set to obtain the precipitation amount estimation model.
In one embodiment, the estimating the precipitation in the preset time period by the precipitation estimation model to obtain second precipitation data includes:
acquiring second radar detection data in a preset time period;
and inputting the second radar detection data into the precipitation estimation model to obtain the second precipitation in a preset time period.
The embodiment of the application provides a precipitation amount pre-estimation device, the precipitation amount pre-estimation device includes:
the acquisition module is used for acquiring first radar detection data and first precipitation data;
the matching module is used for matching the first radar detection data and the first precipitation data to obtain a target matching data set;
the training module is used for training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and the estimation module is used for estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
The embodiment of the application provides a computer device, which comprises a memory and a processor, wherein a computer program capable of running on the processor is stored in the memory, and the processor executes the computer program to realize the following steps:
acquiring first radar detection data and first precipitation data;
matching the first radar detection data and the first precipitation data to obtain a target matching data set;
training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
An embodiment of the application provides a readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the following steps:
acquiring first radar detection data and first precipitation data;
matching the first radar detection data and the first precipitation data to obtain a target matching data set;
training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
The method comprises the steps of obtaining first radar detection data and first precipitation data, matching the first radar detection data and the first precipitation data to obtain a target matching data set, training an initial machine learning model through the target matching data set to obtain a precipitation estimation model, and estimating precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data; the method can adopt a machine learning model to fit the relation between the radar detection data and the discretization precipitation data, so that the purpose of predicting the precipitation in the large coverage range is achieved, the purpose of predicting the precipitation in the large coverage range is achieved through the machine learning model by adopting a machine learning algorithm, and the accuracy of the precipitation prediction result is improved.
Drawings
Fig. 1 is a schematic flow chart of a precipitation estimation method according to an embodiment;
FIG. 2 is a schematic flow chart illustrating a precipitation estimation method according to another embodiment;
FIG. 3 is a schematic structural diagram of a precipitation quantity estimation device according to an embodiment;
fig. 4 is an internal structural diagram of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The rainfall amount estimation method can be applied to a rainfall amount estimation system, and the rainfall amount estimation system comprises a meteorological radar, a meteorological ground observation station and a background server. In this embodiment, the weather radar may be characterized as a radar that monitors strong convection weather, the weather radar may be a rain radar, the radar model is not limited in this embodiment, and the weather radar may detect a radar echo diagram; the meteorological ground observation station can be a meteorological station which can predict information such as wind speed, wind direction, rainfall, temperature, air pressure and the like of a coverage area at regular time; the background server can receive data detected by the meteorological radar and the meteorological ground observation station, process the data and realize the rainfall estimation process; the precipitation amount may include a rainfall amount and/or a snowfall amount, among others. Optionally, the effective radius of the coverage area of the weather radar may be set according to the characteristics of the weather radar; in this embodiment, the effective radius may be set to 250 km; within the coverage of the weather radar, a plurality of weather ground observation stations can be arranged. Optionally, the meteorological radar, the meteorological ground observation station and the background server may communicate with each other through a wireless connection. Optionally, the wireless connection mode may be Wi-Fi, mobile network or bluetooth connection. The specific procedure of the precipitation estimation method will be described in the following examples.
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Fig. 1 is a schematic flow chart of a precipitation estimation method according to an embodiment. The embodiment relates to a process for acquiring a precipitation prediction model and predicting precipitation in a preset time period through the precipitation prediction model. As shown in fig. 1, the method includes:
and S1000, acquiring first radar detection data and first precipitation data.
Specifically, the meteorological radar can detect a target object to obtain an original radar echo map, then analyzes the original radar echo map to obtain original radar echo data, then performs format conversion processing on the original radar echo data to obtain converted radar detection data (namely first radar detection data), and sends the radar detection data in the format to the background server. The radar detection data may be referred to as radar echo data, i.e. the first radar detection data may also be referred to as radar echo data. Optionally, the data format of the original radar echo data may be an iris format, and the radar detection data after the format conversion processing may be equal-height surface data, that is, data in a CAPPI format. For example, the size of the iso-facial data may be represented as 40 × 250 × 720, 40 representing the number of layers, 250 representing the height of each layer, and 720 representing the data resolution (i.e., the resolution of the radar echo map). Alternatively, the target may be clouds, rain, snow, hail, etc. in the atmosphere. Optionally, the weather radar may perform one scan at a certain time interval to obtain one radar echo pattern. Wherein the first radar detection data may reflect the size of precipitation particles in the atmosphere. Optionally, each meteorological ground observation station may detect precipitation data, and send the precipitation data to the background server. Optionally, each meteorological ground observation station can perform detection once at a certain time interval, and a set of precipitation data detected by all meteorological ground observation stations obtains first precipitation data, which can be discretized precipitation data within the coverage range of the meteorological radar, that is, the sum of the precipitation data within the coverage range of each meteorological ground observation station set within the coverage range of the meteorological radar. Wherein, the weather radar coverage area can comprise the sum of the coverage areas of each weather ground observation station.
