Multi-scale sea surface temperature remote sensing data iterative correction method and related equipment thereof

Abstract

The invention relates to a multi-scale sea surface temperature remote sensing data iterative correction method. The method comprises the following steps: acquiring actually measured sea surface temperature data and first remote sensing observation sea surface temperature data of a target sea area; correcting the first remote sensing observation sea surface temperature data based on the actually measured sea surface temperature data to generate second remote sensing observation sea surface temperature data; performing time iteration correction on the second remote sensing observation sea surface temperature data based on a time correction parameter library to generate third remote sensing observation sea surface temperature data; performing spatial correction on the third remote sensing observation sea surface temperature data based on a spatial correction parameter library to generate fourth remote sensing observation sea surface temperature data; and accumulating the third remote sensing observation sea surface temperature data and the fourth remote sensing observation sea surface temperature data to obtain remote sensing observation sea surface temperature correction data corresponding to the target sea area. Based on the method, the quality and reliability of sea surface temperature remote sensing observation data can be improved.

Description

Technical Field

[0001] The present application relates to the field of marine meteorological environment protection technology, and in particular to a multi-scale sea surface temperature remote sensing data iterative correction method and related equipment. Background Art

[0002] The ocean covers more than 70% of the Earth's surface and is a key component of the global climate system. Sea Surface Temperature (SST) directly reflects the thermal energy state of the ocean and has an important impact on climate patterns, weather forecasts, and ecological monitoring. Satellite-borne infrared remote sensing inversion is the main means of obtaining SST, but its results are often affected by atmospheric effects, instrument errors, and other factors. Current SST correction methods mainly include corrections based on radiation transfer algorithms, empirical statistical model corrections, and optimal estimate corrections. These methods correct remote sensing observation data by simulating atmospheric influences, using historical data deviation laws, and solving cost functions.

[0003] However, current methods rely on limited measured data, and the revised model needs to be updated regularly to reflect new data and environmental changes, but it is often difficult to adapt to changes in actual conditions. In addition, the coefficient adjustment of the revised model is usually based on data from a fixed time period, resulting in unstable correction effects and affecting the long-term accuracy of the data. Summary of the invention

[0004] An embodiment of the present invention provides a multi-scale sea surface temperature remote sensing data iterative correction method and related equipment, which at least solves the problems of insufficient correction accuracy and difficulty in model adjustment in related technologies, and improves the quality and reliability of sea surface temperature remote sensing observation data.

[0005] According to a first aspect of an embodiment of the present invention, a method for iteratively correcting multi-scale sea surface temperature remote sensing data is provided, comprising:

[0006] Obtaining the measured sea surface temperature data and the first remote sensing observed sea surface temperature data of the target sea area;

[0007] Based on the measured sea surface temperature data, the first remote sensing observed sea surface temperature data is corrected by an inverse distance weighted interpolation method to generate second remote sensing observed sea surface temperature data;

[0008] performing time iterative correction on the second remotely sensed observed sea surface temperature data based on a time correction parameter library to generate third remotely sensed observed sea surface temperature data, wherein the time correction parameter library is constructed based on the historical reanalysis sea surface temperature data of the target sea area and the first remotely sensed observed sea surface temperature data and is used to record and extract deviation characteristics at different time scales in the first remotely sensed observed sea surface temperature data;

[0009] Performing spatial correction on the third remotely sensed observed sea surface temperature data based on a spatial correction parameter library to generate fourth remotely sensed observed sea surface temperature data, wherein the spatial correction parameter library is constructed based on the historical reanalysis sea surface temperature data and is used to store covariance characteristics of the first remotely sensed observed sea surface temperature data between different spatial positions;

[0010] The third remotely sensed sea surface temperature data and the fourth remotely sensed sea surface temperature data are cumulatively processed to obtain the remotely sensed sea surface temperature correction data corresponding to the target sea area.

[0011] According to a second aspect of an embodiment of the present invention, a multi-scale sea surface temperature remote sensing data iterative correction device is provided, comprising:

[0012] An acquisition module, used to acquire the measured sea surface temperature data of the target sea area and the first remote sensing observed sea surface temperature data;

[0013] A correction module, configured to correct the first remotely sensed observed sea surface temperature data by an inverse distance weighted interpolation method based on the measured sea surface temperature data to generate second remotely sensed observed sea surface temperature data;

[0014] The correction module is further used to perform time iterative correction on the second remotely sensed observed sea surface temperature data based on a time correction parameter library to generate third remotely sensed observed sea surface temperature data, wherein the time correction parameter library is constructed based on the historical reanalysis sea surface temperature data of the target sea area and the first remotely sensed observed sea surface temperature data, and is used to record and extract deviation characteristics at different time scales in the first remotely sensed observed sea surface temperature data;

[0015] The correction module is further used to perform spatial correction on the third remotely sensed observed sea surface temperature data based on a spatial correction parameter library to generate fourth remotely sensed observed sea surface temperature data, wherein the spatial correction parameter library is constructed based on the historical reanalysis sea surface temperature data and is used to store covariance characteristics of the first remotely sensed observed sea surface temperature data between different spatial positions;

[0016] The processing module is used to accumulate the third remotely sensed sea surface temperature data and the fourth remotely sensed sea surface temperature data to obtain the remotely sensed sea surface temperature correction data corresponding to the target sea area.

[0017] According to a third aspect of an embodiment of the present invention, there is provided an electronic device, comprising: a processor, and a memory storing a program, wherein the program comprises instructions, which, when executed by the processor, cause the processor to execute the method according to the first aspect.

[0018] According to a fourth aspect of an embodiment of the present invention, a non-transitory machine-readable medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to execute the method according to the first aspect.

[0019] The embodiment of the present invention provides an iterative correction method for multi-scale sea surface temperature remote sensing data, which can solve the problems of insufficient correction accuracy and frequent data update in the prior art. Specifically, first, measured sea surface temperature data and preliminary remote sensing observation sea surface temperature data are obtained from the target sea area, and the preliminary data is corrected using the inverse distance weighted interpolation method to generate more accurate second remote sensing observation data. Subsequently, a time correction parameter library is constructed, which combines historical reanalysis data and remote sensing data, and extracts deviation characteristics of climate scale, seasonal scale and sub-seasonal scale based on wavelet transform technology. The time correction process uses these features for iterative correction to ensure the consistency and accuracy of the data at different time scales. In terms of spatial correction, a spatial correction parameter library is constructed by a spatial sliding window algorithm, which records the covariance characteristics between different spatial positions. Based on this library, the third remote sensing observation data is spatially corrected to generate fourth remote sensing observation data that is more in line with reality. Finally, the time and space corrected data are accumulated, and after outlier detection and data smoothing, the target remote sensing observation sea surface temperature data with improved quality is finally obtained. In summary, the embodiment of the present invention realizes comprehensive and accurate correction of SST observation data through multi-scale feature fusion, significantly improving the accuracy of the correction results. At the same time, it reduces the dependence on real-time data updates and improves the overall processing efficiency and data stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other embodiments can be obtained based on these drawings without creative work.

