A marine buoy data preprocessing method, system, device and medium

Abstract

The present application relates to the technical field of data processing, and particularly relates to a marine buoy data preprocessing method, system, device and medium, the method comprising: obtaining temperature abnormal sample points in buoy profile sample points; obtaining temperature value and pressure value data corresponding to one-dimensional interpolated buoy profile sample points; obtaining salinity abnormal sample points; obtaining pressure value, temperature value and salinity value data corresponding to two-dimensional interpolated buoy profile sample points. The present application realizes the correction of temperature and salinity of abnormal sample points through twice interpolation processing, so as to ensure that the finally obtained buoy data is more accurate.

Description

[0001] Technology Neighborhood

[0002] This invention relates to the field of data processing technology, specifically to a method, system, device, and medium for preprocessing marine buoy data. Background Technology

[0003] The quality of ocean buoy data can be affected by a variety of factors, including location, equipment, maintenance, and communication. Furthermore, differences in the control processes for buoy deployment and data collection among different countries further contribute to inconsistent data quality. Traditional data quality inspection methods, such as range testing, monotonicity testing, and spike and peak testing, while effective in some aspects, have limitations.

[0004] First, traditional methods struggle to comprehensively consider the impact of multiple factors on the data, thus potentially failing to fully assess data quality. Second, these methods often rely on pre-set thresholds and rules, lacking flexibility in adapting to data variations across different scenarios. Especially in marine environments, where weather and tidal factors can cause significant fluctuations in buoy data, traditional methods cannot effectively adapt to such changes, resulting in substantial errors in the acquired data.

[0005] Therefore, there is a need to provide a method, system, equipment, and medium for preprocessing marine buoy data to solve the above problems. Summary of the Invention

[0006] This invention provides a method, system, device, and medium for preprocessing marine buoy data to address the problem that existing methods, especially in marine environments, cannot effectively adapt to the large fluctuations in buoy data due to weather and tide factors, resulting in large errors in the obtained data.

[0007] The present invention provides a method for preprocessing marine buoy data, which employs the following technical solution: including:

[0008] Based on temperature values, all buoy profile sample points are classified to obtain abnormal temperature clusters;

[0009] The target abnormal temperature cluster is obtained by optimizing the abnormal temperature cluster.

[0010] Based on the pressure values ​​of all buoy profile sample points, and using the radial basis function interpolation method to perform one-dimensional interpolation on the temperature anomaly sample points, the temperature and pressure values ​​corresponding to the one-dimensional interpolated buoy profile sample points are obtained.

[0011] Salinity anomaly sample points are obtained from all buoy profile sample points after one-dimensional interpolation based on salinity values.

[0012] Based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, and using the radial basis function interpolation method to perform two-dimensional interpolation on the salinity anomaly sample points, the pressure, temperature, and salinity values ​​corresponding to the buoy profile sample points after two-dimensional interpolation are obtained.

[0013] Preferably, the step of obtaining temperature anomaly sample points from all buoy profile sample points is as follows: the DBSCAN algorithm based on density clustering is used to classify the temperature values ​​of the buoy profile sample points to obtain multiple abnormal temperature clusters, the k-nearest neighbor distance optimization algorithm is used to optimize the abnormal temperature clusters to obtain the target abnormal temperature cluster, and the sample points in the target abnormal temperature cluster are used as temperature anomaly sample points.

[0014] Preferably, the step of using radial basis function interpolation to perform one-dimensional interpolation on temperature anomaly sample points to obtain the temperature and pressure values ​​corresponding to the one-dimensional interpolated buoy profile sample points is as follows:

[0015] Based on the pressure values ​​of all buoy profile sample points, a one-dimensional vector of temperature anomaly sample points with respect to pressure is constructed.

[0016] Based on the pressure value of each temperature anomaly sample point and the pressure values ​​of all buoy profile sample points, the first distance between the temperature anomaly sample point and each buoy profile sample point is obtained, and a one-dimensional interpolation radial basis function is constructed based on the first distance.

