Short-time wind speed prediction method and device based on satellite image information

Abstract

The invention discloses a short-time wind speed prediction method and short-time wind speed prediction equipment based on satellite image information, belongs to the technical field of meteorological prediction, and realizes high-precision and dynamic prediction of short-time wind speed by combining a satellite remote sensing image and an image feature tracking technology and setting a distance threshold adaptive to an image resolution. According to the method, the motion trend of an extreme point in a wind speed field is extracted and tracked through the SIFT algorithm, the evolution rule of the extreme point is established through regression analysis, and the complete wind speed field is reconstructed in combination with optimal interpolation, so that the limitations of insufficient resolution and obvious hysteresis of traditional numerical weather forecast in short-time local wind speed forecast are effectively overcome, and the wind speed forecast accuracy is improved. The method significantly improves the timeliness and spatial details of forecasting, and is especially suitable for offshore wind plants and other application scenes sensitive to short-time wind speed changes.

Description

Technical Field

[0001] This invention belongs to the field of meteorological forecasting technology, specifically relating to a method and device for short-term wind speed forecasting based on satellite image information. Background Technology

[0002] As a clean and renewable energy source, wind energy continues to increase its share in the global energy structure. The large-scale development of offshore and onshore wind power projects has placed higher demands on the accuracy and timeliness of wind speed forecasting. Accurate forecasting of short-term wind speeds (usually referring to the next 1-6 hours) is directly related to key aspects such as wind power dispatch optimization, wind turbine operation and maintenance safety, and grid stability, and has significant engineering practical value and economic significance.

[0003] Current wind speed forecasting technologies are mainly divided into three categories: First, conventional numerical weather prediction methods, which are based on atmospheric dynamics and thermodynamic models. However, due to limitations in model principles and initial field data, they are more suitable for forecasting mesoscale (e.g., 12 hours or more) meteorological systems and have a weak ability to capture short-term (hourly) small-scale wind speed changes, resulting in large prediction errors. Second, prediction methods based on measured data from ground observation stations and wind towers, which establish statistical models (such as ARIMA, machine learning models, etc.) using historical measured data. However, due to limitations in the density of observation station layout, it is difficult to achieve full coverage monitoring of large areas, and the accuracy of wind speed prediction is insufficient for areas without observation stations. Third, prediction methods based on satellite remote sensing data. Satellites have the advantage of large-scale and high-frequency observation, but existing technologies mostly rely on a single satellite data source or simple feature statistics, failing to fully explore the spatiotemporal evolution of wind speed-related features in satellite images, resulting in the need to improve the accuracy and reliability of short-term wind speed prediction.

[0004] In existing technologies similar to this invention, some methods construct feature sets by extracting atmospheric parameters from satellite observations and use neural network models for prediction. However, these methods rely on a large amount of historical observation data to train the model, have high requirements for data quality, and do not optimize for the spatiotemporal correlation of short-term predictions. In addition, a wind speed estimation method based on satellite cloud image texture features has been proposed, which only extracts information from cloud image textures and does not consider the dynamic evolution of wind speed feature points, resulting in limited prediction timeliness and accuracy. Summary of the Invention

[0005] The purpose of this invention is to provide a short-time wind speed prediction method and device based on satellite image information, so as to solve the technical problems of poor timeliness and low accuracy of existing short-time wind speed prediction methods.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A short-time wind speed prediction method based on satellite imagery includes the following steps: Multiple satellite wind speed inversion images of the target area were acquired and preprocessed. The SIFT feature detection algorithm was used to extract features from each preprocessed satellite wind speed inversion image to obtain the coordinates and wind speed values ​​of the extreme points in each satellite wind speed inversion image. Calculate the pixel distance between extreme points between two adjacent satellite wind speed inversion maps, compare the pixel distance with a distance threshold, perform cross-image matching, and for the successfully matched extreme points, statistically analyze the time-coordinate time series data and the time-wind speed time series data of the extreme points in multiple consecutive satellite wind speed inversion maps. Regression analysis was performed on time-coordinate time series data to obtain the location evolution curve function, and regression analysis was performed on time-wind speed time series data to obtain the wind speed change curve function. Based on the location evolution curve function and the wind speed change curve function, combined with the average time interval of multiple images, the location and wind speed of each extreme point at the next moment are predicted. Based on the wind speed prediction data of the extreme points, the predicted wind speed values ​​of the non-extreme points are estimated by weighting. By combining the predicted wind speed values ​​at extreme points and those at non-extreme points, a complete wind speed prediction map is output.