[0021] Figure 1 A flowchart of a method for iterative correction of multi-scale sea surface temperature remote sensing data provided by an embodiment of the present invention.

[0022] Figure 2 A schematic diagram of a time iterative correction processing method provided by an embodiment of the present invention.

[0023] Figure 3 This is a diagram showing the effect of extracting time scale features of sea surface temperature through historical reanalysis based on wavelet transform according to an embodiment of the present invention.

[0024] Figure 4A flow chart of a method for synthesizing correction results of sea surface temperature observed by remote sensing provided in an embodiment of the present invention.

[0025] Figure 5 A schematic structural diagram of a multi-scale sea surface temperature remote sensing data iterative correction device provided in an embodiment of the present invention.

[0026] Figure 6 A schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] Embodiments of the present embodiment will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present embodiment are shown in the accompanying drawings, it should be understood that the present embodiment can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein, which are instead provided for a more thorough and complete understanding of the present embodiment. It should be understood that the drawings and embodiments of the present embodiment are only for exemplary purposes and are not intended to limit the scope of protection of the present embodiment.

[0028] The processing and correction of sea surface temperature (SST) data is a key step in ocean research, climate change monitoring and weather forecasting. Remote sensing observation technology is an important means to obtain large-scale sea surface temperature information, but due to factors such as sensor accuracy, cloud interference, and atmospheric conditions, remote sensing SST data usually have errors. In addition, the variation characteristics at different time and space scales, as well as local environmental factors, may also cause deviations between the observed data and the actual situation.

[0029] In order to improve the accuracy of remote sensing SST data, data correction methods are usually used. Existing technologies mainly perform data fusion or interpolation correction based on historical reanalysis data, measured data and physical models. However, these methods often have limitations in dealing with deviation characteristics at different time and spatial scales, and cannot fully capture small-scale periodic changes or cross-regional spatial synergy. Therefore, the existing correction methods are insufficient in capturing and processing multi-scale features, resulting in certain deviations in the corrected SST data, which affects the prediction accuracy of ocean and climate models.

[0030] In order to solve the problems of insufficient correction accuracy and frequent data updates of sea surface temperature (SST) data existing in the prior art, an embodiment of the present invention provides an iterative correction method for multi-scale sea surface temperature remote sensing data. Specifically, by constructing a time and space correction parameter library, combining multi-scale feature fusion and iterative correction mechanism, the accuracy of remote sensing sea surface temperature (SST) data is significantly improved. The time correction parameter library extracts deviation characteristics of different time scales, the space correction parameter library captures the covariance characteristics of different locations, and the inverse distance weighted interpolation method is used to preliminarily correct the data. The iterative correction mechanism ensures the gradual optimization of the deviation, and ultimately enhances the stability and accuracy of the data through cumulative processing, making the application of the corrected SST data in ocean and climate forecasting more reliable.

[0031] Figure 1 A flowchart of a multi-scale sea surface temperature remote sensing data iterative correction method provided by an embodiment of the present invention. Figure 1 As shown, the method includes the following steps.

[0032] Step S101, obtaining the measured sea surface temperature data and the first remote sensing observed sea surface temperature data of the target sea area.

[0033] Step S102, based on the measured sea surface temperature data, the first remote sensing observed sea surface temperature data is corrected by an inverse distance weighted interpolation method to generate second remote sensing observed sea surface temperature data.

[0034] Step S103, performing time iterative correction on the second remotely sensed sea surface temperature data based on the time correction parameter library to generate third remotely sensed sea surface temperature data. The time correction parameter library is constructed based on the historical reanalysis sea surface temperature data of the target sea area and the first remotely sensed sea surface temperature data, and is used to record and extract deviation characteristics at different time scales in the first remotely sensed sea surface temperature data.

[0035] Step S104, spatially correcting the third remotely sensed observed sea surface temperature data based on a spatial correction parameter library to generate fourth remotely sensed observed sea surface temperature data, wherein the spatial correction parameter library is constructed based on historical reanalysis sea surface temperature data and is used to store the covariance characteristics of the first remotely sensed observed sea surface temperature data between different spatial positions.

[0036] Step S105, accumulating the third remote sensing observed sea surface temperature data and the fourth remote sensing observed sea surface temperature data to obtain the remote sensing observed sea surface temperature correction data corresponding to the target sea area.

[0037] First, the measured sea surface temperature data and the first remote sensing observed sea surface temperature data of the target sea area are obtained.

[0038] In this embodiment, the target sea area may include a specific area where sea surface temperature observation and correction are required.

[0039] The measured sea surface temperature data is measured by on-site collection equipment (such as buoys or ship-borne thermometers). It is directly derived from the real ocean environment and reflects the actual sea surface temperature conditions. Therefore, it can be used as standard or benchmark data.

[0040] The first remote sensing observation sea surface temperature data is the sea surface temperature data measured by satellite remote sensors. The sensors on the satellite will receive radiation information from the ocean surface and infer the sea surface temperature. However, due to atmospheric interference, sensor errors and other reasons, there is a certain deviation from the measured sea surface temperature data.

[0041] That is to say, the above-mentioned measured sea surface temperature data and the first remote sensing observed sea surface temperature data are both used for subsequent correction operations, wherein the measured sea surface temperature data is the reference standard for correction, and the first remote sensing observed sea surface temperature data is the object that needs to be corrected.

[0042] After obtaining the measured sea surface temperature data and the first remotely sensed sea surface temperature data, the first remotely sensed sea surface temperature data can be corrected by an inverse distance weighted interpolation method based on the measured sea surface temperature data to generate the second remotely sensed sea surface temperature data.