[0017] A one-dimensional interpolation model is obtained based on the radial basis function of one-dimensional interpolation. The one-dimensional vector is then input into the one-dimensional interpolation model to obtain the interpolated temperature values ​​of the temperature anomaly sample points.

[0018] Replace the original temperature value of the temperature anomaly sample point with the interpolated temperature value to obtain the temperature and pressure data corresponding to the buoy profile sample point after one-dimensional interpolation.

[0019] Preferably, the expression for the one-dimensional interpolation model is:

[0020]

[0021]

[0022] In the formula, T(X) represents the temperature value after interpolation of the temperature anomaly sample point X;

[0023] Represents radial basis functions;

[0024] p i This represents the pressure value at the i-th buoy profile sample point;

[0025] p represents the pressure value at temperature anomaly sample point X;

[0026] ||pp i || represents the first distance between temperature anomaly sample point X and the i-th buoy profile sample point;

[0027] σ1 2 The variance of the pressure values ​​at sample points in the buoy profile;

[0028] n represents the number of sample points in the buoy profile;

[0029] w i This represents the weighting coefficient of the i-th buoy profile sample point.

[0030] Preferably, the step of obtaining salinity anomaly sample points among all buoy profile sample points after one-dimensional interpolation is as follows:

[0031] The DBSCAN algorithm based on density clustering classifies the salinity values ​​of sample points in the buoy profile to obtain salinity anomaly clusters. The k-nearest neighbor distance optimization algorithm is used to optimize the salinity anomaly clusters to obtain the target salinity anomaly cluster. The sample points in the target salinity anomaly cluster are used as salinity anomaly sample points.

[0032] Preferably, the step of using the radial basis function interpolation method to perform two-dimensional interpolation on the salinity anomaly sample points to obtain the pressure, temperature, and salinity data corresponding to the two-dimensionally interpolated buoy profile sample points is as follows:

[0033] Based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, a two-dimensional vector of salinity anomaly sample points with respect to pressure and temperature is constructed.

[0034] Based on the pressure and temperature values ​​corresponding to each salinity anomaly sample point and all buoy profile sample points, the second distance between the salinity anomaly sample point and each buoy profile sample point is obtained, and a radial basis function for two-dimensional interpolation is constructed based on the second distance.

[0035] A two-dimensional interpolation model is obtained based on the radial basis function of two-dimensional interpolation. The two-dimensional vector is input into the two-dimensional interpolation model to obtain the interpolated salinity value of the salinity anomaly sample point.

[0036] The salinity values ​​of the salinity anomaly sample points were replaced with the interpolated salinity values ​​to obtain the pressure, temperature, and salinity data corresponding to the buoy profile sample points after two-dimensional interpolation.

[0037] Preferably, the expression for the two-dimensional interpolation model is:

[0038]

[0039]

[0040] In the formula, S(X) represents the salinity value after interpolation of the salinity anomaly sample point X;

[0041] The radial basis functions represent the two-dimensional interpolation model;

[0042] q i Let represent a two-dimensional vector relating pressure and temperature at the i-th buoy profile sample point;

[0043] q represents a two-dimensional vector of the salinity anomaly sample point X with respect to pressure and temperature;

[0044] ‖qq i ‖ represents the second distance between the salinity anomaly sample point X and the i-th sample point;

[0045] σ2 2 This represents the variance of a two-dimensional vector representing the sample points in the buoy profile with respect to pressure and temperature.

[0046] n represents the number of sample points in the buoy profile;

[0047] w i This represents the weighting coefficient of the i-th buoy profile sample point.

[0048] A marine buoy data preprocessing system, comprising:

[0049] The temperature anomaly detection module is used to obtain temperature anomaly sample points in all buoy profile sample points based on temperature values.

[0050] The one-dimensional interpolation module is used to perform one-dimensional interpolation on temperature anomaly sample points based on the pressure values ​​of all buoy profile sample points and using the radial basis function interpolation method to obtain the temperature and pressure data corresponding to the one-dimensional interpolated buoy profile sample points.

[0051] The salinity anomaly detection module is used to obtain salinity anomaly sample points among all buoy profile sample points after one-dimensional interpolation based on salinity values.