[0007] Furthermore, the preprocessing of the multiple satellite wind speed inversion maps of the target area includes: radiometric correction, geometric correction, and noise reduction.

[0008] Furthermore, the extreme points include the number of wind speed peak points and the number of wind speed valley points.

[0009] Furthermore, regression analysis is performed on the time series data, including: If the time series data exhibits a linear distribution, a linear regression model should be used for analysis. If the time series data exhibits a non-linear distribution, a multinomial regression model is used for fitting and analysis.

[0010] Furthermore, using the optimal interpolation method, the predicted wind speed at extreme points is taken as the observed value. Combined with the spatial correlation model of historical wind speed in the target area, the spatial weights of the predicted wind speeds at non-extreme points and each extreme point are calculated, and the wind speed values ​​at non-extreme points are estimated by weighting.

[0011] Furthermore, based on the comparison results between the predicted wind speed map and the actual satellite wind speed inversion map at the next moment, the distance threshold is iteratively optimized.

[0012] Furthermore, the conditions for the comparative analysis include: The root mean square error of all pixels in the predicted wind speed map and the actual satellite wind speed inversion map at the next moment; The wind speed distribution trends of the predicted wind speed map and the actual satellite wind speed inversion map at the next moment are consistent.

[0013] Secondly, the present invention provides a short-time wind speed prediction system based on satellite image information, comprising an acquisition module, a matching module, an analysis module, a prediction module, and an output module, wherein: Acquisition module: Used to acquire multiple satellite wind speed inversion maps of the target area and preprocess them. The SIFT feature detection algorithm is used to extract features from each preprocessed satellite wind speed inversion map to obtain the coordinates and wind speed values ​​of the extreme points of wind speed in each satellite wind speed inversion map; Matching module: used to calculate the pixel distance between extreme points between two adjacent satellite wind speed inversion maps, compare the pixel distance with a distance threshold, perform cross-image matching, and for successfully matched extreme points, statistically analyze the time-coordinate time series data and the time-wind speed time series data of the extreme points in multiple consecutive satellite wind speed inversion maps; Analysis module: Used for regression analysis on time-coordinate time series data to obtain location evolution curve function, and for regression analysis on time-wind speed time series data to obtain wind speed change curve function; Prediction module: Based on the location evolution curve function and the wind speed change curve function, combined with the average time interval of multiple images, it predicts the location and wind speed of each extreme point at the next moment, and estimates the predicted wind speed value of non-extreme points based on the wind speed prediction data of extreme points. Output module: Used to combine the predicted wind speed values ​​at extreme points and the predicted wind speed values ​​at non-extreme points to output a complete wind speed prediction map.

[0014] Thirdly, a terminal device is provided, 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 steps of the above-described method.

[0015] Fourthly, a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method.

[0016] Compared with the prior art, the present invention has the following beneficial technical effects: This invention presents a short-term wind speed prediction method based on satellite image information. Combining satellite remote sensing images with image feature tracking technology, it achieves high-precision and dynamic prediction of short-term wind speed by setting a distance threshold adapted to the image resolution. This method utilizes the SIFT algorithm to extract and track the movement trend of extreme points in the wind speed field, establishes its evolution law through regression analysis, and then reconstructs the complete wind speed field by combining optimal interpolation. It effectively overcomes the limitations of insufficient resolution and significant lag in traditional numerical weather prediction for short-term and local wind speed forecasts, significantly improving the timeliness and spatial detail of the forecast. It is particularly suitable for application scenarios such as offshore wind farms that are sensitive to short-term wind speed changes.

[0017] This invention employs an adaptive regression analysis-based extreme point pattern modeling scheme: based on the actual distribution characteristics of "time-location" and "time-wind speed" data, it adaptively selects linear or nonlinear regression models to accurately quantify the spatiotemporal evolution of wind speed characteristics, thereby improving the adaptability of the prediction model.

[0018] This invention uses predicted extreme points as the core observation values ​​and combines them with a historical wind speed spatial correlation model to complete the wind speed in non-extreme point areas, taking into account both local feature accuracy and global coverage, thus solving the prediction problem in areas without ground observation.