It should be noted that the background server may receive the first radar detection data sent by the weather radar and the first precipitation data sent by all the weather ground observation stations, and further perform data processing on the first radar detection data and the first precipitation data.
And S2000, matching the first radar detection data and the first precipitation data to obtain a target matching data set.
Specifically, the first radar detection data may include a plurality of data, and the first precipitation amount data may also include a plurality of data. Because the radar detection data and the precipitation data are not necessarily in one-to-one correspondence in time and position, the background server can perform matching processing on the first radar detection data and the first precipitation data. Optionally, the matching process may be to match each data in the first radar detection data with each data in the first precipitation data, so as to obtain a one-to-many or many-to-one data set. For example, the first radar detection data a is represented as (a)1，a2) The first precipitation data B is represented by (B)1，b2，b3，b4，b5，b6) Then after matching, a one-to-many data set of { a ] can be obtained1：[b1，b2，b3]，a2：[b4，b5，b6]Either a many-to-one data set of (a)1，a2)：[b1，b2，b3，b4，b5，b6]}; in addition, other forms of one-to-many data sets, or many-to-one data sets, are also possible.
And S3000, training an initial machine learning model through the target matching data set to obtain a precipitation estimation model.
Specifically, the initial machine learning model may be a supervised learning model for classification. Optionally, the initial machine learning model may be a convolutional neural network model, a support vector machine, naive bayes, a decision tree, or an integrated model, which is not limited in this embodiment. It should be noted that the initial machine learning model may implement a classification process using a corresponding machine learning algorithm. Optionally, the target match data set may include a training set, a validation set, and a test set during a training process. Optionally, the background server may train the initial machine learning model through the target matching data set to obtain the precipitation amount estimation model. The precipitation prediction model can represent the relation between radar detection data and precipitation data.
And S4000, estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
Specifically, the preset time period may be any time period to be estimated, and the length of the preset time period is not limited in this embodiment. Optionally, the second precipitation data may be precipitation data with a large coverage area estimated within a preset time period. Optionally, the second precipitation data may be the overall precipitation data within the coverage of the weather radar.
Optionally, the step of predicting the precipitation in the preset time period by using the precipitation prediction model in step S4000 to obtain the second precipitation data may specifically include the following steps: acquiring second radar detection data in a preset time period; and inputting the second radar detection data into the precipitation estimation model to obtain the second precipitation in a preset time period.
It should be noted that the background server may receive second radar detection data obtained by scanning the meteorological radar in the preset time period, then input the second radar detection data in the preset time period to the precipitation amount estimation model, and predict the precipitation amount in the preset time period through the precipitation amount estimation model to obtain the second precipitation amount data.
The rainfall estimation method provided by the embodiment comprises the steps of obtaining first radar detection data and first rainfall data, matching the first radar detection data and the first rainfall data to obtain a target matching data set, training an initial machine learning model through the target matching data set to obtain a rainfall estimation model, and estimating the rainfall in a preset time period through the rainfall estimation model to obtain second rainfall data; the method can adopt a machine learning model to fit the relation between the radar detection data and the discretization precipitation data, thereby achieving the purpose of predicting the precipitation in a large coverage range, reducing the complexity of large data processing by the large data analysis mode and shortening the data processing time; meanwhile, the method can achieve the purpose of estimating the precipitation in the large coverage range by adopting a machine learning algorithm through a machine learning model, so that the accuracy of the precipitation estimation result is improved.
Fig. 2 is a schematic specific flow chart of a precipitation amount estimation method according to another embodiment, and based on fig. 1, the process of acquiring the first radar detection data and the first precipitation amount data in step S1000 may be implemented by the following steps:
step S1100, first radar detection data in a preset area in a preset time period are obtained.