[0052] It also includes a two-dimensional interpolation module, which is used to perform two-dimensional interpolation on the salinity anomaly sample points based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, and to obtain the pressure, temperature and salinity data corresponding to the buoy profile sample points after two-dimensional interpolation.

[0053] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the marine buoy data preprocessing method of the present invention.

[0054] A storage medium having a computer program stored thereon, which, when run, performs the steps of the marine buoy data preprocessing method of the present invention.

[0055] The beneficial effects of this invention are:

[0056] First, temperature anomalies are detected. Based on the pressure value, radial basis function interpolation is used to interpolate the temperature anomaly data, thus correcting the temperature of the anomaly sample points. Then, salinity anomalies are detected at the sample points, resulting in salinity anomaly sample points. Based on the pressure and temperature values, two-dimensional interpolation is used to interpolate the salinity anomaly points, thus correcting the salinity of the anomaly sample points. This allows for a multi-dimensional representation of the structure and relationships of the buoy data. In other words, this invention corrects the temperature and salinity of anomaly sample points through two interpolation processes, ensuring that the final data is more accurate. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1 This is a flowchart of a marine buoy data preprocessing method according to the present invention;

[0059] Figure 2 Pressure-temperature profiles for ocean float data prior to preprocessing;

[0060] Figure 3 The pressure-temperature curve of the ocean float data processed using the method of this invention. Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] An embodiment of the present invention, including a method, system, device, and medium for preprocessing marine buoy data, is provided, for example... Figure 1 As shown, it includes:

[0063] S1. Obtain temperature anomaly sample points from the buoy profile sample points;

[0064] Specifically, the DBSCAN algorithm based on density clustering classifies the temperature values ​​of sample points in the buoy profile to obtain multiple abnormal temperature clusters. The k-nearest neighbor distance optimization algorithm is used to optimize the abnormal temperature clusters to obtain the target abnormal temperature cluster. The sample points in the target abnormal temperature cluster are used as temperature abnormal sample points.

[0065] Step 11: Use the DBSCAN algorithm based on density clustering to classify the temperature values ​​of the buoy profile sample points to obtain abnormal temperature clusters. Specifically, first, the data density is characterized by calculating the distance between each buoy profile sample point and other buoy profile sample points. Then, the abnormal temperature clusters are divided by combining the accessibility of the density to realize the detection of abnormal temperature data. The distance between sample points is Euclidean distance, as shown in formula (1).

[0066]

[0067] The accessibility of data density is based on the ε-neighborhood and the data density threshold MinPts, where N is the ε-neighborhood. ε (x i The definition is: for a buoy profile data space of D1, x i Let x be the i-th sample point in D1. i The ε-neighborhood sample point set is x i The set of all sample points contained in the region centered at ε and with radius ε is shown in formula (2). When the sample points cluster to form a high-density region, MinPts is the minimum number of samples required in the ε neighborhood.

[0068] N ε (x i )={x j ∈D1|Dist(x i ,x j )≤ε} (2)

[0069] When the i-th sample point x i When the amount of data in the ε-neighborhood is greater than or equal to MinPts, x i This can be referred to as core data, i.e., |N ε (x i )|≥MinPts; The direct density attainable of temperature values ​​refers to if the data x i As the core data, x j ∈N ε (x i If the data x is such that it is called a 'data x', then the data x is called a 'data x'. i To x j Direct density attainability. From this, we derive the definition of temperature value density attainability: for a given data sequence (x1, x2, x3, ..., x...), ... nIf any two adjacent data points in a sequence are directly density-reachable, then x1 to x2 are said to be in a sequence with a density of at least 1. n Density can be achieved.

[0070] Therefore, this embodiment performs clustering based on the principle of decreasing density, and only divides the data into two types of clusters: normal temperature clusters and abnormal temperature clusters; it traverses from the sample points with high density, classifying the data within the ε-neighborhood of that point as the normal temperature cluster, and classifying the data outside the ε-neighborhood as the abnormal temperature cluster.

