A sea surface wind speed super-resolution enhancement method, device, medium and product

Abstract

The application discloses a sea surface wind speed super-resolution enhancement method, device, medium and product, relates to the technical field of ocean remote sensing and application, and determines a normalized threshold of sea surface wind speed according to the mean value and the standard deviation of sea surface wind speed data; determines a mapping relationship of sea surface wind speed to a natural image according to the maximum value and the minimum value of sea surface wind speed, a preset natural image threshold and the normalized threshold; determines a target natural image after sea surface wind speed is mapped to a natural image according to the mapping relationship of sea surface wind speed to the natural image; determines a natural image after super-resolution enhancement based on a deep learning model according to the target natural image; and performs inverse mapping calculation on the natural image after super-resolution enhancement to determine sea surface wind speed after super-resolution enhancement. The application can improve the accuracy of sea surface wind speed after super-resolution enhancement.

Description

Technical Field

[0001] This application relates to the field of marine remote sensing and application technology, and in particular to a method, device, medium and product for super-resolution enhancement of sea surface wind speed. Background Technology

[0002] Sea surface wind speed is a crucial input for both atmospheric and ocean wave numerical weather prediction models. With the increasing demand for high-resolution marine environment applications, the horizontal resolution of numerical weather prediction model products (approximately 1 / 30°) is now significantly higher than the horizontal resolution of satellite sea surface wind speed (approximately 1 / 4°). Therefore, improving the horizontal resolution of satellite sea surface wind speed with high precision has become an urgent and important research direction.

[0003] Traditional business systems typically use bilinear interpolation to achieve super-resolution enhancement of sea surface wind speed. However, while interpolation can magnify images, it cannot recover high-frequency detail information. In recent years, research on super-resolution enhancement of sea surface wind speed using deep learning methods has become a hot topic, and its effectiveness compared to other super-resolution enhancement methods has been proven. However, current super-resolution enhancement of sea surface wind speed data differs from that of natural images in the following ways: First, strong wind shear is rare. Sea surface wind speed is fluid and changes much more slowly than in natural images, rarely exhibiting sharp angles or strong shear. This means that although deep learning models are constantly emerging, the improvement in accuracy is not significant enough. Second, sea surface wind speed values ​​vary widely. Sea surface wind speeds can range from 0 to 80 m / s, with most concentrated between 0 and 20 m / s. This means that sea surface wind speed data cannot be directly mapped to natural images. Considering the entire range of 0 to 80 m / s during mapping would result in underestimating the sea surface wind speed values ​​in the 0 to 20 m / s range, leading to an unbalanced image. Conversely, considering only the 0 to 20 m / s range would filter out valuable high-wind-speed data. Therefore, a method to improve the accuracy of sea surface wind speeds after super-resolution enhancement is urgently needed. Summary of the Invention

[0004] The purpose of this application is to provide a method, device, medium, and product for super-resolution enhancement of sea surface wind speed, which can solve the problem of low accuracy in calculating sea surface wind speed after super-resolution enhancement in related technologies.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] In a first aspect, this application provides a method for super-resolution enhancement of sea surface wind speed, including:

[0007] The normalized threshold for sea surface wind speed is determined based on the mean and standard deviation of sea surface wind speed data.

[0008] Based on the maximum and minimum values ​​of sea surface wind speed, a preset natural image threshold and the normalized threshold are used to determine the mapping relationship between sea surface wind speed and natural image.

[0009] Based on the mapping relationship between sea surface wind speed and natural image, determine the target natural image after mapping sea surface wind speed to natural image;

[0010] Based on the target natural image, a super-resolution enhanced natural image is determined using a deep learning model;

[0011] The super-resolution enhanced natural image is subjected to inverse mapping calculation to determine the super-resolution enhanced sea surface wind speed.

[0012] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the sea surface wind speed super-resolution enhancement method described above.

[0013] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for super-resolution enhancement of sea surface wind speed.

[0014] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for super-resolution enhancement of sea surface wind speed.