[0019] This invention uses multi-dimensional verification of predicted and actual graphs to dynamically adjust key parameters and model types, ensuring that the prediction accuracy consistently meets engineering requirements. Attached Figure Description

[0020] Figure 1 This is a flowchart of a short-time wind speed prediction method based on satellite image information in an embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0022] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] Terminology Explanation: SIFT: Scale-Invariant Feature Transform, is an algorithm for detecting and describing local features in images. It can stably identify image feature points under different scales, rotations, and lighting conditions.

[0025] Optimal interpolation: A statistical interpolation method that uses the spatial correlation and error characteristics of observed data to make the optimal estimate of values ​​for unobserved areas.

[0026] Satellite wind speed inversion map: Image data containing wind speed information of the target area, obtained by inverting satellite observation data (such as microwave, infrared and other signals).

[0027] RMSE: Root Mean Square Error, used to quantify the degree of deviation between predicted and actual values. The smaller the value, the higher the prediction accuracy.

[0028] The technical problem to be solved: (1) The problem that conventional numerical weather prediction methods are insufficient in predicting short-term, small-scale wind speeds: Conventional numerical weather prediction is based on atmospheric dynamics models, and its core advantage lies in the medium- and long-term forecasting of mesoscale meteorological systems. However, short-term wind speed changes are significantly affected by small-scale factors such as local topography and thermal convection, and existing numerical models struggle to accurately characterize these rapidly changing local features. This results in large errors in hourly short-term wind speed predictions, failing to meet the accuracy requirements of scenarios such as wind power dispatching and operational safety. Therefore, there is an urgent need for a prediction method that can directly capture the spatiotemporal evolution characteristics of small-scale wind speeds.

[0029] (2) The limited coverage and insufficient accuracy of ground observation and traditional satellite data application methods: While ground-based observation stations and wind towers provide high-precision data, their construction costs and geographical limitations make it difficult to achieve large-scale (especially in offshore and remote areas) dense deployments, resulting in a lack of effective data support for wind speed prediction in areas without observation. Existing satellite-based prediction methods often use single-point-of-time data or simple statistical analysis, failing to fully utilize the spatiotemporal continuity of satellite imagery and thus unable to accurately capture the dynamic changes in wind speed characteristics, making it difficult to improve prediction accuracy.

[0030] (3) Existing prediction methods lack an integrated solution for dynamic tracking of wind speed feature points and accurate completion of non-feature regions: Existing technologies either focus solely on extracting local feature points without systematically analyzing their spatiotemporal evolution trends, or rely on global statistical models, ignoring the dominant role of local features in short-term wind speed changes. Furthermore, wind speed estimation in non-extreme areas of satellite images often employs simple interpolation methods, failing to consider the spatial correlation and local variation patterns of wind speed, resulting in insufficient consistency and accuracy of overall prediction results.

[0031] The present invention will now be described in further detail with reference to the accompanying drawings: like Figure 1 As shown, a short-time wind speed prediction method based on satellite image information includes the following steps: Step 1: Obtain multiple satellite wind speed inversion images of the target area and preprocess them. Use the SIFT feature detection algorithm to extract features from each preprocessed satellite wind speed inversion image to obtain the coordinates and wind speed values ​​of the extreme points in each satellite wind speed inversion image. First, clearly define the target area for prediction (such as a specific offshore wind farm or onshore wind power cluster area, with a suggested coverage range of 10-200km × 10-200km). Obtain historical satellite wind speed inversion maps (wind speed image data) of this area through satellite remote sensing platforms (such as Fengyun-3, Kuihua-8, and other meteorological satellites). The requirements are as follows: ① Number of images n≥5 to ensure traceability of feature evolution patterns; ② Image time interval of 10-30 minutes to ensure the timeliness of short-term feature capture; ③ Image spatial resolution of 0.5-2km / pixel to balance accuracy and data processing efficiency.

[0032] Perform three preprocessing operations on the acquired n satellite wind speed inversion images: Radiometric correction: Eliminates the influence of satellite sensor errors and atmospheric scattering on image grayscale values, ensuring the accuracy of wind speed inversion data; Geometric correction: Based on regional geographic coordinates (such as the WGS84 coordinate system), it corrects the terrain distortion and projection error of satellite images, so that the image pixels correspond accurately to the actual geographical location; Denoising: A Gaussian filtering algorithm (with a kernel size of 3×3 or 5×5, adjusted according to noise intensity) is used to remove random noise from the image while preserving wind speed features. After preprocessing, wind speed color gradient labels (blue: 0-3 m / s, green: 3-8 m / s, red: 8-15 m / s) are uniformly added to all images to facilitate subsequent feature recognition.