In this embodiment, the preset time period may be determined according to actual requirements, and the length of the preset time period is not limited in this embodiment; the preset area may be any one area, and this embodiment is not limited in any way. Alternatively, the preset time period may include a plurality of interval scan times. Optionally, the background server may obtain first radar detection data in a preset region of a preset time period obtained by the detection of the weather radar.
Step S1200, first precipitation data in the preset area in the preset time period are obtained.
Specifically, the background server may obtain first precipitation data in a preset region in a preset time period, which is obtained by detection of the meteorological ground observation station. Optionally, a plurality of meteorological ground observation stations may be disposed in the preset area, and the first precipitation data may be a set of precipitation data detected by the plurality of meteorological ground observation stations.
Further, the process of performing matching processing on the first radar detection data and the first precipitation data in the step S2000 to obtain a target matching data set may include the following steps:
step S2100, perform data conversion processing on the first radar detection data and the first precipitation data in the same sampling time period to obtain the target matching data group including the first radar detection data and the first precipitation data.
Specifically, the meteorological ground observation station used in this embodiment can detect minute-scale data. Illustratively, if the preset time period is 9 to 10 points, the weather radar performs scanning every five minutes, and all radar detection data obtained each time are represented as liThe radar detection data L scanned in the preset time period is expressed as (L)1，l2，...，l11，l12) Detecting every weather ground observation station within the coverage range of the weather radar once every minute, and expressing the precipitation data obtained every time as zn，znRepresenting the set of precipitation data detected by all meteorological ground observation stations at a time, and the precipitation data Z detected in a preset time period is represented as (Z)1，z2，...，z59，z60) Then the background server can divide all radar detection data l from 9 o 'clock to 9 o' clock 05 in five minutes1And all the precipitation data z in five minutes from 9 o 'clock to 9 o' clock 05 min1，z2，z3，z4，z5And performing data conversion, and performing data conversion processing on the radar detection data and the precipitation data within every five minutes by analogy to obtain a target matching data group. Alternatively, the data conversion process may be characterized as a comprehensive arithmetic processing process in which a plurality of types of arithmetic operations process data are converted together, or may be characterized as a simple combination process that does not participate in an actual arithmetic operationAnd (6) processing. In this embodiment, a simple combining process can be understood as a process of combining a plurality of individual data to obtain a data set.
Optionally, the step of performing data conversion processing on the first radar detection data and the first precipitation data in the same sampling time period in step S2100 to obtain the target matching data group including the first radar detection data and the first precipitation data may include the following steps:
and step S2110, combining the first radar detection data and the precipitation data in the same sampling time period to obtain an initial matching data set.
In this embodiment, the data conversion process may be a combination process. Continuing with the previous example, radar detection data l for five minutes, 9 o 'clock to 9 o' clock 051And the precipitation data z from 9 o 'clock to 9 o' clock 05 min1，z2，z3，z4，z5Combining the radar detection data l in five minutes from 9 points 06 to 9 points 101And the precipitation data z from 9 o 'clock to 9 o' clock 05 min1，z2，z3，z4，z5Combining, analogizing in turn, combining the radar detection data and the precipitation data within every five minutes to obtain an initial matching data set A which can be expressed as { l }1：[z1，z2，z3，z4，z5]，l2：[z6，z7，z8，z9，z10]，...，l12：[z56，z57，z58，z59，z60]}。
And S2120, extracting the initial matching data set to obtain the target matching data set.
Specifically, the background server may extract the initial matching data set to obtain the target matching data set. Optionally, the extraction process may include a sampling process and a screening process. Alternatively, the target matching data set may be an initial matching data set including each weather ground viewSurvey station aerial radar detection data dmThe precipitation data g detected by the corresponding meteorological ground observation stationkThe matching data set of (1).