[0043] In this embodiment, the inverse distance weighted interpolation method is a spatial interpolation method. In this embodiment, the principle is: the closer the unknown point (remote sensing observation point, in this embodiment, the remote sensing observation grid point) is to the known point (measured point), the greater the weight of the known point, and the farther the distance, the smaller the weight. Therefore, when interpolating, the influence (i.e., weight) of each known point is calculated based on the distance between the known point and the unknown point, and then the data of the unknown point is corrected by these weights.

[0044] Specifically, the measured sea surface temperature data of multiple measured points in the target sea area can be used to find the measured points adjacent to each remote sensing observation point. Then, weights are assigned according to the distance. The closer the measured points are, the greater the impact on the remote sensing observation points. The temperature value of each remote sensing point is adjusted by weighted average to generate the preliminary corrected "second remote sensing observation sea surface temperature data". Through the interpolation method, the error between the first remote sensing observation sea surface temperature data and the measured sea surface temperature data can be preliminarily reduced to improve the accuracy of the data.

[0045] The second remotely sensed sea surface temperature data obtained in the above steps may then be processed. Specifically, the second remotely sensed sea surface temperature data may be time iteratively corrected based on a time correction parameter library to generate third remotely sensed sea surface temperature data.

[0046] In this embodiment, the second remote sensing observation data is corrected moment by moment, and the historical features in the time correction parameter library are used to reduce the error. The above process is an iterative correction process, that is, multiple corrections are required to gradually adjust the deviation so that the remote sensing data at the current moment gradually coincides with the measured data.

[0047] Based on the above method, the error of the second remote sensing observation data can be reduced in the time dimension, especially the error caused by the fluctuation of sea surface temperature over time. Through time iterative correction, the generated third remote sensing observation sea surface temperature data is more accurate than the second remote sensing observation sea surface temperature data.

[0048] In the above embodiment, the time correction parameter library is constructed based on the historical reanalysis sea surface temperature data of the target sea area and the first remote sensing observation sea surface temperature data, and is used to record and extract the deviation characteristics of the first remote sensing observation sea surface temperature data at different time scales. Optionally, the time scale may include a climate scale, a seasonal scale, and a sub-seasonal scale. Among them, the climate scale may include climate change trends from decades to centuries; the seasonal scale may include temperature fluctuations caused by changes throughout the year; and the sub-seasonal scale may include small fluctuations within a range of weeks to months.

[0049] Figure 2 A schematic diagram of a time iterative correction processing method provided by an embodiment of the present invention.

[0050] like Figure 2 As shown, firstly, the second remote sensing observation sea surface temperature data of the target sea area are obtained (such as T0-15 to T0-09 in the figure), and these data are subjected to preliminary inverse distance weighted interpolation correction processing.

[0051] Extracting time scale series: From the second remote sensing observation sea surface temperature data, the time series of climate scale, seasonal scale and sub-seasonal scale are extracted based on the preset sliding window. The length of the sliding window is determined according to the periodic characteristics of each scale, such as T0-14, T0-13, etc.

[0052] Time deviation correction: Perform time deviation correction on the extracted time scale sequence (such as each layer of data block in the figure). Use the periodic characteristics of the time correction parameter library to iteratively correct the sequence of each time scale. The iterative algorithm continuously adjusts and corrects the deviation at each time point (T0-14 to T0-08, etc.).

[0053] Iteration process and convergence judgment: During the correction process, the algorithm checks whether the iteration value meets the preset convergence condition. If the condition is met, the iteration is stopped and the corrected third remote sensing observation sea surface temperature data of each time scale is generated (such as T0-07, T0-02, etc. after being transferred to the lower layer in the figure).

[0054] Accumulation processing: Finally, all the data layers that have been iteratively corrected by time scale are accumulated to form more accurate time-corrected data. The figure shows that the final output T0 at the bottom gradually accumulates from T0-07, representing the generated third remote sensing observation sea surface temperature data.

[0055] Result output: The final third remote sensing observation sea surface temperature data will be used for subsequent spatial correction and accumulation processing to further improve the temporal and spatial consistency and accuracy of the data.

[0056] Based on the above specific steps of time iterative correction processing, the time deviation of remote sensing data is continuously optimized through iterative algorithms, so that the accuracy of data at various time scales can be improved, ultimately laying the foundation for generating more reliable sea surface temperature data.

[0057] The third remotely sensed sea surface temperature data obtained in the above steps may then be processed. Specifically, the third remotely sensed sea surface temperature data may be spatially corrected based on a spatial correction parameter library to generate fourth remotely sensed sea surface temperature data.

[0058] In this embodiment, the third remote sensing observation sea surface temperature data is further corrected using the spatial covariance information in the parameter library. The spatial correlation between different locations is taken into account in the correction process, so that the spatial error is gradually reduced. Based on the above method, it can ensure that the third remote sensing observation sea surface data is more reasonably distributed in space, reduce the deviation of the spatial dimension, and obtain more accurate fourth remote sensing observation sea surface temperature data.

[0059] In this embodiment, the spatial correction parameter library is constructed based on the historical reanalysis sea surface temperature data, and is used to store the covariance characteristics of the first remote sensing observation sea surface temperature data at different spatial locations. Specifically, the spatial correction parameter library is established based on the historical reanalysis data, and records the spatial relationship between different locations in the target sea area, especially the covariance characteristics of the sea surface temperature at different locations. Covariance can reveal whether the temperature change in a certain area will affect the adjacent area.

[0060] Finally, the third remote sensing observed sea surface temperature data and the fourth remote sensing observed sea surface temperature data are accumulated and processed to obtain the remote sensing observed sea surface temperature correction data corresponding to the target sea area.

[0061] In this embodiment, the accumulation processing may include combining the third remote sensing observation sea surface temperature data and the fourth remote sensing observation sea surface temperature data, and accumulating them to obtain the final remote sensing observation sea surface temperature correction data. Specifically, the results of time correction and space correction are accumulated, and based on the method of fusion multi-scale correction, the overall consistency of the data is ensured by simultaneously considering the errors in time and space. The final generated remote sensing observation sea surface temperature correction data is the result obtained after multiple iterations and multi-scale fusion, and represents the most accurate sea surface temperature data of the target sea area.

[0062] In an optional embodiment, when obtaining the measured sea surface temperature data and the first remote sensing observed sea surface temperature data of the target sea area, the initial measured sea surface temperature data and the initial remote sensing observed sea surface temperature data of the target sea area may be obtained first.