[0071] Step 12: Optimize the abnormal temperature clusters using the k-nearest neighbor distance optimization algorithm to obtain the target abnormal temperature clusters. The significant advantages of the DBSCAN algorithm are its fast clustering speed and ability to discover spatial clusters of arbitrary shapes. However, when the density of spatial clusters is highly uneven, the clustering quality is poor. The fundamental reason is the inappropriate value of the parameter ε. If ε is too small, a large portion of the data will not be clustered; while a large ε value will cause multiple clusters to be merged, resulting in a large number of data points in the same cluster. Therefore, this embodiment uses k-nearest neighbors to dynamically calculate the local radius of the abnormal temperature clusters. First, the sample points in the abnormal temperature clusters are sorted according to their distance from smallest to largest. Then, the parameter k is determined, and the top k numbers are taken. Finally, their average is taken and assigned to ε. This achieves the dynamic setting of the local radius of each abnormal temperature cluster. Thus, the target abnormal temperature cluster is obtained, and all sample points in the obtained target abnormal temperature cluster are recorded as the abnormal temperature sample points.

[0072] S2. Obtain the temperature and pressure values ​​corresponding to the sample points of the buoy profile after one-dimensional interpolation;

[0073] Specifically, the steps for obtaining the temperature and pressure values ​​corresponding to the buoy profile sample points by performing one-dimensional interpolation on the temperature anomaly sample points using radial basis function interpolation based on the pressure values ​​of all buoy profile sample points are as follows: First, construct a one-dimensional vector of the temperature anomaly sample points about pressure based on the pressure values ​​of all buoy profile sample points; second, obtain the first distance between the temperature anomaly sample point and each buoy profile sample point based on the pressure values ​​of each temperature anomaly sample point and all buoy profile sample points, and construct a one-dimensional interpolation radial basis function based on the first distance; third, obtain a one-dimensional interpolation model based on the one-dimensional interpolation radial basis function, input the one-dimensional vector into the one-dimensional interpolation model to obtain the interpolated temperature value of the temperature anomaly sample point; fourth, replace the original temperature value of the temperature anomaly sample point with the interpolated temperature value to obtain the temperature and pressure values ​​corresponding to the buoy profile sample points after one-dimensional interpolation.

[0074] In this embodiment, when the number of buoy profile sample points is n, then the i-th buoy profile sample point x i The pressure value is expressed as pi This yields a one-dimensional vector of temperature anomaly data points with respect to pressure, t(x). i ) represents the sample point x of the i-th buoy profile. i The temperature value; using the temperature value p corresponding to the current interpolation point x and the i-th buoy profile sample point x. i Temperature value p i To calculate the relationship between the current interpolation point x and the i-th buoy profile sample point x. i Distance (||pp) i If ||), then the radial basis function is denoted as The expression for the radial basis function is Equation (4). A one-dimensional interpolation model T(X) is established based on the radial basis function, making T(X) approximate t(x). In this embodiment, the Gaussian function is selected as the one-dimensional interpolation model. Therefore, the expression for the one-dimensional interpolation model T(X) is:

[0075]

[0076]

[0077] In the formula, T(X) represents the temperature value after interpolation of the temperature anomaly sample point X;

[0078] Represents radial basis functions;

[0079] p i This represents the pressure value at the i-th buoy profile sample point;

[0080] p represents the pressure value at temperature anomaly sample point X;

[0081] ||pp i || represents the first distance between temperature anomaly sample point X and the i-th buoy profile sample point;

[0082] σ1 2 The variance of the pressure values ​​at sample points in the buoy profile;

[0083] n represents the number of sample points in the buoy profile;

[0084] w i This represents the weighting coefficient of the i-th buoy profile sample point.

[0085] Based on the one-dimensional input vector of the temperature anomaly data points, a one-dimensional interpolation model is used to interpolate the temperature of the temperature anomaly data points, thereby obtaining the temperature and pressure corresponding to the buoy profile sample points after one-dimensional interpolation.