Optionally, with continuing reference to fig. 2, the step of extracting the initial matching data set in step S2120 to obtain the target matching data set may specifically include:
step S2121, sampling the initial matching data set by adopting a sampling method to obtain an intermediate matching data set;
it should be noted that the above sampling method can be characterized as a process of sampling data around the data therein. Optionally, the background server may detect data d for each radar in the initial matching data setmAnd sampling by adopting a sampling method to obtain an intermediate matching data set. In this embodiment, the radar detection data may be plane data, and the precipitation data may be point data; in order to find out the corresponding data of the precipitation data in the radar detection data without generating errors, the background server can sample all the radar detection data around the target radar detection data by adopting a 9-point method or a 121-point method to obtain the radar detection data above each meteorological ground observation station, and then combine the precipitation data detected by each meteorological ground observation station with the radar detection data above the meteorological ground observation station to obtain a middle matching data set. Optionally, the target radar detection data may be corresponding point data (i.e. one point data in the radar detection data) over the meteorological ground observation station. Alternatively, the sampling method may be other point sampling methods. Continuing with the previous example, the precipitation data detected in five minutes from 9 o 'clock to 9 o' clock 05 of each meteorological ground observation station may be represented as g1，g2，...，g5Wherein z is1May include g1，z2May include g2By analogy with the above, z5May include g5The radar detection data over each meteorological ground observation station obtained by the background server by adopting a sampling method can be represented as dm(ii) a If the number of radar detections above the first meteorological ground observation station is within five minutesAccording to d1And radar survey data d overhead of a second meteorological ground observation station within five minutes2And analogizing in turn, and detecting data d of the radar above the last meteorological ground observation station in five minutesmThen the intermediate matching data set obtained by the background server can be expressed as { d }1：[g1，g2，g3，g4，g5]，d2：[g6，g7，g8，g9，g10]，...，dm：[g5(m-1)+1，g5(m-1)+2，g5(m-1)+3，g5(m-1)+4，g5(m-1)+5]In which d ismRepresenting a data set.
Step S2122, if the intermediate matching data set does not meet the condition of training processing, performing data preprocessing on the intermediate matching data set to obtain the target matching data set; wherein the condition of the training process characterizes the data in the intermediate matched data set as satisfying the initial machine learning model process.
In this embodiment, the initial machine learning model may process data of a fixed size. And if the size of the data in the intermediate matching data group obtained by the background server does not accord with the fixed size of the data which can be processed by the initial machine learning model, performing data preprocessing on the intermediate matching data group to obtain a target matching data group. Optionally, the data preprocessing may be a clipping processing, a normalization processing, a denoising processing, and the like. In this embodiment, the size of the data in the intermediate matching data set may be equal to or greater than the fixed size of the initial machine learning model processable data. Here, the process in step S2122 may be understood as a screening process.
Optionally, after the process of step S2121, the method may further include: and if the intermediate matching data set meets the condition of training processing, taking the intermediate matching data set as the target matching data set.
Further, the step S2122 of performing data preprocessing on the intermediate matching data set to obtain the target matching data set may specifically include: and cutting the data in the intermediate matching data group to obtain the target matching data group.
And if the size of the data in the intermediate matching data group is larger than the fixed size of the data processable by the initial machine learning model, the background server can cut the intermediate matching data group to obtain a target matching data group suitable for the initial machine learning model to process.
According to the rainfall estimation method provided by the embodiment, first radar detection data and first rainfall data can be obtained, the first radar detection data and the first rainfall data are matched to obtain a target matching data set, an initial machine learning model is trained through the target matching data set to obtain a rainfall estimation model, and rainfall in a preset time period is estimated through the rainfall estimation model to obtain second rainfall data; the method can adopt a machine learning model to fit the relation between radar detection data and precipitation data, thereby achieving the purpose of precipitation estimation, reducing the complexity of big data processing by the big data analysis mode, and shortening the data processing time; meanwhile, the method can adopt a machine learning algorithm to estimate the precipitation through a machine learning model, so that the accuracy of the precipitation estimation result is improved.
As an embodiment, with reference to fig. 2, the process of training the initial machine learning model through the target matching data set in step S3000 to obtain the precipitation prediction model may specifically include the following steps:
step S3100, adjusting an initial machine learning model through the target matching data set to obtain an intermediate machine learning model; wherein the adjustment process characterizes an adjustment to a structure of the initial machine learning model and a hyper-parameter in the initial machine learning model.