[0063] In this embodiment, the initial measured sea surface temperature data is usually collected by buoys, ships or other field observation equipment. The initial remote sensing sea surface temperature data is obtained by satellites or other remote sensing equipment. The above two types of data are the basis of the entire iterative correction process, and subsequent operations are processed based on these initial data.

[0064] Then, the initial measured sea surface temperature data are denoised and outliers are eliminated to generate the measured sea surface temperature data of the target sea area.

[0065] In this embodiment, the initial measured sea surface temperature data may contain noise (such as instrument error or natural interference) and outliers (caused by data collection errors or extreme weather conditions). The initial measured data is processed by a denoising algorithm (such as smoothing filtering, low-pass filtering, etc.) and an outlier removal method (such as Z-score detection method, box plot, etc.) to ensure that the generated measured sea surface temperature data is more accurate and can reflect the actual sea surface temperature state.

[0066] Finally, the initial remote sensing observed sea surface temperature data is subjected to denoising, outlier removal and resampling processing to generate the first remote sensing observed sea surface temperature data.

[0067] In this embodiment, similar to the initial measured sea surface temperature data in the above steps, the initial remote sensing observed sea surface temperature data may also contain noise and outliers, and need to be denoised and outlier removed. Since the initial remote sensing observed sea surface temperature data usually has a higher resolution or a different time frequency, the resampling process can adjust the initial remote sensing observed sea surface temperature data to a time or spatial resolution that matches the measured sea surface temperature data, so that the subsequent interpolation and correction steps can proceed more smoothly. The first remote sensing observed sea surface temperature data finally generated will be used as one of the main input data in the iterative correction process.

[0068] In this embodiment, effective measured and remote sensing observation data are extracted from the initial data through the data preprocessing step. The data quality can be ensured and the accuracy of the subsequent iterative correction process can be improved. The data is optimized by denoising, outlier removal, resampling, etc., so that the correction results are more reliable and accurate.

[0069] In an optional embodiment, the second remote sensing sea surface temperature data is generated based on the measured sea surface temperature data by correcting the first remote sensing sea surface temperature data through the inverse distance weighted interpolation method. Specifically:

[0070] First, a plurality of measured points corresponding to the measured sea surface temperature data adjacent to the remote sensing observation point corresponding to the first remote sensing observed sea surface temperature data may be determined.

[0071] In the present embodiment, before performing the inverse distance weighted interpolation calculation, it is necessary to clarify the positions of the observation points in the remote sensing observation sea surface temperature data and the measurement points in the measured sea surface temperature data. Each remote sensing observation point represents the sea surface temperature observation value at that point, and each measured point is the actual measurement data obtained by means of ships or buoys. Through the search algorithm, multiple measured points around each remote sensing observation point (within a certain range of proximity, the search radius R in the present embodiment is 5km) are determined to form a data set, which will be used for the interpolation calculation next. Optionally, it can be implemented by different search algorithms, such as a nearest neighbor search algorithm based on a grid search or based on a spatial tree (such as a KD tree). In specific applications, different search radiuses can also be set according to actual needs, thereby affecting the accuracy and coverage of the interpolation calculation.

[0072] Then, the weight coefficients corresponding to the multiple measured points are determined according to the distances between the remote sensing observation point and the multiple measured points.

[0073] In this embodiment, the core idea of ​​the inverse distance weighted interpolation is that the closer the measured point is to the remote sensing observation point, the greater the influence on the interpolation result. Specifically, the weight of each measured point can be calculated using the inverse distance function. The closer the measured point is, the greater the weight is, while the farther the measured point is, the smaller the weight is, so that the measured point closer to the remote sensing observation point has a greater contribution.

[0074] Optionally, the weight calculation in this embodiment can be implemented based on different inverse distance formulas. In this embodiment, the weight coefficient can be calculated based on the following formula (1):

[0075]

[0076] Among them, W m is the distance weight coefficient of the mth measured point, d mIt is the distance between the remote sensing observation point corresponding to the first remote sensing sea surface temperature data to be corrected and the mth measured point within the corresponding search radius R.

[0077] The above formula (1) can be used to calculate the weight values ​​of different measured points. The measured points closer to the remote sensing observation point use a simple inverse distance weight (i.e. 1 / d_m), while the measured points farther away use a smoothly decreasing weight coefficient to ensure that the weight gradually decreases with the increase of distance. If the distance between the measured point and the observation point exceeds the search radius R, the measured point has no effect on the interpolation result and the weight is 0.

[0078] Finally, based on the weight coefficient, the first remote sensing observation sea surface temperature data is corrected by the inverse distance weighted interpolation method to generate the second remote sensing observation sea surface temperature data.

[0079] Based on the above weight coefficients, the temperature data of each remote sensing observation point is weighted interpolated. Specifically, by multiplying the temperature values ​​of the surrounding measured points by their respective weight coefficients, and then adding these weighted values, an interpolated temperature value is calculated, thereby obtaining the new second remote sensing observation sea surface temperature data.

[0080] In practical applications, the interpolation process can combine multiple data sources (such as buoy data, satellite data, etc.) to improve the accuracy of the interpolation results. Optionally, an error correction model can be introduced to further correct the interpolated temperature data to improve the reliability of the final result for areas with sparse or irregular data distribution.

[0081] In this embodiment, the second remote sensing observation sea surface temperature data can be calculated based on the following formula (2):

[0082]

[0083] Among them, P m Search the observation value of the mth measured point within the radius R for the remote sensing observation point to be corrected.

[0084] In an optional embodiment, a time correction parameter library may be constructed based on the historical reanalyzed sea surface temperature data of the target sea area and the first remote sensing sea surface temperature data by the following method, specifically including:

[0085] Firstly, the historical reanalysis sea surface temperature data and the first remote sensing observation sea surface temperature data are resampled to the same spatial resolution uniform grid matching alignment by bilinear interpolation method, and the deviation between the historical reanalysis sea surface temperature data and the first remote sensing observation sea surface temperature data is calculated to obtain the deviation data set.

[0086] In this embodiment, in order to ensure that the historical reanalysis sea surface temperature data and the first remote sensing observation sea surface temperature data are unified in space, the spatial resolution of the two types of data are resampled by bilinear interpolation and aligned on a unified spatial grid. After alignment, the difference between the two types of data, i.e., the deviation, is calculated, and a deviation data set is generated, which is used for subsequent correction and feature extraction.

[0087] Bilinear interpolation is a method of interpolating the value of a target point by using the values ​​of surrounding grid points based on the distance weight of the location. This allows data of different spatial resolutions to be converted to the same grid. Deviation calculation is to compare historical reanalysis data and remote sensing observation data, calculate the difference between the two at the same location, and obtain the deviation value.

[0088] Then, the deviation dataset is processed based on the wavelet transform method to extract the climate-scale periodic features, seasonal-scale periodic features and sub-seasonal-scale periodic features in the deviation dataset.

[0089] Wavelet transform is a tool suitable for non-stationary signal analysis, which can decompose the local features of time series and extract the changing trends in different frequency ranges. In this embodiment, wavelet transform can be used to effectively extract the different time scale period features in the deviation data.

[0090] Figure 3 A diagram showing the effect of extracting time scale features of sea surface temperature from historical reanalysis based on wavelet transform provided in an embodiment of the present invention.

[0091] like Figure 3 As shown in the figure, the effect of extracting time scale features of historical reanalysis sea surface temperature (SST) based on wavelet transform is demonstrated. By separating the changes in sea surface temperature at different time scales, we can deeply understand the climate trend, seasonal scale features and sub-seasonal scale features. The meanings of the three parts in the figure are as follows:

[0092] Climate Trends (Part 1): The top graph shows the climate trends of sea surface temperature. The black curve represents the raw sea surface temperature data, and the red straight line is the climate trend line, with the equation y=0.00011157x+21.8508.

[0093] This section shows the climate change trend over a long time scale. The equation of the climate trend line shows that during the study period (2001 to 2021), the sea surface temperature showed a slow upward trend, increasing by about 0.00011157℃ per year. This trend reflects the impact of global warming on ocean temperature, and the periodic fluctuations in the original data were successfully separated by the wavelet transform, retaining only the overall trend information.

[0094] SST seasonal scale characteristics (Part 2): The middle figure shows the extracted seasonal scale characteristics, and the curve shows how the sea surface temperature changes with the seasonal cycle.

[0095] This part mainly reflects the seasonal periodic fluctuation characteristics, that is, the change pattern of sea surface temperature within a year. It can be observed that the temperature change shows obvious annual periodic fluctuations, with an amplitude of about ±5℃. Seasonal fluctuations are temperature increases or decreases caused by seasonal changes, such as high temperatures in summer and low temperatures in winter. These periodic characteristics are accurately separated from the original data.

[0096] SST sub-seasonal scale characteristics (Part 3): The bottom figure shows the sub-seasonal scale characteristic curve, showing short-period temperature fluctuations.

[0097] The subseasonal scale reflects fluctuations in sea surface temperature that are shorter than seasonal changes, such as temperature changes caused by storms, ocean currents, and short-term meteorological events. The amplitude of these fluctuations is relatively small, about ±0.5°C, and the frequency of fluctuations is high. Through wavelet transform, subseasonal changes are effectively separated from changes on a larger time scale, providing in-depth analysis capabilities for short-term ocean phenomena.

[0098] The three parts in the figure show the separation effect of different time scale characteristics. The wavelet transform successfully decomposes the original sea surface temperature data into climate trends, seasonal fluctuations and sub-seasonal fluctuations, making each change pattern clearer. This separation method helps to understand the multi-scale characteristics of sea surface temperature changes and can provide strong support for the improvement of climate models, marine environmental monitoring and forecasting. In the long run, sea surface temperatures are gradually rising, while seasonal fluctuations are still significant, and short-term sub-seasonal changes cannot be ignored. This shows that multi-scale analysis is crucial to revealing the complex mechanism of ocean temperature changes.

[0099] In this embodiment, the deviation data set is processed using wavelet transform to extract periodic features at different time scales. In this embodiment, three types of periodic features are included, namely climate scale periodic features, seasonal scale periodic features and sub-seasonal scale periodic features. Among them, climate scale periodic features can reflect the long-term trend of climate change, usually from several years to several decades. Seasonal scale periodic features can reflect the medium-term trend of seasonal changes, usually from several months to one year. Sub-seasonal scale periodic features can reflect the characteristics of short-term changes, usually from several weeks to several months.

[0100] Finally, based on the climate scale periodic characteristics, seasonal scale periodic characteristics and sub-seasonal scale periodic characteristics, the deviation data set is band-pass filtered to generate standard climate scale time series, standard seasonal scale time series and standard sub-seasonal scale time series, so as to generate a time correction parameter library based on the standard climate scale time series, standard seasonal scale time series and standard sub-seasonal scale time series.

[0101] Bandpass filtering can isolate and retain the data information within the target frequency band, while filtering out noise at other frequencies. Each time scale series represents the deviation characteristics of the sea surface temperature data at that scale. Ultimately, these time series are used to construct a time correction parameter library, which records the deviation characteristics at different time scales for subsequent data correction.

[0102] After extracting the periodic features of different scales, the deviation data set is bandpass filtered to generate three types of time series: standard climate scale time series, which contains the correction information of climate scale cycle; standard seasonal scale time series, which contains the correction information of seasonal scale cycle; standard sub-seasonal scale time series, which contains the correction information of sub-seasonal scale cycle.

[0103] In an embodiment of the present invention, a method for constructing a time correction parameter library is provided by processing historical reanalysis sea surface temperature data and first remote sensing observation sea surface temperature data. The method includes resampling by bilinear interpolation, extracting periodic features of different time scales by wavelet transform, and generating multi-scale standard time series by bandpass filtering. The final output time correction parameter library is used to record and extract deviation features of remote sensing observation data at different time scales.

[0104] In this embodiment, the bandpass filtering process can be calculated based on the following formula (3):

[0105] z i,j,n =a1(y i,h,n -y i,j,n-2 )-b1z i,j,n-1 -b2z i,j,n-2 (3)

[0106] Where a1, b1 and b2 are filter coefficients, Z i,j,n is a time series at space (i, j) and time n.

[0107] In an optional embodiment, when processing the deviation data set based on a wavelet transform method to extract climate-scale periodic features, seasonal-scale periodic features and sub-seasonal-scale periodic features in the deviation data set, each deviation value in the deviation data set can first be z-score standardized to obtain a standardized deviation data set; then, based on the standardized deviation data set, a wavelet transform is performed using Morlet wavelet as a wavelet mother function to obtain a decomposition result of the standardized deviation data set in the time-frequency domain, and the decomposition result includes multiple time-scale periodic information; finally, based on multiple time-scale periodic information, the climate-scale periodic features, seasonal-scale periodic features and sub-seasonal-scale periodic features are determined.

[0108] In this embodiment, z-score normalization is performed on each deviation value based on the deviation data set, the purpose of which is to convert the deviation data of different ranges into unified standardized data. Z-score normalization can normalize the deviation data to the same scale, which is convenient for subsequent analysis and processing. Specifically, z-score normalization is achieved by calculating the difference between each deviation value and the mean of the deviation data set and dividing it by the standard deviation of the deviation data set. Such normalization can eliminate the scale differences between different data sets, so that they have the same mean and variance, thereby facilitating the processing of the differences between different data sets in subsequent analysis.

[0109] Morlet wavelet is used as the wavelet mother function. Due to its good time-frequency localization characteristics, Morlet wavelet can effectively decompose the different time scale information in the sea surface temperature deviation data set. Wavelet transform decomposes the standardized deviation data set in two dimensions of time and frequency, thereby obtaining periodic information at multiple time scales, laying the foundation for the subsequent extraction of periodic characteristics at climate scale, seasonal scale and sub-seasonal scale.

[0110] Through the results of the aforementioned wavelet transform, we can obtain periodic information at different time scales. Based on this periodic information, we can distinguish the periodic characteristics of the climate scale, seasonal scale, and sub-seasonal scale. The climate scale periodic characteristics usually refer to fluctuations on a longer time scale (such as many years), the seasonal scale periodic characteristics refer to periodic changes of about one year, and the sub-seasonal scale periodic characteristics refer to fluctuations shorter than the seasonal scale. These characteristics of different time scales help to understand and analyze the changing laws of sea surface temperature deviations in different time periods.

[0111] In an optional embodiment, based on the multiple time scale periodic information, the wavelet variance is calculated, and the first three maximum values ​​can be taken as the main periodic values, that is, the climate scale periodic characteristics, the seasonal scale periodic characteristics and the sub-seasonal scale periodic characteristics are determined.

[0112] In this embodiment, the standardized deviation data set can be calculated based on the following formula (4):

[0113]

[0114] in, For x i,j,n The average value of (n=1,2,…,N), S i,j For x i,j,n Variance of (n=1,2,…,N);

[0115] The wavelet transform formula is shown in the following formula (5):

[0116]

[0117] Where ω0=6.2, a is the time scale transformation coefficient. Optionally, the time scale transformation coefficient can be arithmetic sequence of 0 to 5 with a base of 2 and a power of 0.25 as an interval, to obtain the time scale transformation coefficient sequence 2 0 , 2 0.25 , 2 0.5 ,……,2 5 .

[0118] In an optional embodiment, a spatial correction parameter library can be constructed based on historical reanalysis sea surface temperature data: first, the climate-scale deviation data set, seasonal-scale deviation data set, and sub-seasonal-scale deviation data set are processed separately through a spatial sliding window algorithm to generate a covariance matrix between each spatial position; then, based on the covariance matrix, the covariance values ​​between different spatial positions are determined to construct a spatial correction parameter library based on the covariance values.

[0119] In this embodiment, the deviation data set is composed of the differences between historical reanalysis data and remote sensing observation data, reflecting the spatial deviation characteristics at different time scales. The climate scale, seasonal scale and sub-seasonal scale represent the long-term, medium-term and short-term change laws respectively. The spatial sliding window algorithm selects a certain window size within the spatial range, performs local calculations on the deviation data within the window, and analyzes the correlation of different spatial positions. After the sliding window processing, the covariance matrix between each spatial position is obtained, and the covariance matrix is ​​used to represent the correlation of data changes between different spatial positions.

[0120] After the covariance matrix is ​​generated, the covariance values ​​between the spatial locations can be extracted from the covariance matrix. The covariance value reflects the statistical relationship between different spatial locations, which indicates how the temperature deviation values ​​of two spatial locations are related to each other.

[0121] Based on these covariance values, a spatial correction parameter library is constructed. The spatial correction parameter library saves the covariance information of different locations as the basis for subsequent spatial deviation correction of remote sensing observation sea surface temperature data. By using the covariance parameter library, the temperature deviation at different locations can be corrected more accurately, improving the accuracy of spatial correction.

[0122] The covariance matrix can be calculated based on the following formula (6):

[0123]

[0124] Among them, z l,n 、z l+R,n are the spatial positions z l 、z l+R The sample value at time n, They are z l 、z l+R The sample mean of .

[0125] Through the above steps, the spatial deviation characteristics of the historical reanalysis data are extracted and quantified, forming a parameter library that can reflect the covariance relationship between different spatial locations. The spatial correction parameter library can provide data support for the subsequent spatial correction of remote sensing observation data, which can effectively correct the spatial system deviation and improve the spatial resolution and accuracy of remote sensing data.

[0126] In an optional embodiment, when performing time iterative correction on the second remotely sensed observed sea surface temperature data based on the time correction parameter library to generate the third remotely sensed observed sea surface temperature data, the following steps may be performed:

[0127] Firstly, based on a preset sliding window, remote sensing climate scale time series, remote sensing seasonal scale time series and remote sensing sub-seasonal scale time series are extracted from the second remote sensing observed sea surface temperature data. The lengths of the preset sliding window correspond to the climate scale periodic characteristics, seasonal scale periodic characteristics and sub-seasonal scale periodic characteristics, respectively.

[0128] In this embodiment, the process of extracting different time scale features from the second remote sensing observation sea surface temperature data. Specifically, according to the periodic features of the climate scale, seasonal scale and sub-seasonal scale that have been extracted and determined in the time correction parameter library, the corresponding sliding window is set. The length of each sliding window is consistent with these periodic features to ensure that the data at different time scales are effectively covered. By gradually moving the sliding window, a series of climate, seasonal and sub-seasonal time series are extracted. These time series represent the trend of sea surface temperature changes at different time scales.

[0129] Then, the time deviations of remote sensing climate-scale time series, remote sensing seasonal-scale time series and remote sensing sub-seasonal-scale time series are revised through an iterative algorithm.

[0130] In this embodiment, the iterative algorithm is used to revise the time bias of the extracted climate, seasonal and sub-seasonal time series. The time bias refers to the difference between the remote sensing observation data and the historical reanalysis data. The iterative algorithm gradually adjusts these biases to reduce the errors in the remote sensing observation data and make it closer to the real climate, seasonal and sub-seasonal sea surface temperature changes.

[0131] When the correction value in the iterative process meets the preset convergence condition, the iteration is stopped to generate the third remote sensing observation sea surface temperature data at each time scale.

[0132] In this embodiment, when using the iterative algorithm to revise the deviation, a preset convergence condition is set. This condition usually means that when the iterative correction value (i.e., the adjustment amount of the time deviation) reaches a certain threshold, the effect of the deviation adjustment is no longer significant, that is, the convergence state. At this time, the iteration stops and the final third remote sensing observation sea surface temperature data is generated. After multiple iterative corrections, the third remote sensing observation sea surface temperature data has fully corrected its deviation characteristics on the time scale and can more accurately reflect the actual temperature changes in the target sea area.

[0133] When the time deviation of the remote sensing climate scale time series, remote sensing seasonal scale time series and remote sensing sub-seasonal scale time series is revised by iterative algorithm, the deviation value can be calculated based on the following formula (7):

[0134]

[0135] Where ΔSST t is the deviation value at time t under time scale N, SST S t-n , The time scale is N = a i SST remote sensing observation value and SST reanalysis value at the next time tn.

[0136] In this embodiment, the remote sensing sea surface temperature data at different time scales are iteratively revised based on the time correction parameter library. The time series is extracted by a preset sliding window, the deviation is gradually adjusted using an iterative algorithm, and the difference in time scale is revised based on the deviation calculation formula, and finally the third remote sensing sea surface temperature data with accurate correction is generated.

[0137] In an optional embodiment, when performing spatial correction on the third remotely sensed observed sea surface temperature data based on the spatial correction parameter library to generate fourth remotely sensed observed sea surface temperature data, the following steps may be performed:

[0138] Firstly, the third remote sensing observation sea surface temperature data at different time scales are preprocessed respectively so that the data format of the third remote sensing observation sea surface temperature data is consistent with the spatial correction parameter library.

[0139] In this embodiment, before performing spatial correction, the third remote sensing observation sea surface temperature data obtained from the time correction step needs to be format converted. The data structure of each time scale may be different, and the purpose of preprocessing is to unify the format of these data so that it matches the format of the spatial correction parameter library used for spatial correction. This includes resampling, standardization or reorganization of the data to ensure compatibility between data in subsequent processing. The result of preprocessing is to ensure that data at different time scales can be docked with the spatial correction parameter library in a consistent format, thereby avoiding errors or unprocessable situations caused by inconsistent formats.

[0140] Then, according to the covariance value in the spatial correction parameter library, the deviation value of each remote sensing observation point in the third remote sensing observation sea surface temperature data is corrected to generate fourth remote sensing observation sea surface temperature data corresponding to each of the third remote sensing observation sea surface temperature data at different time scales.

[0141] The core step of spatial correction is to correct the deviation value in the third remote sensing observation sea surface temperature data based on the covariance value in the spatial correction parameter library. Covariance is used to describe the correlation between sea surface temperature observations at different spatial locations. According to the covariance value between different remote sensing observation points and other observation points, the deviation of each remote sensing observation point is corrected. This step helps to eliminate the deviation caused by spatial inconsistency, making the temperature values ​​of each observation point more accurate and consistent, thereby reducing the deviation of observation data in local areas.

[0142] After completing the deviation correction of each remote sensing observation point, the corresponding fourth remote sensing observation sea surface temperature data is generated based on different time scales (such as climate scale, seasonal scale and sub-seasonal scale). In other words, after the data of each time scale is spatially corrected, the corresponding fourth remote sensing observation data will be generated respectively. These data will include corrected temperature observations, reflecting the spatial variation characteristics at different time scales. By processing data of different time scales separately, the spatial variation of sea surface temperature in different time periods can be better captured, thereby improving the accuracy and application value of the data.

[0143] In an optional embodiment, when the third remotely sensed sea surface temperature data and the fourth remotely sensed sea surface temperature data are accumulated to obtain the remotely sensed sea surface temperature correction data corresponding to the target sea area, the following steps may be performed:

[0144] Firstly, based on the third remote sensing observation sea surface temperature data and the fourth remote sensing observation sea surface temperature data, the weighted sum corresponding to each remote sensing monitoring point is calculated respectively to generate a weighted sum data set.

[0145] In this embodiment, the third remote sensing observation sea surface temperature data and the fourth remote sensing observation sea surface temperature data represent two types of temperature data after time and space correction. A new data set is generated by calculating the weighted sum of these two types of data for each remote sensing monitoring point. In this embodiment, the above two sets of data are combined with specific weights by weighted sum, which can comprehensively consider the deviation correction of the two dimensions of time and space to ensure the rationality of the data. The weight can be determined by the correction coefficient at different scales, so that the weighted data can better reflect the real sea surface temperature changes.

[0146] Then, the weighted sum dataset is spatially interpolated to generate the corrected results of remote sensing observed sea surface temperature.

[0147] The spatial interpolation in this embodiment is to fill in missing data between different points of the sea surface temperature data or to improve the spatial resolution. In the previous step, the weighted sum data set of each remote sensing monitoring point has been calculated, but in order to ensure the continuity of these data in space, interpolation processing must be performed. Through the spatial interpolation technology, the data of each observation point of the weighted sum data set is extended to the entire target area, thereby generating a complete and coherent sea surface temperature data set, that is, the remote sensing observation sea surface temperature correction result in this embodiment.

[0148] Finally, outliers are detected and removed from the correction results of remote sensing sea surface temperature observations, and the quality-improved correction data of remote sensing sea surface temperature observations are obtained through smoothing and correction.

[0149] Outlier detection can include identifying and removing extreme values ​​or erroneous observation points that may exist in the data set to improve the accuracy and reliability of the data. These outliers may be caused by factors such as sensor errors and environmental interference. After removing the outliers, smoothing is to reduce the noise in the data to make the changes in temperature data more natural and stable. Correction is to adjust the systematic errors in the data to further improve the accuracy of the data. After these steps, the data has higher quality and can more accurately reflect the true situation of sea surface temperature.