[0086] S3. Obtain salinity anomaly sample points;

[0087] Specifically, salinity anomaly sample points are obtained from all buoy profile sample points after one-dimensional interpolation based on salinity values. That is, in this embodiment, the salinity values ​​of the buoy profile sample points are classified by the DBSCAN algorithm based on density clustering to obtain salinity anomaly clusters. The k-nearest neighbor distance optimization algorithm is used to optimize the salinity anomaly clusters to obtain the target salinity anomaly cluster. The sample points in the target salinity anomaly cluster are used as salinity anomaly sample points.

[0088] S4. Obtain the pressure, temperature and salinity data corresponding to the sample points of the buoy profile after two-dimensional interpolation.

[0089] Specifically, based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, and using the radial basis function interpolation method to perform two-dimensional interpolation on the salinity anomaly sample points, the pressure, temperature, and salinity values ​​corresponding to the buoy profile sample points after two-dimensional interpolation are obtained. The step of using the radial basis function interpolation method to perform two-dimensional interpolation on the salinity anomaly sample points to obtain the two-dimensional interpolated pressure, temperature, and salinity values ​​corresponding to the buoy profile sample points is as follows: Based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, construct the salinity anomaly sample points with respect to... Two-dimensional vectors of pressure and temperature are generated. Based on the pressure and temperature values ​​corresponding to each salinity anomaly sample point and all buoy profile sample points, a second distance is obtained between the salinity anomaly sample point and each buoy profile sample point. A radial basis function for two-dimensional interpolation is constructed based on the second distance. A two-dimensional interpolation model is obtained based on the radial basis function. The two-dimensional vector is input into the two-dimensional interpolation model to obtain the interpolated salinity value of the salinity anomaly sample point. The salinity value of the salinity anomaly sample point is replaced with the interpolated salinity value to obtain the pressure, temperature, and salinity data corresponding to the two-dimensional interpolated buoy profile sample point.

[0090] In this embodiment, salinity interpolation is a multidimensional interpolation. Compared to one-dimensional interpolation, multidimensional radial basis function interpolation can capture the complex structure and relationships of ocean buoy data in multiple dimensions; a two-dimensional vector (p,t) is constructed to represent the pressure and temperature values ​​of the buoy profile sample points; s(x i ) represents the sample point x of the i-th buoy profile. i The salinity value; using the temperature value p corresponding to the current interpolation point X and the i-th buoy profile sample point x. i Temperature value p i And the pressure value is used to calculate the current interpolation point X and the i-th buoy profile sample point x. i If the distance is such that the radial basis function of the two-dimensional interpolation is denoted as , then the radial basis function of the two-dimensional interpolation is denoted as . The expression for the two-dimensional interpolation radial basis function is given by equation (6). Based on the radial basis function, a two-dimensional interpolation model S(X) is established such that S(X) approximates s(x). Then, the expression for the two-dimensional interpolation model is:

[0091]

[0092]

[0093] In the formula, S(X) represents the salinity value after interpolation of the salinity anomaly sample point X;

[0094] The radial basis functions represent the two-dimensional interpolation model;

[0095] q i Let represent a two-dimensional vector relating pressure and temperature at the i-th buoy profile sample point;

[0096] q represents a two-dimensional vector of the salinity anomaly sample point X with respect to pressure and temperature;

[0097] ‖qq i ‖ represents the second distance between the salinity anomaly sample point X and the i-th sample point;

[0098] σ2 2 This represents the variance of a two-dimensional vector representing the sample points in the buoy profile with respect to pressure and temperature.

[0099] n represents the number of sample points in the buoy profile;

[0100] w i This represents the weighting coefficient of the i-th buoy profile sample point.

[0101] The two-dimensional vectors of salinity anomaly data points with respect to pressure and temperature are input into a two-dimensional interpolation model. The two-dimensional interpolation model outputs the salinity of the salinity anomaly data points after interpolation. Therefore, by replacing the salinity of the salinity anomaly data points with the interpolated salinity, the salinity information of all buoy profile sample points can be obtained. Thus, the pressure, temperature and salinity values ​​corresponding to the buoy profile sample points can be obtained.