[0015] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0016] This application first determines a normalized threshold for sea surface wind speed based on the mean and standard deviation of the sea surface wind speed data. Then, based on the maximum and minimum values ​​of the sea surface wind speed, a natural image threshold and the normalized threshold are preset to determine the mapping relationship between sea surface wind speed and the natural image. Based on this mapping relationship, the target natural image after mapping the sea surface wind speed to the natural image is determined. This mapping based on the normalized threshold avoids the current situation of unbalanced natural images caused by direct mapping when strong wind shear is rare or the range of sea surface wind speed values ​​is large. Furthermore, based on a deep learning model, a super-resolution enhanced natural image is determined that can capture high-frequency details of the wind field. Finally, an inverse mapping calculation is performed on the super-resolution enhanced natural image to determine the super-resolution enhanced sea surface wind speed, ultimately improving the accuracy of the super-resolution enhanced sea surface wind speed. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a super-resolution enhancement method for sea surface wind speed provided in this application;

[0019] Figure 2 The super-resolution enhancement process for sea surface wind speed provided in this application;

[0020] Figure 3 The schematic diagram of the super-resolution enhancement principle for sea surface wind speed provided in this application;

[0021] Figure 4 Histogram of global sea surface wind speed data provided for this application;

[0022] Figure 5 A schematic diagram illustrating the normalized and direct mappings between sea surface wind speed and the grayscale of a natural image in Example 1 provided in this application;

[0023] Figure 6 A schematic diagram illustrating the normalized and direct mappings between sea surface wind speed and the grayscale of a natural image in Example 2 provided in this application;

[0024] Figure 7 A schematic diagram of the result of direct mapping of sea surface wind speed to grayscale super-resolution enhancement of natural image provided in Example 1 of this application; Figure 7 In Example (a) of the above, the true value of the sea surface wind speed is directly mapped to the data. Figure 7 Example (b) is a schematic diagram of the result of the direct mapping of the technical solution of this application in Example 1; Figure 7 (c) is a schematic diagram of the bilinear interpolation result corresponding to the direct mapping in Example 1; Figure 7 (d) is a schematic diagram of the nearest neighbor interpolation result corresponding to the direct mapping in Example 1;

[0025] Figure 8 Example 1 provided in this application is the result of grayscale super-resolution enhancement of natural images with normalized sea surface wind speed mapping; Figure 8 Example (a) shows the true sea surface wind speed corresponding to the normalized mapping in Example 1; Figure 8 Example (b) shows the result of the normalized mapping in Example 1 corresponding to the technical solution of this application; Figure 8 In example (c), the normalized mapping corresponds to the bilinear interpolation result. Figure 8 In example (d), the normalized mapping corresponds to the nearest neighbor interpolation result.

[0026] Figure 9 The sea surface wind speed in Example 2 provided in this application is directly mapped to the grayscale super-resolution enhancement result of the natural image; Figure 9 The true value of the sea surface wind speed directly mapped in Example 2 (a) is shown in the figure. Figure 9 Example (b) is a schematic diagram of the result of the direct mapping of the technical solution of this application in Example 2; Figure 9 Example (c) is a schematic diagram of the bilinear interpolation result corresponding to the direct mapping in Example 2; Figure 9 Example (d) is a schematic diagram of the nearest neighbor interpolation result corresponding to the direct mapping in Example 2;

[0027] Figure 10 The result of grayscale super-resolution enhancement of the natural image with normalized mapping of sea surface wind speed in Example 2 provided in this application; Figure 10 Example (a) shows the true sea surface wind speed corresponding to the normalized mapping in Example 2; Figure 10 Example (b) in the second case corresponds to the result of the normalized mapping in this application's technical solution; Figure 10 In example (c), the normalized mapping corresponds to the bilinear interpolation result in example two; Figure 10 In example (d), the normalized mapping corresponds to the nearest neighbor interpolation result. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] like Figure 1 As shown, this application provides a super-resolution enhancement method for sea surface wind speed, including:

[0031] Step 11: Determine the normalized threshold for sea surface wind speed based on the mean and standard deviation of the sea surface wind speed data.