In this embodiment, the initial machine learning model may be a support vector machine or a convolutional neural network model. Because the initial machine learning model may include a convolutional layer, a full connection layer, a pooling layer, and the like, before the precipitation amount estimation model is obtained through training, the background server needs to adjust the initial machine learning model through the target matching data set to obtain an intermediate machine learning model. Optionally, the adjusting process may be understood as adjusting the structure in the initial machine learning model to the optimal machine learning model, and adjusting the hyper-parameters of the optimal machine learning model; that is, a part of the structure of the initial machine learning model is deleted, the remaining structure is used as the optimal machine learning model, i.e., the intermediate machine learning model, and meanwhile, the hyper-parameters of the intermediate machine learning model are adjusted to be the optimal parameters. The hyper-parameters may be parameters that cannot be obtained based on data and are set only by human experience, and the hyper-parameters may include a learning rate of a model, a number of model channels, a number of network layers, and the like. It should be noted that, the parameters of the initial machine learning model and the initial state of the hyper-parameters may be set arbitrarily, and are not limited.
Step S3200, training the intermediate machine learning model through the target matching data set to obtain the precipitation amount estimation model.
Specifically, the background server can train the intermediate machine learning model through the target matching data set to obtain the precipitation amount estimation model. Optionally, the target match data set may include a training set, a validation set, and a test set during a training process. In the training process, the network loss function can be set arbitrarily according to the actual network requirements, and the current intermediate machine learning model is subjected to back propagation by using the current network loss function so as to update the model parameters of the current intermediate machine learning model; when the difference between the current network loss function and the network loss function constructed in the last iteration process does not exceed the preset difference threshold, the training process can be ended. After training is finished, the obtained machine learning model can be defined as a precipitation estimation model, and parameters of the model are optimal at the moment. Alternatively, the preset difference threshold may be a small datum, and may be set arbitrarily, for example, 0.001, 0.0008, or the like.
According to the rainfall estimation method provided by the embodiment, a machine learning model can be adopted to fit the relation between radar detection data and rainfall data, so that the purpose of rainfall estimation is achieved, the complexity of big data processing is reduced through the big data analysis mode, and the data processing time is shortened; meanwhile, the method can adopt a machine learning algorithm to estimate the precipitation through a machine learning model, so that the accuracy of the precipitation estimation result is improved.
It should be understood that although the steps in the flowcharts of fig. 1-2 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 1-2 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.
For the specific limitation of the precipitation amount estimation device, reference may be made to the above limitation on the precipitation amount estimation method, and details are not described herein again. All or part of the modules in the precipitation estimation device of the computer equipment can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
Fig. 3 is a schematic structural diagram of a precipitation amount estimation device according to an embodiment. As shown in fig. 3, the system may include: an acquisition module 11, a matching module 12, a training module 13 and an estimation module 14.
Specifically, the obtaining module 11 is configured to obtain first radar detection data and first precipitation data;
the matching module 12 is configured to perform matching processing on the first radar detection data and the first precipitation data to obtain a target matching data set;
the training module 13 is configured to train an initial machine learning model through the target matching data set to obtain a precipitation amount estimation model;
the estimation module 14 is configured to estimate the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
The precipitation amount estimation apparatus provided in this embodiment may implement the method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
In one embodiment, the matching module 12 includes: and a data conversion unit.
The data conversion unit is configured to perform data conversion processing on the first radar detection data and the first precipitation data within the same sampling time period to obtain the target matching data group including the first radar detection data and the first precipitation data.
The precipitation amount estimation apparatus provided in this embodiment may implement the method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
In one embodiment, the data conversion unit includes: a combining subunit and a sampling subunit.
The combining subunit is configured to combine the first radar detection data and the precipitation data within the same sampling time period to obtain an initial matching data set;
and the sampling subunit is used for extracting the initial matching data group to obtain the target matching data group.
The precipitation amount estimation apparatus provided in this embodiment may implement the method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
In one embodiment, the sampling subunit is specifically configured to perform sampling processing on the initial matching data set by using a sampling method to obtain an intermediate matching data set; if the intermediate matching data set is judged not to be in accordance with the training processing conditions, performing data preprocessing on the intermediate matching data set to obtain the target matching data set; wherein the condition of the training process characterizes data satisfying the initial machine learning model process.
The precipitation amount estimation apparatus provided in this embodiment may implement the method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
In one embodiment, the training module 13 includes: an adjusting unit and a training unit.
The adjusting unit is used for adjusting an initial machine learning model through the target matching data set to obtain the intermediate machine learning model; wherein the adjustment process characterizes an adjustment to a structure of the initial machine learning model and a hyper-parameter in the initial machine learning model;
and the training unit is used for training the intermediate machine learning model through the target matching data set to obtain the precipitation amount estimation model.