[0102] It should be noted that, as Figure 2 As shown, Figure 2 It is raw ocean profile data, from Figure 2 It can be seen that there are many spikes and burrs; Figure 3 This refers to the data preprocessed using the preprocessing method described in this embodiment (anomaly detection and interpolation), from... Figure 3 As can be seen, spikes and peaks in the data are well identified, and through interpolation algorithms, fluctuations and abrupt changes in the original data are eliminated, resulting in complete and smooth curves, a more reasonable data distribution trend, and a significant improvement in data quality compared to before.

[0103] A marine buoy data preprocessing system includes: a temperature anomaly detection module, a one-dimensional interpolation module, a salinity anomaly detection module, and a two-dimensional interpolation module. The temperature anomaly detection module is used to acquire temperature anomaly sample points from all buoy profile sample points based on temperature values. The one-dimensional interpolation module is used to perform one-dimensional interpolation on the temperature anomaly sample points based on the pressure values ​​of all buoy profile sample points using radial basis function interpolation, obtaining the temperature and pressure values ​​corresponding to the one-dimensional interpolated buoy profile sample points. The salinity anomaly detection module is used to acquire salinity anomaly sample points from all buoy profile sample points after one-dimensional interpolation based on salinity values. The two-dimensional interpolation module is used to perform two-dimensional interpolation on the salinity anomaly sample points based on the pressure and temperature values ​​corresponding to the one-dimensional interpolated buoy profile sample points using radial basis function interpolation, obtaining the pressure, temperature, and salinity values ​​corresponding to the two-dimensional interpolated buoy profile sample points.

[0104] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the marine buoy data preprocessing method of the present invention.

[0105] A storage medium having a computer program stored thereon, which, when run, performs the steps of the marine buoy data preprocessing method of the present invention.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for preprocessing ocean buoy data, characterized in that, include: Temperature anomaly sample points are obtained from all buoy profile sample points based on temperature values; Based on the pressure values ​​of all buoy profile sample points, and using the radial basis function interpolation method to perform one-dimensional interpolation on the temperature anomaly sample points, the temperature and pressure values ​​corresponding to the one-dimensional interpolated buoy profile sample points are obtained. Salinity anomaly sample points are obtained from all buoy profile sample points after one-dimensional interpolation based on salinity values. Based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, and using the radial basis function interpolation method to perform two-dimensional interpolation on the salinity anomaly sample points, the pressure, temperature, and salinity values ​​corresponding to the buoy profile sample points after two-dimensional interpolation are obtained.

2. The method for preprocessing ocean buoy data according to claim 1, characterized in that, The steps to obtain temperature anomaly sample points from all buoy profile sample points are as follows: The DBSCAN algorithm based on density clustering classifies the temperature values ​​of sample points in the buoy profile to obtain multiple abnormal temperature clusters. The k-nearest neighbor distance optimization algorithm is used to optimize the abnormal temperature clusters to obtain the target abnormal temperature cluster. The sample points in the target abnormal temperature cluster are used as temperature abnormal sample points.

3. The method for preprocessing ocean buoy data according to claim 1, characterized in that, The steps for using radial basis function interpolation to perform one-dimensional interpolation on temperature anomaly sample points to obtain the temperature and pressure values ​​corresponding to the one-dimensional interpolated buoy profile sample points are as follows: Based on the pressure values ​​of all buoy profile sample points, a one-dimensional vector of temperature anomaly sample points with respect to pressure is constructed. Based on the pressure value of each temperature anomaly sample point and the pressure values ​​of all buoy profile sample points, the first distance between the temperature anomaly sample point and each buoy profile sample point is obtained, and a one-dimensional interpolation radial basis function is constructed based on the first distance. A one-dimensional interpolation model is obtained based on the radial basis function of one-dimensional interpolation. The one-dimensional vector is then input into the one-dimensional interpolation model to obtain the interpolated temperature values ​​of the temperature anomaly sample points. Replace the original temperature value of the temperature anomaly sample point with the interpolated temperature value to obtain the temperature and pressure data corresponding to the buoy profile sample point after one-dimensional interpolation.