[0032] In some embodiments, before step 11, the method further includes: calculating the mean and standard deviation of the sea surface wind speed data.

[0033] The mean μ is: v i (i∈[0,p]); where p is the number of sea surface wind speed data; v i This represents the i-th sea surface wind speed data.

[0034] The standard deviation σ is:

[0035] In some embodiments, step 11 specifically includes: determining a normalized threshold for sea surface wind speed based on the mean and standard deviation of the sea surface wind speed data, using the three Sigma rule of thumb. The normalized threshold is: v 3σ =μ+3σ; where, v 3σ σ is the normalization threshold; μ is the mean; σ is the standard deviation.

[0036] Step 12: Based on the maximum and minimum values ​​of sea surface wind speed, preset the natural image threshold and the normalization threshold to determine the mapping relationship between sea surface wind speed and natural image.

[0037] In some embodiments, the mapping relationship between the sea surface wind speed and the natural image is as follows:

[0038]

[0039] Among them, v max v represents the maximum wind speed at sea surface. min The minimum wind speed at sea surface; g 3σ A preset natural image threshold; v i For the i-th sea surface wind speed data; g i This represents a mapping relationship.

[0040] In some embodiments, before step 12, the method further includes: determining a preset natural image threshold corresponding to the normalization threshold based on the maximum and minimum gray values ​​of the natural image; the maximum gray value is 255; the minimum gray value is 0; and the preset natural image threshold is 200.

[0041] Step 13: Based on the mapping relationship between sea surface wind speed and natural image, determine the target natural image after mapping the sea surface wind speed to natural image.

[0042] Step 14: Based on the target natural image, determine the super-resolution enhanced natural image using a deep learning model.

[0043] In some embodiments, the deep learning model is a Fast SuperResolution using Convolutional Neural Network (FSRCNN).

[0044] Among them, the fast super-resolution convolutional neural network is also called the fast convolutional neural network super-resolution.

[0045] The deep learning model in this application can also be a super resolution using a convolutional neural network (SRCNN), a very deep convolutional network super resolution (VDSR), an efficient sub-pixel convolutional neural network (ESPCNN), or a generative adversarial network (GAN).

[0046] Step 15: Perform inverse mapping calculation on the super-resolution enhanced natural image to determine the super-resolution enhanced sea surface wind speed.

[0047] In some embodiments, step 15 specifically includes:

[0048] according to Inverse mapping calculations are performed on the target natural image to determine the sea surface wind speed after super-resolution enhancement; among which... The natural image after super-resolution enhancement; The sea surface wind speed after super-resolution enhancement; g 3σ A preset natural image threshold; v 3σ To standardize the threshold.

[0049] Where 255 is the maximum gray value of a natural image.

[0050] Specifically, spatial resolution is a key indicator for evaluating the target resolution capability of an imaging system. To improve spatial resolution, both hardware and software methods can be employed. Hardware methods involve increasing the size of the sensor or the density of photosensitive elements to improve resolution; however, frequency aliasing, manufacturing processes, and investment costs limit the advancement of hardware technology. Software methods involve obtaining higher resolution data through super-resolution enhancement methods, offering the advantage of controllable costs. However, low-resolution signals from the same imaging scene are generated by degrading high-resolution signals, losing some high-frequency information. According to Shannon's information theory sampling theorem, when the sampling frequency is less than twice the signal frequency, frequency aliasing will occur in the sampled signal. Therefore, the goal of super-resolution enhancement is to overcome the constraints of the sampling theorem and recover as much high-frequency information as possible from the acquired signal. Interpolation-based super-resolution methods assume that the value of the point to be interpolated has a linear or nonlinear relationship with the values ​​of its surrounding points, estimating the value of the point to be interpolated using known point values ​​and their relationships. Commonly used methods include nearest-neighbor interpolation and bilinear interpolation. Interpolation-based super-resolution enhancement methods are simple in principle and have low computational cost, but they cannot recover high-frequency detail information.