The precipitation amount estimation apparatus provided in this embodiment may implement the method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
In one embodiment, the estimation module 14 includes: a third acquisition unit and a processing unit.
The third acquiring unit is used for acquiring second radar detection data in a preset time period;
and the processing unit is used for inputting the second radar detection data into the precipitation estimation model to obtain the second precipitation in a preset time period.
The precipitation amount estimation apparatus provided in this embodiment may implement the method embodiments, and the implementation principle and the technical effect are similar, which are not described herein again.
In one embodiment, a computer device is provided, the internal structure of which may be as shown in FIG. 4. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external computer device through a network connection. The computer program is executed by a processor to implement a precipitation estimation method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 4 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring first radar detection data and first precipitation data;
matching the first radar detection data and the first precipitation data to obtain a target matching data set;
training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
In one embodiment, a readable storage medium is provided, having stored thereon a computer program which, when executed by a processor, performs the steps of:
acquiring first radar detection data and first precipitation data;
matching the first radar detection data and the first precipitation data to obtain a target matching data set;
training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A precipitation estimation method is characterized by comprising the following steps:
acquiring first radar detection data and first precipitation data;
matching the first radar detection data and the first precipitation data to obtain a target matching data set;
training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
2. The method of claim 1, wherein the matching the first radar detection data and the first precipitation data to obtain a target match data set comprises:
and performing data conversion processing on the first radar detection data and the first precipitation data in the same sampling time period to obtain the target matching data group comprising the first radar detection data and the first precipitation data.
3. The method of claim 2, wherein the data conversion processing of the first radar detection data and the precipitation data within the same sampling period to obtain the target matching data set including the first radar detection data and the first precipitation data comprises:
combining the first radar detection data and the precipitation data in the same sampling time period to obtain an initial matching data set;
and extracting the initial matching data set to obtain the target matching data set.
4. The method of claim 3, wherein the extracting the initial matching data set to obtain the target matching data set comprises:
sampling the initial matching data set by adopting a sampling method to obtain an intermediate matching data set;
if the intermediate matching data set does not meet the condition of training processing, performing data preprocessing on the intermediate matching data set to obtain the target matching data set; wherein the condition of the training process characterizes the data in the intermediate matched data set as satisfying the initial machine learning model process.
5. The method of claim 4, further comprising: and if the intermediate matching data set meets the condition of training processing, taking the intermediate matching data set as the target matching data set.
6. The method of claim 1, wherein the training an initial machine learning model through the target matching data set to obtain a precipitation prediction model comprises:
adjusting the initial machine learning model through the target matching data set to obtain an intermediate machine learning model; wherein the adjustment process characterizes an adjustment to a structure of the initial machine learning model and a hyper-parameter in the initial machine learning model;
and training the intermediate machine learning model through the target matching data set to obtain the precipitation amount estimation model.
7. The method according to claim 1, wherein the estimating precipitation in a preset time period by the precipitation estimation model to obtain second precipitation data comprises:
acquiring second radar detection data in a preset time period;
and inputting the second radar detection data into the precipitation estimation model to obtain the second precipitation data in a preset time period.
8. A precipitation estimator, the estimator comprising:
the acquisition module is used for acquiring first radar detection data and first precipitation data;
the matching module is used for matching the first radar detection data and the first precipitation data to obtain a target matching data set;
the training module is used for training an initial machine learning model through the target matching data set to obtain a precipitation estimation model;
and the estimation module is used for estimating the precipitation in a preset time period through the precipitation estimation model to obtain second precipitation data.
