CN111753461B

CN111753461B - Tidal water level correction method, target residual water level acquisition method, device and equipment

Info

Publication number: CN111753461B
Application number: CN202010396754.9A
Authority: CN
Inventors: 任磊; 潘广维; 姬进财; 杨清书
Original assignee: Sun Yat Sen University
Current assignee: Sun Yat Sen University
Filing date: 2020-05-12
Publication date: 2024-07-05
Anticipated expiration: 2040-05-12

Abstract

The application relates to a tidal water level correction method, a target residual water level acquisition method, a device and equipment. The method comprises the following steps: acquiring current target data of a river basin to be measured; the current target data are environment data of target data types which have influence on the residual water level; inputting the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data; the residual water level prediction model is a neural network model obtained through training according to historical target data and actual residual water levels of the target data type. By adopting the method, the accuracy of the output target residual water level can be improved.

Description

Tidal water level correction method, target residual water level acquisition method, device and equipment

Technical Field

The application relates to the technical field of ocean science, in particular to a tidal water level correction method, a target residual water level acquisition device and target residual water level acquisition equipment.

Background

The residual water level is sea surface water level variation caused by weather factors and seasonal weather factors, and can be obtained by removing astronomical tide level and average sea surface according to the measured water level.

The residual water level prediction is favorable for accurately predicting the ocean tide level, and the residual water level prediction in the prior art is often inaccurate, so that the prediction accuracy of the tide level is affected, and great difficulty is caused to the subsequent ocean engineering construction and ocean scientific research according to the tide level.

Disclosure of Invention

Based on this, it is necessary to provide a tidal water level correction method, a target residual water level acquisition method, a device and equipment in view of the above technical problems.

A target water level acquisition method, the method comprising:

acquiring current target data of a river basin to be measured; the current target data are environment data of target data types which have influence on the residual water level;

Inputting the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data; the residual water level prediction model is a neural network model obtained through training according to historical target data and actual residual water levels of the target data type.

A tidal level correction method, the method comprising:

acquiring a target residual water level obtained according to the target residual water level acquisition method;

And correcting the initial tidal water level according to the target residual water level to obtain a corrected tidal water level.

A residual water level obtaining device, the device comprising:

The data acquisition module is used for acquiring current target data of the river basin to be detected; the current target data are environment data of target data types which have influence on the residual water level;

The target prediction module is used for inputting the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data; the residual water level prediction model is a neural network model obtained through training according to historical target data and actual residual water levels of the target data type.

A computer device comprising a memory storing a computer program and a processor which when executing the computer program performs the steps of:

A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:

The tidal water level correction method, the target residual water level acquisition device and the equipment are characterized in that the target residual water level acquisition method comprises the following steps: and acquiring environmental data of a target data type influencing the residual water level of the river basin to be measured as current target data, inputting the current target data into a residual water level prediction model obtained through the historical target data of the target data type and the actual residual water level training, and carrying out residual water level prediction to obtain the target residual water level of the current target data in a corresponding time period. Because the environmental data of the target data type has influence on the change of the residual water level, a residual water level prediction model for predicting the residual water level is obtained by adopting historical target data and actual residual water level training, the target residual water level can be accurately obtained according to input data, and the residual water level prediction model is a neural network model based on actual measured data, can accurately reflect the relation between the input data and output data, and further improves the accuracy of the output target residual water level.

Drawings

FIG. 1 is an application environment diagram of a target residual water level acquisition method in one embodiment;

FIG. 2 is a flow chart of a target water level obtaining method according to an embodiment;

FIG. 3 is a flow chart of a residual water level prediction model training process in one embodiment;

FIG. 4 is a schematic flow chart of a residual water level prediction model training process in another embodiment;

FIG. 5 is a flow chart of a target water level obtaining method according to another embodiment;

FIG. 6 is a flow chart of a residual water level prediction model training process in another embodiment;

FIG. 7 is a flow chart of a target water level obtaining method according to another embodiment;

FIG. 8 is a flow chart of a method for obtaining a target data type in one embodiment;

FIG. 9 is a flow chart of a tidal level correction method in one embodiment;

FIG. 10 is a block diagram showing a structure of a target water level obtaining apparatus according to an embodiment;

FIG. 11 is an internal block diagram of a computer device in one embodiment.

Detailed Description

The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.

The target residual water level obtaining method provided by the application can be applied to an application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The terminal 102 acquires current target data of a river basin to be detected from the server 104, wherein the current target data is environment data of a target data type which has an influence on the residual water level; the terminal 102 inputs the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data; the residual water level prediction model is a neural network model obtained through training according to historical target data and actual residual water levels of the target data type. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smartphones, tablet computers, and portable wearable devices, and the server 104 may be implemented by a stand-alone server or a server cluster composed of a plurality of servers.

In one embodiment, as shown in fig. 2, a target residual water level obtaining method is provided, and the method is applied to the terminal in fig. 1 for illustration, and includes the following steps:

s201, acquiring current target data of a river basin to be detected.