[0051] This application utilizes the three-sigma empirical rule in normal distribution theory to calculate the normalization threshold and proposes a cross-mapping method between sea surface wind speed and natural images. Furthermore, it leverages mature deep learning models for natural images, such as FSRCNN, for super-resolution enhancement of sea surface wind speed, significantly improving the accuracy of this enhancement. Specifically, it maps sea surface wind speed data to the grayscale range of natural images, roughly achieving grayscale balance without losing high-speed information. Then, it uses the mature FSRCNN for super-resolution enhancement, and finally calculates the final sea surface wind speed result through inverse mapping. (Refer to...) Figure 2 The specific steps of the high-resolution enhancement method for sea surface wind speed provided in this application include the following in practical applications.

[0052] The sea surface wind speed data is represented as a matrix with dimensions m×n, containing p = m×n sea surface wind speed values ​​v. i (i∈[0,p]). m is the number of rows in the matrix; n is the number of columns in the matrix.

[0053] The average wind speed at sea surface is:

[0054]

[0055] The standard deviation of sea surface wind speed is:

[0056]

[0057] Within the range of (μ-3σ, μ+3σ), the sea surface wind speed accounts for 99.73%.

[0058] Let the normalization threshold be:

[0059] v 3σ =μ+3σ (3)

[0060] When v i Less than or equal to v 3σ The sea surface wind speed is the main component of the sea surface wind speed, and is greater than v. 3σ The sea surface wind speed is high due to extreme weather events such as typhoons.

[0061] The minimum grayscale value of a natural image is 0, and the maximum grayscale value is 255. Let g be the threshold for the natural image corresponding to the normalization threshold. 3σ (g 3σ =200).

[0062] Let the maximum wind speed at the sea surface be v max The minimum value is v min The mapping relationship between sea surface wind speed and natural images is defined as follows:

[0063]

[0064] This yields a natural image mapped from sea surface wind speed. The FSRCNN deep learning model is then used to perform super-resolution enhancement on this natural image, resulting in the enhanced natural image g. i SR .

[0065] This natural image Inverse mapping calculations are performed to obtain the super-resolution enhanced sea surface wind speed.

[0066]

[0067] Received This refers to the sea surface wind speed after super-resolution enhancement.

[0068] In practical applications, such as Figure 2 As shown, the specific implementation process is as follows:

[0069] The first step is to calculate the mean and standard deviation of the sea surface wind speed using the input and sea surface wind speed data, formula (1) and formula (2).

[0070] The second step is to calculate the normalization threshold using formula (3).

[0071] The third step is to introduce a natural image threshold and use formula (4) and normalized threshold to map the sea surface wind speed to the natural image.

[0072] The fourth step involves using a deep learning model (FSRCNN) to perform super-resolution enhancement on the natural image, resulting in a super-resolution enhanced natural image; for example... Figure 3 As shown.

[0073] The fifth step involves introducing a natural image threshold and a normalization threshold to inversely map the enhanced natural image onto the sea surface wind speed, thus obtaining the super-resolution enhanced sea surface wind speed.

[0074] This application enhances the sea surface wind speed in the range of 0 to 20 m / s, which accounts for a large proportion, and assigns a certain weight to high wind speeds. The sea surface wind speed is mapped onto a natural image and super-resolution enhancement is performed using the deep learning model FSRCNN. Finally, the image is inversely mapped back to the sea surface wind speed to obtain the final super-resolution enhanced sea surface wind speed.