[0033] Step 2: Calculate the pixel distance between extreme points between two adjacent satellite wind speed inversion maps, compare the pixel distance with the distance threshold, perform cross-image matching, and for the successfully matched extreme points, statistically analyze the time series data of the extreme points in multiple consecutive satellite wind speed inversion maps in terms of time-coordinate and time-wind speed. The SIFT feature detection algorithm was used to extract features from each preprocessed satellite wind speed inversion map, focusing on identifying "local wind speed extreme points"—that is, peak points in high wind speed areas (within the red gradient) and valley points in low wind speed areas (within the blue gradient). The geographic coordinates (x, y, in km) and corresponding wind speed value (v, in m / s) of each extreme point were recorded, and high wind speed extreme points were marked with solid red marks (e.g., P1, P2), while low wind speed extreme points were marked with solid blue marks (e.g., Q1, Q2). Then, cross-image matching is performed: a distance threshold ΔL is set (based on the image resolution, such as ΔL=3 pixels when 1km / pixel, corresponding to an actual distance of 3km). The pixel distance is calculated for extreme points in two adjacent images (such as the i-th image and the (i+1)-th image). Two extreme points with a distance < ΔL are identified as "continuation points of the same wind speed feature" (such as P1 in the i-th image and P1' in the (i+1)-th image) and marked by connecting them with a dashed line. Temporary extreme points that "have no matching points in subsequent images" (such as P3 in the i-th image having no corresponding point in the (i+1)-th image) are removed to avoid accidental wind speed changes interfering with the analysis results.

[0034] For the successfully matched stable extreme points (such as P1-P1'-P1'', Q2-Q2'-Q2''), their time series data in n consecutive images are statistically analyzed: with the image acquisition time as the horizontal axis (t, unit min), and the geographical coordinates (x, y) and wind speed (v) of the extreme points as the vertical axes, two sets of data sequences are constructed: "time-location" and "time-wind speed".

[0035] Step 3: Perform regression analysis on the time series data related to time-coordinates to obtain the location evolution curve function, and perform regression analysis on the time series data related to time-wind speed to obtain the wind speed change curve function. Regression analysis was performed on the two sets of data sequences respectively: Location evolution analysis: If the "time-location" data is linearly distributed (goodness of fit R² ≥ 0.95), a linear regression model (y = kx + b) is used to generate the location evolution curve function f(x,y,t); if it is non-linearly distributed, a multinomial regression model (y = ax² + bx + c) is used to fit the data, ensuring that the goodness of fit R² ≥ 0.95. Wind speed evolution analysis: Similarly, based on the distribution characteristics of "time-wind speed" data, an adaptive linear or nonlinear regression model is selected to generate a wind speed change curve function g(v,t), with a goodness of fit satisfying R²≥0.95. Through regression analysis, the location movement trend of extreme points and the wind speed change trend are quantified, providing a mathematical basis for subsequent predictions.

[0036] Step 4: Based on the location evolution curve function and the wind speed change curve function, and combined with the average time interval of multiple images, predict the location and wind speed of each extreme point at the next moment. Based on the wind speed prediction data of the extreme points, weighted estimate the predicted wind speed values ​​of the non-extreme points. Based on the obtained position evolution curve function f(x,y,t) and wind speed change curve function g(v,t), combined with the average time interval Δt of n images (if the time span of n images is T, then Δt=T / (n-1)), the next moment is predicted. = +Δt, i.e., the time corresponding to n+1 images, the predicted positions of each stable extreme point ( , ) and predicted wind speed ( ).

[0037] For wind speed prediction in non-extreme areas of satellite imagery, the optimal interpolation method is used: the predicted wind speed at extreme points is used as the basis for prediction. Based on the observed values, a spatial correlation model of historical wind speeds in the target area is used. The spatial correlation model is obtained by statistical analysis of historical satellite data from the past three months. The spatial weights of non-extreme points and each predicted extreme point are calculated, and the wind speed values ​​of non-extreme points are estimated by weighting.

[0038] Step 5: Combine the predicted wind speed values ​​at extreme points with those at non-extreme points to output a complete wind speed prediction map. By combining the predicted wind speed values ​​at extreme points and those at non-extreme points, a complete wind speed prediction map for the next moment is generated, achieving full-area wind speed coverage.