The current target data is environment data of a target data type which has an influence on the residual water level.

Further, the target data type may be weather data, such as rainfall, snowfall, etc., weather data, such as temperature, humidity, etc., and basin runoff data, such as runoff amount, runoff flow rate, runoff water level, etc.

In particular, the computer device may obtain the current target data of the target data type through a related test device/mechanism or from a storage device/database. For example, meteorological data such as precipitation, snowfall and the like of a basin to be measured and climatic data such as temperature, humidity and the like are obtained from a meteorological station of the basin to be measured, and basin runoff data such as runoff quantity, runoff flow speed, runoff water level and the like of the basin to be measured are obtained from a hydrological station of the basin to be measured.

S202, inputting the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data.

The residual water level prediction model is a neural network model obtained through training according to historical target data and actual residual water levels of the target data type.

Further, the residual water level prediction model is a neural network model which is trained based on a machine learning method and used for predicting the residual water level.

Specifically, the computer equipment adopts historical target data of the target data type and actual residual water levels of corresponding time periods to train to obtain the residual water level prediction model, and inputs the current target data into the trained residual water level prediction model to obtain the target residual water level of the corresponding time period of the current target data. For example, when the computer device obtains the current target data of month 1 in 2020, the current target data of this month is correspondingly input into the residual water level prediction model, so as to obtain the target residual water level of the basin to be measured corresponding to the time period (month 1 in 2020).

In this embodiment, the computer device obtains environmental data of a target data type having an influence on the residual water level of the basin to be measured as current target data, inputs the current target data into a residual water level prediction model obtained by training historical target data of the target data type and an actual residual water level, and performs residual water level prediction to obtain a target residual water level of a time period corresponding to the current target data. According to the method, the environmental data of the target data type with influence on the residual water level of the river basin to be measured is used as the input data of the residual water level prediction model for predicting the residual water level, and because the environmental data of the target data type has influence on the change of the residual water level, the residual water level prediction model for carrying out residual water level prediction is obtained by adopting the historical target data and the actual residual water level training, the target residual water level can be accurately obtained according to the input data, and the residual water level prediction model is a neural network model based on the measured data, so that the relation between the input data and the output data can be accurately reflected, and the accuracy of the output target residual water level is further improved.

In one embodiment, as shown in fig. 3, before the step S202 of inputting the current target data into a residual water level prediction model to obtain a target residual water level in a time period corresponding to the current target data, the step of training the actual residual water level corresponding to the time of the historical target data with the historical target data to obtain the residual water level prediction model includes the following steps:

s301, dividing the historical target data and the actual residual water level corresponding to the time of the historical target data into a training set and a testing set.

The training set and the testing set both comprise the historical target data and the actual residual water level corresponding to the time of the historical target data.

Specifically, the computer equipment acquires the historical target data obtained by actual measurement of the basin to be measured and the actual residual water level corresponding to the time of the historical target data in a past period of time. For example, the historical target data of the basin 2011 to 2015 to be measured in five years and the actual residual water level corresponding to the acquisition time of the historical target data in the five years are acquired. The computer room device further divides the historical target data and the actual residual water level corresponding to the time of the historical target data into a training set and a testing set according to a certain distribution proportion. For example, the training set and the test set are divided according to 4:1.

Also included before S301 is a data process for the history target data, the data process including:

And carrying out data interpolation on the initial historical target data according to a preset time scale to obtain the historical target data of the preset time scale.

Specifically, due to different modes and different devices for acquiring the historical target data of different types, the difference exists in time scale, and the computer device performs data interpolation on the acquired initial historical target data according to the time scale of the target residual water level to be acquired so as to acquire the historical target data which is the same as the time scale of the target residual water level to be acquired. The time scale of the historical target data and the target residual water level is unified, the consistency and fineness of the data are improved, and comprehensive data support is provided for obtaining the target residual water level.

S302, inputting the training set into an initial prediction model for training to obtain the residual water level prediction model.

Wherein the initial predictive model may be a long-short term memory (LSTM, long-short term memory) model.

Further, the LSTM model is composed of an input layer, a circulation hiding layer and an output layer, and a basic unit of the LSTM model hiding layer is a memory module. The memory module contains a cell state unit (CEC) and 3 special arithmetic units called gates. The memory module comprises 3 gate structures of an input gate, a forget gate and an output gate, which can control the information flow in the memory module. The forget gate determines which information the cell state throws out, the input gate determines which new information is stored in the cell state, and the output gate determines which information of the cell state as output. If the model input is noted as: x= (X ₁,X₂,…X_T), then the hidden layer can be obtained by the following cyclic training:

f₁＝σ(X_tU_f+S_t-1W_f+b_f)

i_t＝σ(X_tU_i+S_t-1W_i+b_i)

o_t＝σ(X_tU_o+S_t-1W_o+b_o)

Wherein W _f、W_i、W_o、W_c is the input weight matrix of the forget gate, the input gate, the output gate and the CEC respectively; u _f、U_i、U_o、U_c is the cyclic weight matrix of the forget gate, the input gate, the output gate and CEC respectively; b _f、b_i、b_o、b_c are bias vectors of the forget gate, the input gate, the output gate and the CEC, respectively; x _t represents the input of the first layer LSTM model memory module; f _t、i_t、o_t、C_t、S_t is the output of forget gate, input gate, output gate, CEC and unit at time t respectively; s _t-1 represents the output of the t-1 time unit; c _t-1 represents the output of CEC at time t-1; And Representing vector summation and vector dot product operation respectively; sigma (·) is a standard sigmoid function; tan h (·) is the hyperbolic tangent activation function.