[0075] Furthermore, to clearly compare the root mean square error of super-resolution enhanced sea surface wind speeds obtained by different methods, the sea surface wind speeds in the following text are all accurate to three decimal places. For example... Figure 4 As shown, the calculated mean global sea surface wind speed data is 8.000 m / s, the standard deviation is 3.496 m / s, and the standardized threshold is 18.498 m / s (i.e., Figure 4(The vertical line in the image shows that) sea surface wind speeds of 20 m / s or less account for 99.86%, and sea surface wind speeds of 99.45% are less than or equal to the standardized threshold. Sea surface wind speeds of 20 m / s to 80 m / s account for a very small percentage, but they have a significant impact on the safety of navigation at sea and are marine meteorological elements that require special attention.

[0076] The validity of this application is verified below using two sets of Radarsat satellite synthetic aperture radar sea surface wind speed data:

[0077] For example, Example 1 shows a representative case of sea surface wind speed data without a typhoon. Example 2 shows a representative case of sea surface wind speed data with a typhoon. To facilitate verification of the super-resolution enhancement accuracy, the sea surface wind speeds in Examples 1 and 2 are first used as ground truth values. Then, the ground truth values ​​in Examples 1 and 2 are downsampled, reducing the resolution to 1 / 4 of the original resolution. The proposed sea surface wind speed super-resolution enhancement method is then used to super-resolution the sea surface wind speed by 4 times. Finally, the super-resolution sea surface wind speed is compared with the ground truth sea surface wind speed to verify the accuracy.

[0078] As shown in Table 1, Example 1 has no typhoon, the maximum wind speed is 33 m / s, and the normalization threshold is 17.1 m / s. Example 2 includes a typhoon, the maximum wind speed reaches 65.8 m / s, and the normalization threshold reaches 23.4 m / s. It can be seen from the two examples that the difference between the maximum wind speed and the normalization threshold is much larger than the range of sea surface wind speeds below the normalization threshold. In other words, the relatively small proportion of high wind speeds is distributed over a large range of values. Carefully handling, rather than directly removing, high-wind-speed sea surface winds is an important issue that super-resolution enhancement needs to consider.

[0079] Table 1. Maximum wind speed, minimum wind speed, and normalized threshold for Examples 1 and 2.

[0080] Example 1 Example 2 Maximum wind speed (m / s) 33 65.8 Minimum wind speed (m / s) 1.9 0.2 Normalization threshold (m / s) 17.1 23.4

[0081] Using formula (4), the normalized and direct mappings between Radarsat synthetic aperture radar sea surface wind speed and natural image grayscale can be established in Examples 1 and 2, see [link to relevant documentation]. Figure 5 and Figure 6 The normalization mapping converts sea surface wind speeds below the normalization threshold to a grayscale value range of 0 to 200 in the natural image, and converts sea surface wind speeds above the normalization threshold to a grayscale value range of 200 to 255 in the natural image. It can be seen that the normalization mapping adjusts the polyline parameters according to different implementation methods.

[0082] like Figure 7 As shown, the top left corner displays the true sea surface wind speed; the top right corner displays the super-resolution enhancement result of sea surface wind speed in this application; the bottom left corner displays the bilinear interpolation result; and the bottom right corner displays the nearest neighbor interpolation result. Similarly, Figure 8 for Figure 7The result of normalized mapping.

[0083] Overall, Figure 7 Comparison Figure 8 dark, Figure 9 Comparison Figure 10 The image appears darker than the image obtained through direct mapping, meaning the values ​​are smaller and less balanced compared to normalized mapping. This is because a few maxima in the image deviate significantly from the normalization threshold, causing the wind speed, which accounts for a larger proportion in the case of direct linear mapping, to be concentrated in the low-grayscale region. In contrast, the image obtained through normalized mapping exhibits clearer and more balanced brightness levels.

[0084] Based on the four results—the true value of sea surface wind speed, the super-resolution enhancement result of sea surface wind speed in this application, bilinear interpolation, and nearest neighbor interpolation—the true value of sea surface wind field is detailed. The super-resolution enhancement result and bilinear interpolation result of sea surface wind speed in this application are close to the true value of sea surface wind speed. However, the nearest neighbor interpolation shows a significant mosaic effect and has a large error.