4. The method for preprocessing ocean buoy data according to claim 3, characterized in that, The expression for the one-dimensional interpolation model is: In the formula, T(X) represents the temperature value after interpolation of the temperature anomaly sample point X; Represents radial basis functions; p i pi represents the pressure value of the i-th buoy profile sample point; p represents the pressure value at temperature anomaly sample point X; ||p-p i || represents the first distance of the temperature anomaly sample point X from the i-th buoy profile sample point; σ1 2 σ1 2 σ1 2 σ1 2 σ1 2 σ1 2 σ1 < n represents the number of sample points in the buoy profile; w i wi represents the weight coefficient of the i-th buoy profile sample point.

5. The method for preprocessing ocean buoy data according to claim 1, characterized in that, The steps to obtain salinity anomaly sample points among all buoy profile sample points after one-dimensional interpolation are as follows: The DBSCAN algorithm based on density clustering classifies the salinity values ​​of sample points in the buoy profile to obtain salinity anomaly clusters. The k-nearest neighbor distance optimization algorithm is used to optimize the salinity anomaly clusters to obtain the target salinity anomaly cluster. The sample points in the target salinity anomaly cluster are used as salinity anomaly sample points.

6. The method for preprocessing ocean buoy data according to claim 1, characterized in that, The steps for using radial basis function interpolation to perform two-dimensional interpolation on salinity anomaly sample points, and obtaining the pressure, temperature, and salinity values ​​corresponding to the two-dimensional interpolated buoy profile sample points, are as follows: Based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, a two-dimensional vector of salinity anomaly sample points with respect to pressure and temperature is constructed. Based on the pressure and temperature values ​​corresponding to each salinity anomaly sample point and all buoy profile sample points, the second distance between the salinity anomaly sample point and each buoy profile sample point is obtained, and a radial basis function for two-dimensional interpolation is constructed based on the second distance. A two-dimensional interpolation model is obtained based on the radial basis function of two-dimensional interpolation. The two-dimensional vector is input into the two-dimensional interpolation model to obtain the interpolated salinity value of the salinity anomaly sample point. The salinity values ​​of the salinity anomaly sample points were replaced with the interpolated salinity values ​​to obtain the pressure, temperature, and salinity data corresponding to the buoy profile sample points after two-dimensional interpolation.

7. The method for preprocessing ocean buoy data according to claim 5, characterized in that, The expression for the two-dimensional interpolation model is: In the formula, S(X) represents the salinity value after interpolation of the salinity anomaly sample point X; The radial basis functions represent the two-dimensional interpolation model; q i a two-dimensional vector representing the i-th buoy profile sample point with respect to pressure and temperature; q represents a two-dimensional vector of the salinity anomaly sample point X with respect to pressure and temperature; ‖qq i ‖ represents the second distance between the salinity anomaly sample point X and the i-th sample point; σ2 2 This represents the variance of a two-dimensional vector representing the sample points in the buoy profile with respect to pressure and temperature. n represents the number of sample points in the buoy profile; w i This represents the weighting coefficient of the i-th buoy profile sample point.

8. A marine buoy data preprocessing system, characterized in that, include: The temperature anomaly detection module is used to obtain temperature anomaly sample points in all buoy profile sample points based on temperature values. The one-dimensional interpolation module is used to perform one-dimensional interpolation on temperature anomaly sample points based on the pressure values ​​of all buoy profile sample points and using the radial basis function interpolation method to obtain the temperature and pressure data corresponding to the one-dimensional interpolated buoy profile sample points. The salinity anomaly detection module is used to obtain salinity anomaly sample points among all buoy profile sample points after one-dimensional interpolation based on salinity values. It also includes a two-dimensional interpolation module, which is used to perform two-dimensional interpolation on the salinity anomaly sample points based on the pressure and temperature values ​​corresponding to the buoy profile sample points after one-dimensional interpolation, and to obtain the pressure, temperature and salinity data corresponding to the buoy profile sample points after two-dimensional interpolation.

9. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method as described in any one of claims 1-6.

10. A storage medium, characterized in that, It stores a computer program that, when run, performs the steps of the method according to any one of claims 1-6.

Images