Further, the training set may be further divided into a training subset and a verification subset, both of which also include the historical target data and the actual residual level corresponding to the historical target data time. And the computer equipment trains the initial prediction model by adopting the training subset, and verifies the trained initial prediction model by adopting the verification subset to obtain the residual water level prediction model.

Specifically, the computer equipment inputs all the training subsets into the initial prediction model for multiple times, adjusts model parameters in the initial prediction model according to the mean square error between the training residual water level obtained by the initial prediction model and the corresponding actual residual water level each time, and adjusts the model parameters once for each training to determine the model parameters in the initial prediction model, so that multiple times of training on the initial prediction model are realized. And the computer equipment inputs all the test subsets into the initial prediction model after each training, and determines whether the training of the initial prediction model is stopped according to the change condition of the mean square error between the training residual water level obtained by the initial prediction model obtained once and the corresponding actual residual water level. And when the mean square error between the training residual water level obtained according to the initial prediction model and the corresponding actual residual water level is minimum and is not reduced, taking the initial prediction model as the residual water level prediction model.

Further, the test set is adopted to test the accuracy of the residual water level prediction model, and the accuracy test process specifically comprises the following steps:

s303, inputting the historical target data in the test set into the residual water level prediction model to obtain a test residual water level corresponding to the time of the historical target data.

Specifically, the computer equipment inputs the historical target data in the test set into the residual water level prediction model, predicts the residual water level in a time period corresponding to the historical target data, and obtains the predicted residual water level. For example, the computer device inputs each of the historical target data of 2015 in the test set into the residual water level prediction model to obtain the predicted residual water level corresponding to the collection time of each of the historical target data of 2015.

S304, comparing the plurality of the test residual water levels with the plurality of the actual residual water levels in the corresponding time periods to obtain the prediction accuracy.

The prediction accuracy is the percentage of the number of times that the predicted residual water level and the actual residual water level are in an error range, which is obtained by adopting the residual water level prediction model, to the total number of times of prediction.

Specifically, the computer equipment compares a plurality of the test residual water levels with a plurality of the actual residual water levels in corresponding time periods, and obtains the percentage of the number of times that the predicted residual water level and the actual residual water level are in an error range to the number of the obtained test residual water levels (namely, the predicted total number of times) as the prediction accuracy. For example, the computer device performs 100 times of residual water level prediction by using a residual water level prediction model, wherein if 80 times of predicted residual water levels and actual residual water levels are within an error range, the prediction accuracy of the residual water level prediction model is 80%.

And S305, if the prediction accuracy is greater than or equal to a preset threshold, executing the step of inputting the current target data into a residual water level prediction model to obtain the target residual water level of the time period corresponding to the current target data.

And S306, if the prediction accuracy is smaller than the preset threshold, executing the step of receiving the new historical target data and the actual residual water level corresponding to the new historical target data time, and dividing the historical target data and the actual residual water level corresponding to the historical target data time into a training set and a testing set until the prediction accuracy is larger than or equal to the preset threshold.

The preset threshold is a preset accuracy threshold, and can be used for determining whether the residual water level prediction model can accurately predict the residual water level.

Specifically, the computer equipment compares the prediction accuracy obtained by inputting the test set into the residual water level prediction model with the preset threshold, and if the prediction accuracy is greater than or equal to the preset threshold, the computer equipment executes S202 to input the current target data into the residual water level prediction model to obtain the target residual water level of the current target data in a corresponding time period; if the prediction accuracy is smaller than the preset threshold, the computer equipment receives the new historical target data and the actual residual water level corresponding to the new historical target data time, and continues to execute S301, and the historical target data and the actual residual water level corresponding to the historical target data time are divided into a training set and a testing set.

In this embodiment, before the computer device performs the residual water level prediction using the residual water level prediction model, the computer device is trained in advance using the historical target data to obtain the residual water level prediction model for performing the residual water level prediction. The computer equipment divides the historical target data into a training set and a testing set, the training set is adopted to train and verify model parameters in the initial prediction model to obtain a residual water level prediction model, and then the testing set is adopted to test the obtained residual water level prediction model so as to further determine whether the obtained residual water level prediction model is accurately available or needs to be retrained. The residual water level prediction model obtained by training and testing the initial prediction model through the method improves the accuracy and reliability of the residual water level prediction of the to-be-detected flow area.