[0085] The root mean square errors (RMS) of sea surface wind speed super-resolution enhancement using the proposed technical solution, bilinear interpolation, and nearest-neighbor interpolation methods are shown in Table 2. First, nearest-neighbor interpolation exhibits the largest RMS error. Under direct mapping, bilinear interpolation does not show a significant disadvantage. However, under normalized mapping, the RMS error of the proposed solution is smaller than that of bilinear interpolation, demonstrating the effectiveness of this application.

[0086] Table 2 shows the root mean square error (m / s) of sea surface wind speed super-resolution enhancement using three methods: this application, bilinear interpolation, and nearest neighbor interpolation.

[0087]

[0088] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the methods described above.

[0089] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the methods described above.

[0090] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the methods described above.

[0091] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0092] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRdM), magnetic random access memory (MRdM), ferroelectric random access memory (FRdM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RdM) or external cache memory, etc. By way of illustration and not limitation, RdM can take many forms, such as static random access memory (SRdM) or dynamic random access memory (DRdM).

[0093] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0094] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0095] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A super-resolution enhancement method for sea surface wind speed, characterized in that, include: The normalized threshold for sea surface wind speed is determined based on the mean and standard deviation of sea surface wind speed data. Based on the maximum and minimum values ​​of sea surface wind speed, a preset natural image threshold and the normalized threshold are used to determine the mapping relationship between sea surface wind speed and natural image; the mapping relationship between sea surface wind speed and natural image is as follows: ； in, This represents the maximum wind speed over the sea surface. This represents the minimum wind speed over the sea surface. A preset threshold for natural images; For the first Sea surface wind speed data; This is a mapping relationship; To standardize the threshold; Based on the mapping relationship between sea surface wind speed and natural image, determine the target natural image after mapping sea surface wind speed to natural image; Based on the target natural image, a super-resolution enhanced natural image is determined using a deep learning model; Perform inverse mapping calculations on the super-resolution enhanced natural image to determine the super-resolution enhanced sea surface wind speed, specifically including: according to Inverse mapping calculations are performed on the target natural image to determine the sea surface wind speed after super-resolution enhancement; in, The natural image after super-resolution enhancement; The sea surface wind speed after super-resolution enhancement; A preset threshold for natural images; To standardize the threshold.

2. The super-resolution enhancement method for sea surface wind speed according to claim 1, characterized in that, Before determining the normalized threshold for sea surface wind speed based on the mean and standard deviation of sea surface wind speed data, the following steps are also included: Calculate the mean and standard deviation of sea surface wind speed data; The mean for: ;in, The number of sea surface wind speed data; For the first Sea surface wind speed data ; The standard deviation for: .

3. The super-resolution enhancement method for sea surface wind speed according to claim 1, characterized in that, Based on the mean and standard deviation of sea surface wind speed data, a normalized threshold for sea surface wind speed is determined, specifically including: Based on the mean and standard deviation of sea surface wind speed data, and using the three Sigma rule of thumb, a normalized threshold for sea surface wind speed is determined. The normalization threshold is: ;in, To standardize the threshold; The mean; The standard deviation is denoted as .

4. The super-resolution enhancement method for sea surface wind speed according to claim 1, characterized in that, Before determining the mapping relationship between sea surface wind speed and natural image based on the maximum and minimum values ​​of sea surface wind speed, a preset natural image threshold, and the normalized threshold, the process further includes: Based on the maximum and minimum gray values ​​of the natural image, a preset natural image threshold corresponding to the normalization threshold is determined; the maximum gray value is 255; the minimum gray value is 0; and the preset natural image threshold is 200.

5. The super-resolution enhancement method for sea surface wind speed according to claim 3, characterized in that, The deep learning model is a fast super-resolution convolutional neural network.

6. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the sea surface wind speed super-resolution enhancement method according to any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the super-resolution enhancement method for sea surface wind speed as described in any one of claims 1-5.

8. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the super-resolution enhancement method for sea surface wind speed as described in any one of claims 1-5.

Images