[0039] In one optional embodiment, after the actual satellite wind speed inversion map is obtained at the next moment, it is compared and verified with the predicted wind speed map: Single-point error analysis: Select typical extreme points (such as high wind speed point A and low wind speed point B) and calculate the difference between the predicted wind speed and the actual wind speed (such as predicted v=9.2m / s at point A, actual v=8.9m / s, error 0.3m / s). Global accuracy assessment: Calculate the root mean square error (RMSE) of all pixels in the entire image, and simultaneously calculate the zone errors of high wind speed area (8-15m / s), medium wind speed area (3-8m / s), and low wind speed area (0-3m / s); Trend consistency assessment: Evaluate the wind speed distribution trend (such as the location of high wind speed areas and the direction of wind speed gradient) between the predicted map and the actual map, and calculate the area proportion of areas with consistent trends.

[0040] If the verification results meet the condition of "RMSE≤0.5m / s and trend consistency rate>85%", the prediction result is output to complete a short-term wind speed prediction. If the condition is not met, the distance threshold (e.g., ±1 pixel) or the regression model type is adjusted for iterative optimization until the prediction accuracy meets the standard.

[0041] The present invention has the following technical effects: (1) High prediction accuracy: By using SIFT feature tracking and adaptive regression modeling, combined with closed-loop error correction, the root mean square error of short-term wind speed prediction is controlled within 0.5m / s. In particular, the prediction error in high wind speed areas (8-15m / s) can be as low as 0.35m / s, which meets the stringent requirements of wind power dispatch and wind turbine operation and maintenance for accuracy. (2) Wide coverage: Relying on the advantages of satellite imagery for wide-area observation (10-200km×10-200km), it can predict wind speed in areas without ground observation stations, such as at sea and in remote areas, thus solving the problem of insufficient coverage of traditional ground observation. (3) High timeliness: The entire process processing cycle is ≤1 hour (including data acquisition, analysis, prediction and verification), which can support short-term wind speed prediction for the next 1-6 hours and can quickly respond to real-time needs such as wind power scheduling and extreme wind speed warning; (4) Highly practical: It does not rely on complex numerical models or large-scale historical training data. It can make predictions using only n≥5 short-interval satellite images. The process is simple, the parameters are easy to adjust, and it is easy to implement and promote in engineering. It is suitable for wind power projects of different scales, meteorological early warning and other scenarios. (5) Good robustness: Through the closed-loop mechanism of removing temporary extreme points and dynamically adjusting model parameters, it can adapt to the changes in wind speed characteristics in different regions (sea / land) and different meteorological conditions (sunny / cloudy), ensuring the stability and reliability of the prediction results.

[0042] In another embodiment of the present invention, a short-time wind speed prediction system based on satellite image information is provided, comprising an acquisition module, a matching module, an analysis module, a prediction module, and an output module, wherein: Acquisition module: Used to acquire multiple satellite wind speed inversion maps of the target area and preprocess them. The SIFT feature detection algorithm is used to extract features from each preprocessed satellite wind speed inversion map to obtain the coordinates and wind speed values ​​of the extreme points of wind speed in each satellite wind speed inversion map; Matching module: used to calculate the pixel distance between extreme points between two adjacent satellite wind speed inversion maps, compare the pixel distance with a distance threshold, perform cross-image matching, and for successfully matched extreme points, statistically analyze the time-coordinate time series data and the time-wind speed time series data of the extreme points in multiple consecutive satellite wind speed inversion maps; Analysis module: Used for regression analysis on time-coordinate time series data to obtain location evolution curve function, and for regression analysis on time-wind speed time series data to obtain wind speed change curve function; Prediction module: Based on the location evolution curve function and the wind speed change curve function, combined with the average time interval of multiple images, it predicts the location and wind speed of each extreme point at the next moment, and estimates the predicted wind speed value of non-extreme points based on the wind speed prediction data of extreme points. Output module: Used to combine the predicted wind speed values ​​at extreme points and the predicted wind speed values ​​at non-extreme points to output a complete wind speed prediction map.

[0043] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can be implemented in one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs) containing computer-usable program code. The form of a computer program product implemented on ROM, optical memory, etc.

[0044] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0045] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0046] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit its scope of protection. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading the present invention, they can still make various changes, modifications or equivalent substitutions to the specific implementation of the invention, but these changes, modifications or equivalent substitutions are all within the scope of protection of the pending claims of the invention.

Claims

1. A short-time wind speed prediction method based on satellite image information, characterized by, The method comprises the following steps: obtaining a plurality of satellite wind speed retrieval maps of a target area and performing preprocessing, performing feature extraction on each of the preprocessed satellite wind speed retrieval maps by using a SIFT feature detection algorithm to obtain coordinates and wind speed values of wind speed extreme points in each of the satellite wind speed retrieval maps; calculating pixel distances between the extreme points in adjacent two satellite wind speed retrieval maps, comparing the pixel distances with a distance threshold, performing cross-image matching, and for the matched extreme points, counting time series data of the extreme points in the plurality of continuous satellite wind speed retrieval maps with respect to time-coordinates and time series data with respect to time-wind speed; performing regression analysis on the time series data with respect to time-coordinates to obtain a position evolution curve function, and performing regression analysis on the time series data with respect to time-wind speed to obtain a wind speed change curve function; based on the position evolution curve function and the wind speed change curve function, combining an average time interval of the plurality of images, predicting positions and wind speeds of the extreme points at a next time, and based on the predicted wind speed values of the extreme points, weightedly estimating predicted wind speed values of non-extreme points; integrating the predicted wind speed values of the extreme points and the predicted wind speed values of the non-extreme points to output a complete wind speed prediction map.

2. The short-time wind speed prediction method based on satellite image information according to claim 1, characterized in that, The preprocessing of the plurality of satellite wind speed retrieval maps of the target area comprises radiation correction, geometric correction and denoising processing. 3.The short-time wind speed prediction method based on satellite image information according to claim 1, characterized in that, The extreme points comprise wind speed peak points and wind speed valley points.

4. The short-time wind speed prediction method based on satellite image information according to claim 1, characterized in that, The regression analysis on the time series data comprises: if the time series data is linearly distributed, a linear regression model is used for analysis; if the time series data is nonlinearly distributed, a polynomial regression model is used for fitting analysis.

5. The short-time wind speed prediction method based on satellite image information according to claim 1, characterized in that, An optimal interpolation method is used to take the predicted wind speed of the extreme points as an observation value, combine a spatial correlation model of historical wind speed of the target area, calculate spatial weights of the non-extreme points and the predicted wind speed of each of the extreme points, and weightedly estimate the wind speed values of the non-extreme points.

6. The short-time wind speed prediction method based on satellite image information according to claim 1, characterized in that, Based on a comparison result of the predicted wind speed map and an actual satellite wind speed retrieval map at a next time, the distance threshold is iteratively optimized.

7. The short-time wind speed prediction method based on satellite image information according to claim 6, characterized in that, The comparison analysis conditions comprise: a root mean square error of all pixel points of the predicted wind speed map and the actual satellite wind speed retrieval map at the next time; consistency of wind speed distribution trends of the predicted wind speed map and the actual satellite wind speed retrieval map at the next time.

8. A short-term wind speed prediction system based on satellite image information, characterized by, The method comprises an obtaining module, a matching module, an analysis module, a prediction module and an output module, wherein: the obtaining module is configured to obtain a plurality of satellite wind speed retrieval maps of a target area and perform preprocessing, perform feature extraction on each of the preprocessed satellite wind speed retrieval maps by using a SIFT feature detection algorithm to obtain coordinates and wind speed values of wind speed extreme points in each of the satellite wind speed retrieval maps; the matching module is configured to calculate pixel distances between the extreme points in adjacent two satellite wind speed retrieval maps, compare the pixel distances with a distance threshold, perform cross-image matching, and for the matched extreme points, count time series data of the extreme points in the plurality of continuous satellite wind speed retrieval maps with respect to time-coordinates and time series data with respect to time-wind speed; The analysis module is configured to perform regression analysis on time series data about time-coordinates to obtain a position evolution curve function, and perform regression analysis on time series data about time-wind speed to obtain a wind speed change curve function; The prediction module is configured to predict the position and wind speed of each extreme point at the next moment based on the position evolution curve function and the wind speed change curve function, and the average time interval of multiple images, and estimate the predicted wind speed value of the non-extreme point based on the predicted wind speed data of the extreme point; The output module is configured to output a complete wind speed prediction map by integrating the predicted wind speed value of the extreme point and the predicted wind speed value of the non-extreme point.

9. A terminal device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor implements the computer program when the computer program is executed by the processor to realize the method of claim 1 7. The step of the method of any one.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program, which is executed by a processor, implements the method as claimed in claim 1 7. The step of the method of any one.

Images