In one embodiment, as shown in fig. 4, the step S302 of inputting the training set into an initial prediction model to perform training to obtain the residual water level prediction model includes:

s401, dividing the historical target data in the training set into season historical data corresponding to seasons according to a season schedule according to the collection time of the historical target data.

Wherein the season history data includes spring history data, summer history data, autumn history data, and winter history data.

Further, the season schedule is used for representing the correspondence between time and season. In this embodiment, the season schedule is used to represent the correspondence between time month and season, for example, 3-5 months are spring, 6-8 months are summer, 9-11 months are autumn, and 12-2 months are winter.

Specifically, the computer device divides the month corresponding to the historical target data acquisition time in the training set into the corresponding seasonal historical data according to the corresponding relation between the month and the season in the seasonal schedule. For example, when the collection time corresponding to the historical target data in the training set is 2011-2015, the historical target data with the collection time of 3-5 months in the historical target data of five years is taken as spring historical data, the historical target data with the collection time of 3-5 months is taken as spring historical data, the historical target data with the collection time of 6-8 months is taken as summer historical data, the historical target data with the collection time of 9-11 months is taken as autumn historical data, and the historical target data with the collection time of 12-2 months is taken as winter historical data.

S402, respectively inputting the season history data in the training set into an initial prediction model for training to obtain a residual water level prediction model corresponding to seasons.

The residual water level prediction model corresponding to the season comprises a spring residual water level prediction model, a summer residual water level prediction model, an autumn residual water level prediction model and a winter residual water level prediction model.

Specifically, the computer equipment respectively inputs the spring history data, the summer history data, the autumn history data and the winter history data obtained by the division into an initial prediction model for training to obtain a residual water level prediction model corresponding to seasons. Training an initial prediction model by using the spring historical data to obtain a spring residual water level prediction model, training the initial prediction model by using the summer historical data to obtain a summer residual water level prediction model, training the initial prediction model by using the autumn historical data to obtain an autumn residual water level prediction model, and training the initial prediction model by using the winter historical data to obtain a winter residual water level prediction model.

The training set data of the corresponding seasons are adopted to conduct the training process to obtain the residual water level prediction model of the corresponding season types, and the testing set data of the corresponding seasons are adopted to conduct the accuracy testing process to obtain the accuracy testing result of the corresponding season types. For each seasonal surplus water level prediction model, the specific training process is the same as that in S302, and the specific accuracy testing process is the same as that in S303 to S306, and will not be described again.

In this embodiment, the computer device specifically divides the historical target data in the training set for training the residual water level prediction model into the season history data of the corresponding seasons according to the acquisition time, and inputs the initial residual water level prediction models respectively for training to obtain the season residual water level prediction models of the corresponding seasons, thereby realizing the refined season classification of the residual water level prediction models for residual water level prediction, and further adopting the season residual water level prediction models of the corresponding seasons for residual water level prediction, so as to improve the accuracy of residual water level prediction.

In one embodiment, as shown in fig. 5, the step S202 of inputting the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data includes:

S501, matching the acquisition time of the current target data with the season schedule, and dividing the current target data into season data types.

Wherein the season data type comprises a spring data type, a summer data type, an autumn data type and a winter data type.

Specifically, the computer device matches the month of the current target data acquisition time with the season schedule to determine which season the month of the current target data acquisition time belongs to in the season schedule, and further determines the season data type of the current target data. For example, the current target data is acquired in the period of 1 month to 4 months in 2020, the acquisition time is matched with the season schedule, the current target data acquired in the period of 1 month to 2 months in 2020 is divided into winter data types, and the current target data acquired in the period of 3 months to 4 months in 2020 is spring data types.

S502, inputting the current target data of the spring data type into the spring residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the spring data type; and/or

S503, inputting the current target data of the summer data type into the summer residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the summer data type; and/or

S504, inputting the current target data of the autumn data type into the autumn residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the autumn data type; and/or

S505, inputting the current target data of the autumn data type into the winter residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the autumn data type.

Specifically, after dividing the current target data according to the acquisition time, the computer equipment respectively inputs the obtained current target data in different seasons into a seasonal surplus water level prediction model in the corresponding season to obtain the corresponding target surplus water level. For example, the current target data with the acquisition time of 1 month to 4 months in 2020 is divided into a winter data type (1 month to 2 months) and a spring data type (3 months to 4 months) according to the acquisition time, the computer equipment inputs the current target data with the winter data type (1 month to 2 months) into a winter residual water level prediction model to obtain a target residual water level corresponding to 1 month to 2 months, and inputs the current target data with the spring data type (3 months to 4 months) into a spring residual water level prediction model to obtain a target residual water level corresponding to 3 months to 4 months.

In this embodiment, the computer device specifically divides the current target data for inputting the residual water level prediction model into current target data of corresponding seasons according to the acquisition time, so as to input the current target data of the same season into the corresponding seasonal residual water level prediction model, so as to obtain the target residual water level of the corresponding season. The current target data acquisition time span is longer, when a plurality of seasons are included, the difference between the environmental data in different seasons is larger, the influence degree on the residual water level is also larger, and if the environmental data in different seasons are treated uniformly, the residual water level prediction model is input to perform residual water level prediction, and the residual water level prediction accuracy is poor. According to the method, the current target data are subdivided into seasons according to the acquisition time, so that the divided current target data corresponding to different seasons are input into the seasonal surplus water level prediction model corresponding to the seasons, and the prediction accuracy of the surplus water level is improved.

In one embodiment, as shown in fig. 6, when the target data type includes wind speed, the step S302 of inputting the training set into an initial prediction model to perform training includes:

S601, dividing the historical target data in the training set into season historical data corresponding to seasons according to a season schedule according to the collection time of the historical target data.

The specific process is referred to S401.

S602, dividing each type of the seasonal historical data into a preset number of seasonal wind direction data according to the wind direction of the wind speed in the seasonal historical data.

The wind direction is the wind direction in the meteorological standard, and the wind direction comprises 16 wind directions. The wind speed and the wind direction can be measured by a comprehensive meteorological instrument in the meteorological field.

In particular, the computer device may divide each of the season history data into 16 types of season wind direction data according to wind direction of wind speed in the season history data. For example, the 16 types of wind direction markers are ① wind directionsWind direction, the historical spring data includes ① wind directionsThe 16 kinds of spring wind direction data of the wind direction comprise ① wind directions over-high in summer history dataThe 16 kinds of summer wind direction data of the wind direction comprise ① wind directions over-high in autumn history dataThe wind direction data of the No. 16 autumn is characterized in that the historical winter data comprises ① wind directions over the wind direction16 Kinds of winter wind direction data of the No. wind direction.

S603, inputting the seasonal wind direction data in the training set in the same season and the same wind direction corresponding time period into an initial prediction model for training, and obtaining a residual water level prediction model in the same season and the same wind direction corresponding time period.

Specifically, the computer equipment respectively inputs the seasonal wind direction data of the same season and the same wind direction obtained by the division into an initial prediction model for training, and correspondingly obtains a residual water level prediction model of the same season and the same wind direction in a corresponding time period. For example, the seasonal wind direction data (including wind speed) corresponding to the ② wind direction time period in the spring history data in the training set is input into an initial prediction model to perform training, and a residual water level prediction model corresponding to the time period is obtained.

The training set data of the corresponding seasons and the corresponding wind direction time periods are adopted to conduct the training process to obtain the residual water level prediction model of the corresponding seasons and the corresponding wind directions, and the testing set data of the corresponding seasons and the corresponding wind direction time periods are adopted to conduct the accuracy testing process to obtain the accuracy testing results of the corresponding seasons and the corresponding wind directions. The specific training process of each residual water level prediction model is the same as the content in the above step S302, and the specific accuracy testing process is the same as the content in the above steps S303 to S306, and will not be described again here.

As shown in fig. 7, S202, inputting the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data, includes:

And S701, matching the acquisition time of the current target data with the season schedule, and dividing the current target data into target season data types of the current target data.

Wherein the process of dividing the target seasonal data type is the same as the process of dividing the target seasonal data type, and the specific process is S501.

S702, acquiring a target wind direction of the wind speed in the current target data.

Specifically, the computer equipment acquires the wind direction of each wind speed in the current target data acquired by the weather synthesizer as the target wind direction. The target wind direction may be at least one of the 16 types of wind directions described above as being obtained by meteorological standards.

S703, inputting the current target data into the seasonal wind direction residual water level prediction model corresponding to the target seasonal data type and the target wind direction, and obtaining a target residual water level of a time period corresponding to the current target data.

Specifically, the computer equipment inputs the target current data corresponding to each wind direction time period under each target seasonal data type into a seasonal wind direction residual water level prediction model corresponding to the same season and the same wind direction to obtain a corresponding target residual water level. For example, the current target data with the acquisition time of 1 month to 4 months in 2020 is divided into a winter data type (1 month to 2 months) and a spring data type (3 month to 4 months) according to the acquisition time, wherein the winter data type and the spring data type are the target season data type, the 1 month wind direction in the winter data type (1 month to 2 months) is ① # wind direction, the 2 month wind direction is ② # wind direction, and the 3 month to 4 months wind direction in the spring data type (3 month to 4 months) is ⑤ # wind direction. The computer equipment inputs the current target data corresponding to the ① wind direction time period (1 month) of the winter data type (1-2 months) into a corresponding ① wind direction residual water level prediction model in winter to obtain a target residual water level corresponding to 1 month, inputs the current target data corresponding to the ② wind direction time period (2 months) of the winter data type (1-2 months) into a corresponding ② wind direction residual water level prediction model in winter to obtain a2 month corresponding target residual water level, and inputs the current target data corresponding to the ⑤ wind direction time period (3-4 months) of the spring data type (3-4 months) into a ⑤ wind direction residual water level prediction model in spring to obtain a target residual water level corresponding to 3-4 months.

In this embodiment, the computer device specifically divides the current target data for inputting the residual water level prediction model into current target data corresponding to seasons according to the collection time, and obtains the target wind direction of the wind speed in the current target data, and inputs the current target data corresponding to each target wind direction time in the same target seasonal data type into the seasonal residual water level prediction model corresponding to the seasons and corresponding wind directions. The current target data acquisition time span is longer, when a plurality of seasons are included, the difference between the environmental data in different seasons is larger, the influence degree of the current target data acquisition time span on the residual water level is also larger, the influence degree of different wind directions on the residual water level is also different, if the current target data acquisition time span is uniformly treated, the current target data acquisition time span is input into the same residual water level prediction model to predict the residual water level, and the residual water level prediction accuracy is poor. According to the method, not only is the current target data refined in seasons according to the acquisition time, but also the wind direction is further refined, so that the current target data is further refined, the further refined data is input into a further residual water level prediction model, and the prediction precision of the residual water level is further improved.

In one embodiment, as shown in fig. 8, before the step S201 of obtaining the current target data of the basin to be measured, the method includes:

s801, acquiring historical environment data of a plurality of data types of the to-be-detected river basin.

Specifically, the computer device may obtain, through a related test device/mechanism, or obtain, from a storage device/database, weather data such as precipitation, snowfall, and the like of the basin to be tested over a period of time, and weather data such as temperature, humidity, and the like, and obtain, from a hydrological station of the basin to be tested, basin runoff data such as runoff amount, runoff flow rate, runoff water level, and the like of the basin to be tested, as the historical environmental data.

S802, acquiring variable importance values of each data type for the residual water level of the basin to be tested by adopting a random forest algorithm based on the historical environment data of the plurality of data types.

Wherein said random forest algorithm is a classification method comprising a plurality of decision trees, each of said decision trees being operable to characterize one of said historical environmental data. The importance of each monitoring data on the sea surface flow speed, namely the influence degree, can be obtained according to the random forest algorithm, and the influence degree can be specifically represented by the variable importance value.

S803, sorting the obtained variable importance values from big to small to obtain a variable importance table.

The variable importance assessment is carried out by adopting a random forest algorithm, namely, how much each variable contributes to each tree in the constructed random forest model is quantified, then average values are obtained, and finally the contribution sizes among the variables are compared and ranked. In the random forest model, firstly, samples are extracted by a Bootstrap method for each tree to train, but 1/3 of data is not extracted, the data is changed into out-of-bag data OOB (out of bag), OOB is brought into a decision tree, error1 is calculated, noise interference is carried out on values corresponding to variables (feature degrees) of all samples in the OOB, namely, the values of the variables (features) are randomly changed, and then the data is brought into the decision tree, so that error2 is calculated. For N trees, the importance of variable X is the average of error2-error 1.

S804, acquiring data types corresponding to the variable importance values of at least the first two bits in the variable importance table as the target data types.

The variable importance value is used for representing the importance, namely the influence degree, of the historical environmental data of the corresponding data type on the residual water level. The historical environmental data of different data types have different degrees of influence on the sea surface flow rate.

Specifically, the computer equipment acquires variable importance values of the historical environment data of different data types for residual water levels by adopting a random forest algorithm, sorts the acquired variable importance values from large to small to obtain variable importance tables of the historical environment data of different data types for the sea table flow speed, and acquires data types of the historical environment data corresponding to at least the first two variable importance values in the variable importance tables as the target data types. The computer device may also acquire the target data type by adopting a preset variable importance threshold, for example, acquire a data type of the historical environmental data corresponding to the variable importance value greater than the variable importance threshold in the variable importance table, as the target data type.

In this embodiment, the computer device determines, as the target data type, a data type having a greater influence on the sea level flow rate in the historical environmental data by using variable importance values of the historical environmental data of different data types on the residual water level, so as to obtain the residual water level according to the data of the target data type, thereby removing the data having a smaller influence on the residual water level, reducing the amount of data required for obtaining the residual water level, and improving the data processing efficiency, so as to improve the obtaining efficiency of the residual water level obtaining method as a whole.

In one embodiment, as shown in fig. 9, there is provided a tidal water level correction method, the method comprising:

S901, acquiring a target residual water level obtained according to the target residual water level acquisition method.

S902, correcting the initial tidal water level according to the target residual water level to obtain a corrected tidal water level.

Specifically, the initial tidal water level may be a tidal water level actually measured by a tidal measurement device in the meteorological field, and the computer apparatus may correct the initial tidal water level according to the target residual water level obtained by any one of the target residual water level obtaining methods described above to obtain a corrected tidal water level. For example, the computer apparatus adds the measured initial tidal water level to the target residual water level obtained by any of the target residual water level obtaining methods described above to obtain the corrected tidal water level.

In this embodiment, the computer device corrects the initial tidal level by using the target residual water level, thereby eliminating the influence of the residual water level and the tidal water level, improving the accuracy of the obtained tidal water level, and further improving the accuracy and the accuracy of the obtained tidal water level by using the target residual water level acquisition method, wherein the accuracy of the tidal prediction information is good for ocean engineering construction and ocean scientific research.

It should be understood that, although the steps in the flowcharts of fig. 2-9 are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in fig. 2-9 may include multiple steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor does the order in which the steps or stages are performed necessarily performed in sequence, but may be performed alternately or alternately with at least a portion of the steps or stages in other steps or other steps.

In one embodiment, as shown in fig. 10, there is provided a target residual water level obtaining apparatus including: a data acquisition module 1010 and a target prediction module 1020, wherein:

The data acquisition module 1010 is configured to acquire current target data of a to-be-detected drainage basin; the current target data are environment data of target data types which have influence on the residual water level;

the target prediction module 1020 is configured to input the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data; the residual water level prediction model is a neural network model obtained by training according to historical target data and actual residual water levels of the target data type

In one embodiment, the target prediction module 1020 is further configured to:

dividing the historical target data and the actual residual water level corresponding to the time of the historical target data into a training set and a testing set; wherein the training set and the testing set both comprise the historical target data and the actual residual water level corresponding to the time of the historical target data;

inputting the training set into an initial prediction model for training to obtain the residual water level prediction model;

inputting the historical target data in the test set into the residual water level prediction model to obtain a test residual water level corresponding to the time of the historical target data;

comparing the plurality of the test residual water levels with the plurality of the actual residual water levels in the corresponding time periods to obtain prediction accuracy;

if the prediction accuracy is greater than or equal to a preset threshold, executing the step of inputting the current target data into a residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data;

And if the prediction accuracy is smaller than the preset threshold, executing the step of receiving the new historical target data and the actual residual water level corresponding to the new historical target data time, and dividing the historical target data and the actual residual water level corresponding to the historical target data time into a training set and a testing set until the prediction accuracy is larger than or equal to the preset threshold.

In one embodiment, the target prediction module 1020 is further configured to:

Dividing the historical target data in the training set into season historical data corresponding to seasons according to a season schedule according to the collection time of the historical target data; the season history data comprises spring history data, summer history data, autumn history data and winter history data;

respectively inputting the season history data in the training set into an initial prediction model for training to obtain a residual water level prediction model corresponding to seasons; the residual water level prediction model corresponding to the season comprises a spring residual water level prediction model, a summer residual water level prediction model, an autumn residual water level prediction model and a winter residual water level prediction model.

In one embodiment, the target prediction module 1020 is further configured to:

Matching the acquisition time of the current target data with the season schedule, and dividing the current target data into season data types; the season data type comprises a spring data type, a summer data type, an autumn data type and a winter data type;

Inputting the current target data of the spring data type into the spring residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the spring data type; and/or

Inputting the current target data of the summer data type into the summer residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the summer data type; and/or

Inputting the current target data of the autumn data type into the autumn residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the autumn data type; and/or

And inputting the current target data of the winter data type into the winter residual water level prediction model to obtain a target residual water level of a time period corresponding to the current target data of the winter data type.

In one embodiment, the target prediction module 1020 is further configured to:

When the target data type includes a wind speed,

Dividing the historical target data in the training set into season historical data corresponding to seasons according to a season schedule according to the collection time of the historical target data;

Dividing each type of the seasonal historical data into a preset number of seasonal wind direction data according to the wind direction of the wind speed in the seasonal historical data;

inputting the seasonal wind direction data in the training set in the same season and the same wind direction corresponding time period into an initial prediction model for training to obtain a residual water level prediction model in the same season and the same wind direction corresponding time period;

The step of inputting the current target data into a residual water level prediction model to obtain the target residual water level of the current target data in the corresponding time period comprises the following steps:

Matching the acquisition time of the current target data with the season schedule, and dividing the current target data into target season data types of the current target data;

acquiring a target wind direction of the wind speed in the current target data;

And inputting the current target data into the seasonal wind direction residual water level prediction model corresponding to the target season and the target wind direction to obtain the target residual water level of the current target data in the corresponding time period.

In one embodiment, the data acquisition module 1010 is further configured to:

Acquiring historical environment data of a plurality of data types of the to-be-detected drainage basin;

based on the historical environment data of the data types, acquiring a variable importance value of each data type for the residual water level of the basin to be detected by adopting a random forest algorithm;

sequencing the obtained variable importance values from big to small to obtain a variable importance table;

And acquiring the data type corresponding to the variable importance value of at least the first two bits in the variable importance table as the target data type.

The specific limitation of the target residual water level obtaining device may be referred to as limitation of the target residual water level obtaining method hereinabove, and will not be described herein. The above-mentioned respective modules in the target residual water level obtaining device may be realized in whole or in part by software, hardware, and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a computer device is provided, which may be a terminal, and the internal structure thereof may be as shown in fig. 11. The computer device includes a processor, a memory, a communication 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 includes a non-volatile 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 the operating system and computer programs in the non-volatile storage media. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program, when executed by a processor, implements a target residual water level acquisition 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, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.

It will be appreciated by those skilled in the art that the structure shown in FIG. 11 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the 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 stored therein a computer program, the processor when executing the computer program performing the steps of:

In one embodiment, the processor when executing the computer program further performs the steps of:

When the target data type includes a wind speed,

acquiring a target wind direction of the wind speed in the current target data;

In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:

In one embodiment, the computer program when executed by the processor further performs the steps of:

When the target data type includes a wind speed,

acquiring a target wind direction of the wind speed in the current target data;

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, or the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory. By way of illustration, and not limitation, RAM can be in various forms such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The above examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.

Claims

1. A target water level obtaining method, characterized in that the method comprises:

dividing historical target data and actual residual water levels corresponding to the time of the historical target data into a training set and a testing set; the training set and the testing set both comprise the historical target data and the actual residual water level corresponding to the time of the historical target data;

inputting the training set into an initial prediction model for training to obtain a residual water level prediction model;

If the prediction accuracy is smaller than a preset threshold, executing the step of receiving the new historical target data and the actual residual water level corresponding to the new historical target data time, and dividing the historical target data and the actual residual water level corresponding to the historical target data time into a training set and a testing set until the prediction accuracy is larger than or equal to the preset threshold;

If the prediction accuracy is greater than or equal to the preset threshold, the current target data is input into a residual water level prediction model, and a target residual water level of a time period corresponding to the current target data is obtained; the residual water level prediction model is a neural network model obtained through training according to historical target data and actual residual water levels of the target data type;

inputting the training set into an initial prediction model for training to obtain a residual water level prediction model, wherein the method comprises the following steps:

respectively inputting the season history data in the training set into an initial prediction model for training to obtain a residual water level prediction model corresponding to seasons; the residual water level prediction model corresponding to the seasons comprises a spring residual water level prediction model, a summer residual water level prediction model, an autumn residual water level prediction model and a winter residual water level prediction model.

2. The method according to claim 1, wherein the inputting the current target data into a residual water level prediction model to obtain the target residual water level of the time period corresponding to the current target data includes:

3. The method of claim 1, wherein the step of determining the position of the substrate comprises,

Before the training set is input into an initial prediction model to be trained to obtain a residual water level prediction model, the method comprises the following steps:

When the target data type includes a wind speed,

And inputting the seasonal wind direction data in the training set in the same season and the same wind direction corresponding time period into an initial prediction model for training to obtain a residual water level prediction model in the same season and the same wind direction corresponding time period.

4. The method of claim 3, wherein the inputting the current target data into a residual water level prediction model to obtain the target residual water level for the time period corresponding to the current target data comprises:

acquiring a target wind direction of the wind speed in the current target data;

And inputting the current target data into a seasonal wind direction residual water level prediction model corresponding to the target season and the target wind direction to obtain a target residual water level of a time period corresponding to the current target data.

5. The method of claim 1, wherein the initial predictive model is a long-term memory model.

6. The method according to claim 1, comprising, prior to said obtaining current target data for a basin to be measured:

7. A tidal level correction method, the method comprising:

obtaining a target residual water level according to the method of claims 1-6;

8. A target residual water level obtaining device, characterized in that the device comprises:

The target prediction module is used for dividing the historical target data and the actual residual water level corresponding to the time of the historical target data into a training set and a testing set; the training set and the testing set both comprise the historical target data and the actual residual water level corresponding to the time of the historical target data; inputting the training set into an initial prediction model for training to obtain a residual water level prediction model; inputting the historical target data in the test set into the residual water level prediction model to obtain a test residual water level corresponding to the time of the historical target data; comparing the plurality of the test residual water levels with the plurality of the actual residual water levels in the corresponding time periods to obtain prediction accuracy; if the prediction accuracy is smaller than a preset threshold, executing the step of receiving the new historical target data and the actual residual water level corresponding to the new historical target data time, and dividing the historical target data and the actual residual water level corresponding to the historical target data time into a training set and a testing set until the prediction accuracy is larger than or equal to the preset threshold; if the prediction accuracy is greater than or equal to the preset threshold, the current target data is input into a residual water level prediction model, and a target residual water level of a time period corresponding to the current target data is obtained; the residual water level prediction model is a neural network model obtained through training according to historical target data and actual residual water levels of the target data type;

Inputting the training set into an initial prediction model for training to obtain a residual water level prediction model, wherein the method comprises the following steps: dividing the historical target data in the training set into season historical data corresponding to seasons according to a season schedule according to the collection time of the historical target data; the season history data comprises spring history data, summer history data, autumn history data and winter history data; respectively inputting the season history data in the training set into an initial prediction model for training to obtain a residual water level prediction model corresponding to seasons; the residual water level prediction model corresponding to the seasons comprises a spring residual water level prediction model, a summer residual water level prediction model, an autumn residual water level prediction model and a winter residual water level prediction model.

9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 7 when the computer program is executed.

10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 7.