9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010065651.4A CN111291903A (en) | 2020-01-20 | 2020-01-20 | Precipitation amount estimation method and device, computer equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010065651.4A CN111291903A (en) | 2020-01-20 | 2020-01-20 | Precipitation amount estimation method and device, computer equipment and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111291903A true CN111291903A (en) | 2020-06-16 |
Family
ID=71031071
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010065651.4A Pending CN111291903A (en) | 2020-01-20 | 2020-01-20 | Precipitation amount estimation method and device, computer equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111291903A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112926664A (en) * | 2021-03-01 | 2021-06-08 | 南京信息工程大学 | Feature selection and CART forest short-time strong rainfall forecasting method based on evolutionary algorithm |
| CN113111567A (en) * | 2021-02-26 | 2021-07-13 | 中国科学院空天信息创新研究院 | Evapotranspiration component separation method, component separation device, apparatus, and medium |
| US11727303B2 (en) * | 2019-08-22 | 2023-08-15 | Kyndryl, Inc. | Precipitation detection using mobile devices |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107703564A (en) * | 2017-10-13 | 2018-02-16 | 中国科学院深圳先进技术研究院 | A kind of precipitation predicting method, system and electronic equipment |
| CN110346844A (en) * | 2019-07-15 | 2019-10-18 | 南京恩瑞特实业有限公司 | Quantitative Precipitation estimating and measuring method of the NRIET based on cloud classification and machine learning |
-
2020
- 2020-01-20 CN CN202010065651.4A patent/CN111291903A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107703564A (en) * | 2017-10-13 | 2018-02-16 | 中国科学院深圳先进技术研究院 | A kind of precipitation predicting method, system and electronic equipment |
| CN110346844A (en) * | 2019-07-15 | 2019-10-18 | 南京恩瑞特实业有限公司 | Quantitative Precipitation estimating and measuring method of the NRIET based on cloud classification and machine learning |
Non-Patent Citations (1)
| Title |
|---|
| 肖晨: "人工神经网络在雷达定量测量降水中的研究与实现" * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11727303B2 (en) * | 2019-08-22 | 2023-08-15 | Kyndryl, Inc. | Precipitation detection using mobile devices |
| CN113111567A (en) * | 2021-02-26 | 2021-07-13 | 中国科学院空天信息创新研究院 | Evapotranspiration component separation method, component separation device, apparatus, and medium |
| CN112926664A (en) * | 2021-03-01 | 2021-06-08 | 南京信息工程大学 | Feature selection and CART forest short-time strong rainfall forecasting method based on evolutionary algorithm |
| CN112926664B (en) * | 2021-03-01 | 2023-11-24 | 南京信息工程大学 | Feature selection and CART forest short-time strong precipitation prediction method based on evolutionary algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11249221B2 (en) | Weather forecasting systems and methods | |
| US20220043182A1 (en) | Spatial autocorrelation machine learning-based downscaling method and system of satellite precipitation data | |
| US11353625B1 (en) | Systems and methods for forecasting lightning and severe storms | |
| CN111291903A (en) | Precipitation amount estimation method and device, computer equipment and readable storage medium | |
| US8818029B1 (en) | Weather forecasting systems and methods | |
| CN111523414A (en) | Face recognition method and device, computer equipment and storage medium | |
| WO2020044127A1 (en) | Atmospheric pollution forecasting method | |
| CN109946762A (en) | A kind of method and system based on probability distribution Short-term Forecast precipitation | |
| CN110705759B (en) | Water level early warning monitoring method and device, storage medium and electronic equipment | |
| Chen et al. | A statistical downscaling model for forecasting summer rainfall in China from DEMETER hindcast datasets | |
| CN111105068B (en) | Numerical mode correction method based on sequential regression learning | |
| Ng et al. | Uncertainty analysis of rainfall depth duration frequency curves using the bootstrap resampling technique | |
| US20200355846A1 (en) | Systems and methods for forecasting weather | |
| Haddad et al. | Comparing three methods to form regions for design rainfall statistics: two case studies in Australia | |
| Boni et al. | Extreme rainfall events: Learning from raingauge time series | |
| Sokol et al. | Nowcasting of precipitation–advective statistical forecast model (SAM) for the Czech Republic | |
| CN111913236A (en) | Meteorological data processing method, meteorological data processing device, computer equipment and storage medium | |
| Mamadjanova et al. | Assessment of mudflow risk in Uzbekistan using CMIP5 models | |
| van der Plas et al. | A comparative verification of high-resolution precipitation forecasts using model output statistics | |
| CN110309855B (en) | Training method for image segmentation, computer device and storage medium | |
| CN113269805B (en) | Rainfall event guided remote sensing rainfall inversion training sample self-adaptive selection method | |
| Hu | An algorithm for converting weather radar data into GIS polygons and its application in severe weather warning systems | |
| KR101870743B1 (en) | Tidal flat experience index forecasting apparatus and the method thereof | |
| CN108055639A (en) | A kind of density of stream of people Forecasting Methodology and system based on signaling data analysis | |
| CN113075751A (en) | Method and system for fusing observation data in short-term forecasting |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |