CN112632868A - Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar - Google Patents

Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar Download PDF

Info

Publication number
CN112632868A
CN112632868A CN202011538074.2A CN202011538074A CN112632868A CN 112632868 A CN112632868 A CN 112632868A CN 202011538074 A CN202011538074 A CN 202011538074A CN 112632868 A CN112632868 A CN 112632868A
Authority
CN
China
Prior art keywords
sea surface
radial flow
filling
data
observation
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.)
Granted
Application number
CN202011538074.2A
Other languages
Chinese (zh)
Other versions
CN112632868B (en
Inventor
任磊
杨凌娜
卢梓君
刘李哲
苗建明
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sun Yat Sen University
Southern Marine Science and Engineering Guangdong Laboratory Zhuhai
Original Assignee
Sun Yat Sen University
Southern Marine Science and Engineering Guangdong Laboratory Zhuhai
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Sun Yat Sen University, Southern Marine Science and Engineering Guangdong Laboratory Zhuhai filed Critical Sun Yat Sen University
Priority to CN202011538074.2A priority Critical patent/CN112632868B/en
Publication of CN112632868A publication Critical patent/CN112632868A/en
Application granted granted Critical
Publication of CN112632868B publication Critical patent/CN112632868B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/08—Fluids
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • 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

Abstract

The invention discloses a filling and correcting method and a system for a high-frequency ground wave radar observation radial flow missing value, wherein the method comprises the following steps: obtaining a sea surface radial flow missing point through statistical analysis; filling the defect of the sea surface radial flow; laying unmanned boats and collecting observation data to obtain observation data; superposing the filled radial flow field diagram on the corresponding radar observation area to obtain a sea surface flow distribution diagram; revising the sea surface current distribution map according to the filling data and the observation data to obtain a revised sea surface current distribution map; and carrying out quantitative comprehensive evaluation on the revised ocean current distribution map. The system comprises: the system comprises a missing point acquisition module, a filling module, an unmanned ship observation module, a visual processing module, a revision module and an evaluation module. The method can be used for filling and correcting the missing data of the high-frequency ground wave radar. The method and the system for filling and correcting the observed radial flow missing value of the high-frequency ground wave radar can be widely applied to the field of data correction of the high-frequency ground wave radar.

Description

Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar
Technical Field
The invention belongs to the field of high-frequency ground wave radar data correction, and particularly relates to a method and a system for filling and correcting a radial flow missing value observed by a high-frequency ground wave radar.
Background
The sea surface flow data observed by the high-frequency ground waves has a missing condition, and a data set with missing values directly influences research results, so that errors and losses in research and application are finally caused. In practical application, the observed deficiency value of the high-frequency ground wave radar may influence the high-frequency ground wave radar to play a key role in emergency response such as coast rescue, oil spill treatment and the like, so that the quality, density and timeliness of hydrodynamic field information provided for marine ecological environment protection and treatment are reduced. In the process of observing radial flow by a high-frequency ground wave radar, the integrity and the accuracy of regional data are guaranteed as much as possible, but the discontinuity of ocean current space-time detection caused by external factors cannot be avoided. Therefore, when the observation data of the high-frequency ground wave radar is actually applied, effective measures need to be taken to reduce the occurrence of data missing, and the missing data needs to be filled and revised.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a method for filling and correcting a missing value of a radial flow observed by a high-frequency ground wave radar, which is helpful for comprehensively and comprehensively knowing observed data of the high-frequency ground wave radar.
The first technical scheme adopted by the invention is as follows: a filling and correcting method for a high-frequency ground wave radar observation radial flow missing value comprises the following steps:
acquiring historical data of the radial sea surface flow of the ground wave radar and carrying out statistical analysis on the historical data of the sea surface flow to obtain a missing point of the radial sea surface flow;
filling the missing point of the sea surface radial flow to obtain filling data and a filled radial flow field diagram;
laying unmanned boats according to the sea surface radial flow missing points and collecting observation data to obtain observation data;
based on the radial flow field diagram which is superposed and filled in the corresponding radar observation area by CAD, obtaining a sea surface flow distribution diagram;
revising the sea surface current distribution map according to the filling data and the observation data to obtain a revised sea surface current distribution map;
and carrying out quantitative comprehensive evaluation on the revised ocean current distribution map.
Further, the step of obtaining the historical data of the radial sea surface current of the ground wave radar and performing statistical analysis on the historical data of the sea surface current to obtain the missing point of the radial sea surface current specifically includes:
acquiring radial ocean current historical data detected by a high-frequency ground wave radar in a research area;
obtaining observation data and a radial flow field diagram according to the radial sea surface flow historical data, and establishing a spatial two-dimensional matrix of a radial flow field;
judging high-density points and low-density points according to a space two-dimensional matrix of the radial flow field and observation data;
and carrying out deletion detection on the high-density points and the low-density points according to a preset rule to obtain the sea surface radial flow field deletion points.
Further, the specific judgment rule for judging the high-density point and the low-density point according to the spatial two-dimensional matrix of the radial flow field and the observation data is as follows:
recording the historical observation period T of the ith observation point0The observation frequency in the time period is M, T0The number of effective observation data obtained in the time period is NiAnd, K is 100% × (N)i/M), taking the point with the K of more than or equal to 85 percent as a high-density point
Figure BDA0002853764020000021
The rest is low density points
Figure BDA0002853764020000022
On a time scale, when delta T is more than or equal to 0.2TkTime scale of Tk+1＝TkAt + Δ t, the high and low density points should be subdivided.
Further, the preset rule is specifically as follows:
when the number of spatio-temporal adjacent sea surface radial flow missing points is N (N is more than or equal to 3), the number is recorded at the r-thjArc of equal diameter of strip, thetajThe coordinate of a point on the radial line is (r)j，θj) Defining the coordinate as (r)j±1，θi±1) The points of (1) are adjacent points of the point, and the area formed by connecting the observation points is judged as a missing area;
when the number of continuous missing values of the same observation point in the Y time period is n (n is more than or equal to 2), the observation point is judged as a sea surface radial flow continuous missing point.
Further, the step of filling the defect point of the sea surface radial flow to obtain a filled radial flow field map specifically includes:
fitting radial flow data on a high-density point time scale to obtain a smooth curve;
filling missing values of the sea surface radial flow high-density points according to the smooth curve;
carrying out interpolation filling on missing values of low-density points of the sea surface radial flow based on a preset cascade neural network;
and obtaining the filling data and the filled radial flow field diagram.
Further, the preset cascade neural network comprises a first-stage artificial neural network and a second-stage LSTM model.
Further, according to the sea surface radial flow missing point, laying of the unmanned ship and collection of observation data are carried out, and the step of obtaining the observation data specifically comprises the following steps:
and counting the sea surface radial flow missing points, respectively arranging a preset number of unmanned boats on the missing points, and performing long-period backward survey patrol or fixed-point observation on radial branch low-density points with a resolution angle as an interval of the high-frequency ground wave radar through a background remote real-time control unmanned boat to obtain observation data.
Further, the unmanned ship is provided with an ultrasonic wind measuring sensor, an ADCP flow measuring sensor, a ship-based laser wave meter and a tide level sensor.
Further, the step of performing quantitative comprehensive evaluation on the revised ocean surface current distribution map specifically includes:
obtaining five statistical values of a correlation coefficient, a negative correlation coefficient, a root mean square error, an average absolute error and an average absolute percentage error according to filling data and observation data corresponding to the revised ocean current distribution map;
and carrying out quantitative comprehensive evaluation according to the five statistical values.
The second technical scheme adopted by the invention is as follows: the filling and correcting system for the high-frequency ground wave radar observation radial flow missing value comprises the following modules:
the missing point acquisition module is used for acquiring the historical data of the radial sea surface flow of the ground wave radar and carrying out statistical analysis on the historical data of the sea surface flow to obtain a missing point of the radial sea surface flow;
the filling module is used for filling the sea surface radial flow missing points to obtain filling data and a filled radial flow field diagram;
the unmanned ship observation module is used for laying unmanned ships according to the sea surface radial flow missing points and collecting observation data to obtain observation data;
the visualization processing module is used for obtaining a sea surface flow distribution map based on the radial flow field map which is overlapped and filled in the corresponding radar observation area by the CAD;
the revision module is used for revising the sea surface current distribution map according to the filling data and the observation data to obtain a revised sea surface current distribution map;
and the evaluation module is used for carrying out quantitative comprehensive evaluation on the revised sea surface flow distribution map.
The method and the system have the beneficial effects that: the method and the device combine the surface unmanned ship ocean current observation data to fill and revise the missing data of the high-frequency ground wave radar, solve the problems of high observation and movement cost and low precision of the filled data in the prior filling technology, and finally evaluate the filling and revising effects of the missing values, so that the filled data are more accurate.
Drawings
FIG. 1 is a flowchart illustrating steps of a method for filling and correcting missing values of observed radial flows of a high-frequency ground wave radar according to an embodiment of the present invention;
FIG. 2 is a block diagram of a system for filling and correcting missing values of observed radial flows of a high-frequency ground wave radar according to an embodiment of the present invention;
FIG. 3 is a distribution diagram of radial ocean current missing values of a high-frequency ground wave radar at a certain time according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating a predetermined cascaded neural network test procedure according to an embodiment of the present invention.
Detailed Description
The invention is described in further detail below with reference to the figures and the specific embodiments. The step numbers in the following embodiments are provided only for convenience of illustration, the order between the steps is not limited at all, and the execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.
Referring to fig. 1, the invention provides a ratio evaluation method for observing a wind wave flow field by a high-frequency ground wave radar, which comprises the following steps:
s1, acquiring historical data of the radial sea surface flow of the ground wave radar and carrying out statistical analysis on the historical data of the sea surface flow to obtain a missing point of the radial sea surface flow;
s2, filling the defect points of the sea surface radial flow to obtain filling data and a filled radial flow field diagram;
s3, laying unmanned boats according to the sea surface radial flow missing points and collecting observation data to obtain observation data;
s4, superposing and filling the radial flow field diagram in the corresponding radar observation area based on CAD to obtain a sea surface flow distribution diagram;
s5, revising the sea surface current distribution map according to the filling data and the observation data to obtain a revised sea surface current distribution map;
in particular, by using TaData such as satellites, ship loads, buoys and the like in a space range with the radius R being less than or equal to R (R is the distance between two adjacent observation points on the same radial observation line of the high-frequency ground wave radar) and the filling values obtained by the cascade network are subjected to mean value processing (if no relevant data source exists, the filling values are revision values), so that low-density points of the radial flow missing values of the high-frequency ground wave radar are revised.
And S6, carrying out quantitative comprehensive evaluation on the revised sea surface flow distribution map.
Further, as a preferred embodiment of the method, the step of obtaining historical data of the radial sea surface flow of the ground wave radar and performing statistical analysis on the historical data of the sea surface flow to obtain a point of missing the radial sea surface flow specifically includes:
acquiring radial ocean current historical data detected by a high-frequency ground wave radar in a research area;
obtaining observation data and a radial flow field diagram according to the radial sea surface flow historical data, and establishing a spatial two-dimensional matrix of a radial flow field;
judging high-density points and low-density points according to a space two-dimensional matrix of the radial flow field and observation data;
and carrying out deletion detection on the high-density points and the low-density points according to a preset rule to obtain the sea surface radial flow field deletion points.
Further, as a preferred embodiment of the method, the specific judgment rule for judging the high-density point and the low-density point according to the spatial two-dimensional matrix of the radial flow field and the observation data is as follows:
recording the historical observation period T of the ith observation point0The observation frequency in the time period is M, T0The number of effective observation data obtained in the time period is NiAnd, K is 100% × (N)i/M), taking the point with the K of more than or equal to 85 percent as a high-density point
Figure BDA0002853764020000041
The rest is low density points
Figure BDA0002853764020000042
On a time scale, when delta T is more than or equal to 0.2TkTime scale of Tk+1＝TkAt + Δ t, the high and low density points should be subdivided.
Further, the number a of radial observation branches (when the maximum sector angle is 160 ° and the high-frequency ground wave radar resolution angle is 10 °, a is 160 °/10 ° /) in the entire probe plane and the index of each observation point on the radial line (from 1 to b, the observable point on the longest radial line is b) are determined from the high-frequency ground wave radar observation resolution angle θ, and t is establishedaThe spatial two-dimensional matrix of the radial flow field at the moment can be recorded as:
Figure BDA0002853764020000043
as a preferred embodiment of the present invention, the preset rule is specifically:
when the number of spatio-temporal adjacent sea surface radial flow missing points is N (N is more than or equal to 3), the number is recorded at the r-thjArc of equal diameter of strip, thetajThe coordinate of a point on the radial line is (r)j，θj) Defining the coordinate as (r)j±1，θi±1) The points of (1) are adjacent points of the point, and the area formed by connecting the observation points is judged as a missing area;
when the number of consecutive missing values of the same observation point in the Y time period is n (n is more than or equal to 2), the observation point is judged as a sea surface radial flow consecutive missing point, and the reference is made to FIG. 3.
Further, as a preferred embodiment of the present invention, the step of filling the defect point of the sea surface radial flow to obtain a filled radial flow field map specifically includes:
fitting radial flow data on a high-density point time scale to obtain a smooth curve;
filling missing values of the sea surface radial flow high-density points according to the smooth curve;
carrying out interpolation filling on missing values of low-density points of the sea surface radial flow based on a preset cascade neural network;
and obtaining the filling data and the filled radial flow field diagram.
Further as a preferred embodiment of the present invention, the preset cascade neural network includes a first-stage artificial neural network and a second-stage LSTM model.
Specifically, since the dynamic process of ocean currents is influenced by multiple factors, different interaction relationships and information mechanisms may exist among the factor variables or indexes, a data sequence rule cannot be completely described by using a single model, and meanwhile, too many factors are considered by a single neural network, which results in complex network topology structure, slow learning rate and the like. Based on the limitations, the application provides a high-frequency ground wave radar observation data missing value filling model adopting a cascade neural network. The model is composed of two serially connected neural networks, the differently cascaded neural networks relatively independently strengthen processing of different types of data, the output of the first-stage network is used as partial input of the next-stage network, the multi-stage neural networks cooperate with each other to form complementary advantages, the generalization level is higher, and a better filling effect can be achieved.
The first-stage artificial neural network modeling specifically comprises the following procedures: (1) data cleaning: and (4) reexamining and checking the data, deleting repeated information and abnormal values, correcting existing errors and providing data consistency. (2) Normalization treatment: in order to make the optimization process of the optimal solution more gradual and make convergence easier to obtain the optimal solution, the method adopts a (0, 1) standardization method to traverse each data, records Max and Min, and performs data normalization processing by using Max-Min as a base number (Max is equal to 1, and Min is equal to 0). (3) Selecting characteristics: irrelevant features are removed, important features are selected to reduce the difficulty of learning tasks, the problem of dimension is relieved, and the efficiency of the model is improved. (4) Model structure: the timeliness and the accuracy of filling missing values by using the model are comprehensively considered, the common 3-layer artificial neural network is adopted as a model structure of a first-layer network, and the 3 layers are divided into an input layer, a hidden layer and an output layer. (5) Adjusting parameters: the neural network threshold value is determined to select the excitation function, based on the existing research, the S-shaped function has a good effect as the excitation function, and the S-shaped function is selected as the excitation function. The number of nodes of the hidden layer, the maximum number of iterations, the connection weight of the first layer and the second layer of the neural network, the learning rate of the neural network, the number of input nodes of the input layer and the number of output nodes of the output layer are all parameters needing to be considered and adjusted in the modeling process, through sensitivity testing, the correlation coefficient and the root mean square error are calculated, and a model with the maximum correlation coefficient and the minimum root mean square error is selected as an optimal model through comparison. (6) And (4) checking: and comparing the obtained preliminary data with the observation value of the unmanned ship, and selecting a model with a large correlation coefficient and a small root mean square error for the next prediction.
In addition, the second-level network in the preset cascade neural network model adopts ocean current data within a time length range of delta T before the observation time point T of the missing value (delta T is T)1) And the tide (Δ t ═ t) having a strong correlation with ocean currents2) Wave (t ═ t)3) Wind (Δ t ═ t)4) The data of the parameters and the initial filling data of the missing values obtained through the first-stage network are used as input variables, and the data sources of wind, waves, flow, tides and the like are the same as those of the first-stage network. The output value obtained by inputting the data through the second-level network is the missing filling value considering the space-time continuity of the ocean current and the interaction of the ocean current with wind, waves and tides.
Referring to fig. 4, the inspection of the entire shim model: and (3) checking the finally obtained model of the observation point by using the data observed at the point by the unmanned ship, if the check meets the requirement, the model can be directly used, and if the check does not meet the requirement, the parameter adjusting step is required to be carried out again, and the check is carried out again until the model meets the requirement.
Further, as a preferred embodiment of the method, the step of laying unmanned ships according to the sea surface radial flow missing points and collecting observation data to obtain the observation data specifically includes:
and counting the sea surface radial flow missing points, respectively arranging a preset number of unmanned boats on the missing points, and performing long-period backward survey patrol or fixed-point observation on radial branch low-density points with a resolution angle as an interval of the high-frequency ground wave radar through a background remote real-time control unmanned boat to obtain observation data.
Further as a preferred embodiment of the method, the unmanned boat is equipped with an ultrasonic anemometer sensor, an ADCP flow sensor, a ship-based laser wave meter and a tide level sensor.
Specifically, unmanned vessels are equipped with wind, wave, flow, and tide sensors: the ultrasonic wind measuring sensor is fixedly arranged right above the hull of the unmanned ship; the ADCP flow measurement sensor is arranged at the water depth of about 0.5m to 1.0m in the center of the ship bottom; the ship-based laser wave meter and the tide level sensor are arranged at the vertical water surface of the ship head, the 4 sensors are debugged respectively before being laid on the spot, test data are collected, processing and comprehensive evaluation are carried out, and the sensors are continuously debugged according to the evaluation result until the accuracy required by the observation standard is achieved. The obtained observation data of the synchronous wind element, the wave element, the flow element and the tide level element can be formatted and output through a wind sensor, a wave sensor, a flow sensor and a tide sensor which are arranged on the unmanned ship.
Further, as a preferred embodiment of the method, the step of performing quantitative comprehensive evaluation on the revised ocean surface current distribution map specifically includes:
obtaining five statistical values of a correlation coefficient, a negative correlation coefficient, a root mean square error, an average absolute error and an average absolute percentage error according to filling data and observation data corresponding to the revised ocean current distribution map;
and carrying out quantitative comprehensive evaluation according to the five statistical values.
Specifically, the formula is as follows:
correlation coefficient r:
Figure BDA0002853764020000071
complex correlation coefficient R:
Figure BDA0002853764020000072
root mean square error RMSE:
Figure BDA0002853764020000073
mean absolute error MAE:
Figure BDA0002853764020000074
mean absolute percent error MAPE:
Figure BDA0002853764020000075
in the formula, x and y are respectively ocean current data obtained by actual measurement of an unmanned ship and data obtained by filling a model based on a high-frequency ground wave radar observation ocean current data missing value of a cascade neural network,
Figure BDA0002853764020000076
respectively, the data mean values of two corresponding groups of data, m represents the data quantity of each group of data,
Figure BDA0002853764020000077
expressing the predicted value, x, of the regression equationiThe value y of the ith item in the ocean current data obtained by actual measurement of the unmanned shipiAnd the value of the ith item in the data obtained by the model for filling the missing value of the high-frequency ground wave radar observation ocean current data based on the cascade neural network is represented.
The larger the correlation coefficient R and the complex correlation coefficient R is, the smaller the root mean square error RMSE, the smaller the average absolute error MAE and the average absolute percentage error MAPE are, and the better the missing value filling and revising effects of the high-frequency ground wave radar are.
As shown in fig. 2, the system for filling and correcting missing values of radial flow observed by high-frequency ground wave radar includes the following modules:
the missing point acquisition module is used for acquiring the historical data of the radial sea surface flow of the ground wave radar and carrying out statistical analysis on the historical data of the sea surface flow to obtain a missing point of the radial sea surface flow;
the filling module is used for filling the sea surface radial flow missing points to obtain filling data and a filled radial flow field diagram;
the unmanned ship observation module is used for laying unmanned ships according to the sea surface radial flow missing points and collecting observation data to obtain observation data;
the visualization processing module is used for obtaining a sea surface flow distribution map based on the radial flow field map which is overlapped and filled in the corresponding radar observation area by the CAD;
the revision module is used for revising the sea surface current distribution map according to the filling data and the observation data to obtain a revised sea surface current distribution map;
and the evaluation module is used for carrying out quantitative comprehensive evaluation on the revised sea surface flow distribution map.
The contents in the system embodiments are all applicable to the method embodiments, the functions specifically realized by the method embodiments are the same as the system embodiments, and the beneficial effects achieved by the method embodiments are also the same as the beneficial effects achieved by the system embodiments.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims (10)

1. The filling and correcting method for the high-frequency ground wave radar observation radial flow missing value is characterized by comprising the following steps of:
acquiring historical data of the radial sea surface flow of the ground wave radar and carrying out statistical analysis on the historical data of the sea surface flow to obtain a missing point of the radial sea surface flow;
filling the missing point of the sea surface radial flow to obtain filling data and a filled radial flow field diagram;
laying unmanned boats according to the sea surface radial flow missing points and collecting observation data to obtain observation data;
based on the radial flow field diagram which is superposed and filled in the corresponding radar observation area by CAD, obtaining a sea surface flow distribution diagram;
revising the sea surface current distribution map according to the filling data and the observation data to obtain a revised sea surface current distribution map;
and carrying out quantitative comprehensive evaluation on the revised ocean current distribution map.
2. The method for filling and correcting the observed radial flow missing value of the high-frequency ground wave radar according to claim 1, wherein the step of obtaining historical data of the radial sea surface flow of the ground wave radar and performing statistical analysis on the historical data of the sea surface flow to obtain the sea surface radial flow missing point specifically comprises the steps of:
acquiring radial ocean current historical data detected by a high-frequency ground wave radar in a research area;
obtaining observation data and a radial flow field diagram according to the radial sea surface flow historical data, and establishing a spatial two-dimensional matrix of a radial flow field;
judging high-density points and low-density points according to a space two-dimensional matrix of the radial flow field and observation data;
and carrying out deletion detection on the high-density points and the low-density points according to a preset rule to obtain the sea surface radial flow field deletion points.
3. The method for filling and correcting the missing value of the high-frequency ground wave radar observation radial flow according to claim 2, wherein the specific judgment rule for judging the high-density point and the low-density point according to the spatial two-dimensional matrix of the radial flow field and the observation data is as follows:
recording the historical observation period T of the ith observation point0The observation frequency in the time period is M, T0The number of effective observation data obtained in the time period is NiAnd, K is 100% × (N)i/M), taking the point with the K of more than or equal to 85 percent as a high-density point
Figure FDA0002853764010000011
The rest is low density points
Figure FDA0002853764010000012
On a time scale, when delta T is more than or equal to 0.2TkTime scale of Tk+1＝TkAt + Δ t, the high and low density points should be subdivided.
4. The method for filling and correcting the observed radial flow missing value of the high-frequency ground wave radar according to claim 3, wherein the preset rule is specifically as follows:
when the number of spatio-temporal adjacent sea surface radial flow missing points is N (N is more than or equal to 3), the number is recorded at the r-thjArc of equal diameter of strip, thetajThe coordinate of a point on the radial line is (r)j，θj) Defining the coordinate as (r)j±1，θi±1) The points of (1) are adjacent points of the point, and the area formed by connecting the observation points is judged as a missing area;
when the number of continuous missing values of the same observation point in the Y time period is n (n is more than or equal to 2), the observation point is judged as a sea surface radial flow continuous missing point.
5. The method for filling and correcting the observed radial flow missing value of the high-frequency ground wave radar according to claim 4, wherein the step of filling the sea surface radial flow missing point to obtain the filled data and the filled radial flow field map specifically comprises:
fitting radial flow data on a high-density point time scale to obtain a smooth curve;
filling missing values of the sea surface radial flow high-density points according to the smooth curve;
carrying out interpolation filling on missing values of low-density points of the sea surface radial flow based on a preset cascade neural network;
and obtaining the filling data and the filled radial flow field diagram.
6. The method for filling and correcting the missing value of the high-frequency ground wave radar observation radial flow according to claim 5, wherein the preset cascade neural network comprises a first-stage artificial neural network and a second-stage LSTM model.
7. The method for filling and correcting the observed radial flow missing value of the high-frequency ground wave radar according to claim 6, wherein the step of laying unmanned boats according to sea surface radial flow missing points and collecting observation data to obtain observation data specifically comprises:
and counting the sea surface radial flow missing points, respectively arranging a preset number of unmanned boats on the missing points, and performing long-period backward survey patrol or fixed-point observation on radial branch low-density points with a resolution angle as an interval of the high-frequency ground wave radar through a background remote real-time control unmanned boat to obtain observation data.
8. The method for filling and correcting the observed radial flow missing value of the high-frequency ground wave radar according to claim 7, wherein the unmanned ship is provided with an ultrasonic wind measuring sensor, an ADCP flow measuring sensor, a ship-based laser wave measuring instrument and a tide level sensor.
9. The method for filling and correcting the missing value of the radial flow observed by the high-frequency ground wave radar according to claim 8, wherein the step of quantitatively and comprehensively evaluating the revised sea surface flow distribution map specifically comprises the following steps:
obtaining five statistical values of a correlation coefficient, a negative correlation coefficient, a root mean square error, an average absolute error and an average absolute percentage error according to filling data and observation data corresponding to the revised ocean current distribution map;
and carrying out quantitative comprehensive evaluation according to the five statistical values.
10. The system for filling and correcting the radial flow missing value observed by the high-frequency ground wave radar is characterized by comprising the following modules:
the missing point acquisition module is used for acquiring the historical data of the radial sea surface flow of the ground wave radar and carrying out statistical analysis on the historical data of the sea surface flow to obtain a missing point of the radial sea surface flow;
the filling module is used for filling the sea surface radial flow missing points to obtain filling data and a filled radial flow field diagram;
the unmanned ship observation module is used for laying unmanned ships according to the sea surface radial flow missing points and collecting observation data to obtain observation data;
the visualization processing module is used for obtaining a sea surface flow distribution map based on the radial flow field map which is overlapped and filled in the corresponding radar observation area by the CAD;
the revision module is used for revising the sea surface current distribution map according to the filling data and the observation data to obtain a revised sea surface current distribution map;
and the evaluation module is used for carrying out quantitative comprehensive evaluation on the revised sea surface flow distribution map.
CN202011538074.2A 2020-12-23 2020-12-23 Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar Active CN112632868B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011538074.2A CN112632868B (en) 2020-12-23 2020-12-23 Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011538074.2A CN112632868B (en) 2020-12-23 2020-12-23 Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar

Publications (2)

Publication Number Publication Date
CN112632868A true CN112632868A (en) 2021-04-09
CN112632868B CN112632868B (en) 2023-01-24

Family

ID=75321609

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011538074.2A Active CN112632868B (en) 2020-12-23 2020-12-23 Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar

Country Status (1)

Country Link
CN (1) CN112632868B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114997250A (en) * 2022-08-01 2022-09-02 北京海兰信数据科技股份有限公司 Radar radial flow post-processing method and system

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0234391A2 (en) * 1986-02-13 1987-09-02 SELENIA INDUSTRIE ELETTRONICHE ASSOCIATE S.p.A. Device for the evaluation of azimuth within radar systems, in particular for secondary radar systems with a single receiving channel
JP2011033529A (en) * 2009-08-04 2011-02-17 Japan Radio Co Ltd Radar ocean wave analysis device
CN106019257A (en) * 2016-08-05 2016-10-12 中国船舶重工集团公司第七〇九研究所 Interpolation method based on high-frequency ground-wave radar ocean current observation result space-time characteristics
CN109782247A (en) * 2019-01-28 2019-05-21 中船重工鹏力（南京）大气海洋信息系统有限公司 A method of utilizing track Information revision radar return
CN110134675A (en) * 2019-05-23 2019-08-16 大连海事大学 A kind of data cleaning method and system towards oceanographic data stream
CN111381215A (en) * 2020-03-25 2020-07-07 中国科学院地质与地球物理研究所 Phase correction method and meteor position acquisition method

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0234391A2 (en) * 1986-02-13 1987-09-02 SELENIA INDUSTRIE ELETTRONICHE ASSOCIATE S.p.A. Device for the evaluation of azimuth within radar systems, in particular for secondary radar systems with a single receiving channel
JP2011033529A (en) * 2009-08-04 2011-02-17 Japan Radio Co Ltd Radar ocean wave analysis device
CN106019257A (en) * 2016-08-05 2016-10-12 中国船舶重工集团公司第七〇九研究所 Interpolation method based on high-frequency ground-wave radar ocean current observation result space-time characteristics
CN109782247A (en) * 2019-01-28 2019-05-21 中船重工鹏力（南京）大气海洋信息系统有限公司 A method of utilizing track Information revision radar return
CN110134675A (en) * 2019-05-23 2019-08-16 大连海事大学 A kind of data cleaning method and system towards oceanographic data stream
CN111381215A (en) * 2020-03-25 2020-07-07 中国科学院地质与地球物理研究所 Phase correction method and meteor position acquisition method

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
LEI REN ET AL.: "Forecasting of Surface Currents via Correcting Wind Stress with Assimilation of High-Frequency Radar Data in a Three-Dimensional Model", 《ADVANCES IN METEOROLOGY》 *
夏文梅等: "基于VAD技术的迭代法在实测多普勒速度缺测区域的填补研究", 《气象学报》 *
赵飞等: "一种大域数据流中缺失值的填充方法", 《南京大学学报(自然科学版)》 *
黄奇华等: "基于BP神经网络的高频地波雷达海流空间插值", 《海洋学报》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114997250A (en) * 2022-08-01 2022-09-02 北京海兰信数据科技股份有限公司 Radar radial flow post-processing method and system

Also Published As

Publication number Publication date
CN112632868B (en) 2023-01-24

Similar Documents

Publication Publication Date Title
WO2022198899A1 (en) Submarine cable fault diagnosis method and apparatus, and device
CN112598113A (en) Ocean sound velocity profile acquisition method based on self-organizing competitive neural network
CN112632868B (en) Filling and correcting method and system for radial flow missing value observed by high-frequency ground wave radar
CN113240199A (en) Port ship track prediction method based on DILATE _ TLSTM
De Masi et al. Ship motion prediction by radial basis neural networks
CN115342814B (en) Unmanned ship positioning method based on multi-sensor data fusion
CN114935759A (en) Wave field missing value filling method and system based on high-frequency ground wave radar observation
CN105136126B (en) The method that tsunami ripple detecting is carried out using deep seafloor pressure data
CN113962426A (en) Offshore water navigation safety intelligent forecasting method and device
CN115660137B (en) Accurate estimation method for wind wave navigation energy consumption of ship
CN112307410A (en) Seawater temperature and salinity information time sequence prediction method based on shipborne CTD measurement data
Li et al. Ship traffic flow prediction based on AIS data mining
CN112731382B (en) Ratio evaluation method and system for high-frequency ground wave radar observation wind wave flow field
Mikulić et al. A review of artificial intelligence applications in ship structures
Ozsari Predicting main engine power and emissions for container, cargo, and tanker ships with artificial neural network analysis
CN115410419A (en) Ship mooring early warning method and system, electronic device and storage medium
CN114331042A (en) Ship collision risk analysis method
JP2018034585A (en) Analysis of propulsion performance of ship
Tsompopoulou et al. On the Evaluation of Uncertainty of AI models for Ship Powering and its effect on Power Estimates for non-ideal conditions.
CN108595484B (en) Marine atmosphere waveguide data acquisition and visualization processing method
CN116522446B (en) Method for defining main channel of estuary bay tidal current sediment
CN114722130B (en) Digital ocean submarine cable state monitoring method based on GIS system
CN116028510B (en) Ocean jump layer searching method and system
CN116499532B (en) Complex marine environment deep water pile group construction monitoring system constructed based on hydrologic model
Kwon et al. Inverse wave estimation from measured FPSO motions through artificial neural networks

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
GR01 Patent grant
GR01 Patent grant