[0150] In this embodiment, the correction result of the remote sensing observation sea surface temperature can be calculated based on the following formula (8):

[0151]

[0152] Among them, SST S is the remote sensing observation value, B is the inverse distance weight correction coefficient, ΔSSTk , A k , C l,k The time correction deviation, wavelet decomposition coefficient and spatial covariance correction coefficient at scale k are respectively used.

[0153] Figure 4 A flow chart of a method for synthesizing the correction results of sea surface temperature observed by remote sensing provided by an embodiment of the present invention. Figure 4 As shown, the method comprises the following steps:

[0154] On-site measurement correction: First, the measured sea surface temperature data of the target sea area is obtained, and the initial remote sensing observation sea surface temperature data is initially corrected. This process corrects the first remote sensing observation data through the inverse distance weighted interpolation method to generate the second remote sensing observation sea surface temperature data.

[0155] According to the time correction parameter library, the second remote sensing observation sea surface temperature data at the climate scale, seasonal scale and sub-seasonal scale are respectively corrected in time iteration:

[0156] Climate-scale time iterative correction: Extract climate-scale time series and use the climate cycle characteristics in the parameter library for iterative correction.

[0157] Seasonal scale time iterative correction: Extract seasonal scale time series and use seasonal cycle characteristics to adjust time deviation.

[0158] Sub-seasonal time iterative correction: extract sub-seasonal time series and revise the deviation based on the sub-seasonal cycle characteristics.

[0159] Based on the spatial correction parameter library, the spatial deviation of the time-corrected data of each scale is corrected:

[0160] Climate-scale spatial correction: The climate-scale covariance matrix is ​​used to correct the spatial deviations between different locations in the third-party remote sensing observation data.

[0161] Seasonal scale spatial correction: The spatial deviation after seasonal scale correction is adjusted through the seasonal scale covariance matrix.

[0162] Sub-seasonal scale spatial correction: Relying on the covariance characteristics of the sub-seasonal scale, the spatial bias of the sub-seasonal correction results is corrected.

[0163] Synthesis of deviation correction results at different scales: Synthesize the remote sensing sea surface temperature data that have been corrected at different time and space scales. Accumulate the deviation correction results at different scales, combine the correction effects in time and space, and obtain the comprehensive sea surface temperature correction results.

[0164] Output correction results: The final output is the synthesized remote sensing observation sea surface temperature correction data, which fully considers the deviation characteristics of each time scale and spatial position, and significantly improves the accuracy and consistency of remote sensing data.

[0165] This embodiment demonstrates the overall process of multi-scale time and space correction, and describes in detail how each step combines the time correction parameter library and the space correction parameter library to synthesize a more accurate sea surface temperature correction result through iteration and correction.

[0166] The following is a detailed description of a multi-scale sea surface temperature remote sensing data iterative correction method provided by the present invention in conjunction with a specific embodiment. The method is based on the measured, reanalyzed and remotely sensed observation data of the target sea area (taking sea area A as an example in this embodiment), performs multi-scale temporal and spatial correction processing, and finally generates high-precision sea surface temperature data.

[0167] Taking a certain sea area A as the target sea area, we first obtain the historical reanalysis sea surface temperature data, measured sea surface temperature data and remote sensing observation sea surface temperature data of the sea area in the past 10 years. The specific data sources are as follows:

[0168] Historical reanalysis of sea surface temperature: using the Global Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) provided by the UK Met Office.

[0169] Measured sea surface temperature: Derived from the United States measured data quality monitoring system (iQuam).

[0170] Remote sensing observations of sea surface temperature: provided by OSI-SAF.

[0171] Based on the above data, a multi-scale temporal and spatial correction parameter library of sea surface temperature is constructed.

[0172] Build time correction parameter library:

[0173] Data alignment: The historical reanalysis sea surface temperature and the remote sensing observed sea surface temperature were spatially aligned to the same uniform grid of 0.01°×0.01° using the bilinear interpolation method.

[0174] Deviation extraction: Through the alignment results, the deviation between the remote sensing observed sea surface temperature and the reanalysis sea surface temperature is extracted.

[0175] Wavelet analysis: The wavelet analysis method is used to extract the different time scale characteristics of the sea surface temperature deviation, and obtain the climate scale period a1, seasonal scale period a2, and sub-seasonal scale period a3, where the corresponding wavelet coefficient transformations are T(a1,b), T(a2,b), and T(a3,b), respectively.

[0176] Bandpass filtering: The Butterworth filter is used to bandpass filter the historical reanalysis sea surface temperature series to obtain climate scale time series, seasonal scale time series, and sub-seasonal scale time series respectively.

[0177] Construct a spatial correction parameter library: Construct a spatial correction parameter library based on climate scale time series, seasonal scale time series and sub-seasonal scale time series.

[0178] The parameter library contains the spatial covariance characteristics between remote sensing sea surface temperature deviations and historical reanalysis sea surface temperatures at different time scales, which are used for subsequent spatial corrections.

[0179] Preliminary processing of remote sensing sea surface temperature data: Obtain the measured sea surface temperature data of sea area A and the initial remote sensing sea surface temperature data.

[0180] Quality control: De-noising and outlier removal are performed on the measured sea surface temperature to ensure data quality.

[0181] Resampling: The initial remote sensing observation sea surface temperature data is subjected to denoising, outlier removal and resampling processing to generate the first remote sensing observation sea surface temperature data.

[0182] Time iteration correction: Based on the measured sea surface temperature data, the inverse distance weighted interpolation method is used to make a preliminary correction to the first remote sensing observation sea surface temperature data to generate the second remote sensing observation sea surface temperature data.

[0183] The time correction parameter library is used to iteratively correct the second remote sensing observation sea surface temperature data on a time scale to generate the third remote sensing observation sea surface temperature data.

[0184] Spatial correction: Based on the spatial correction parameter library, the third remote sensing observation sea surface temperature data is corrected on a spatial scale to obtain the fourth remote sensing observation sea surface temperature data.

[0185] Synthesis of corrected data: The third remote sensing observation sea surface temperature data and the fourth remote sensing observation sea surface temperature data are cumulatively processed, and the multi-scale correction results are integrated to finally generate the remote sensing observation sea surface temperature correction data of the target sea area.

[0186] The multi-scale iterative correction method of this embodiment can effectively reduce the deviation between the remotely sensed sea surface temperature and the measured value, improve the temporal and spatial accuracy of the sea surface temperature, and enhance the monitoring and analysis capabilities of ocean changes at the climate, seasonal and sub-seasonal scales. Experimental results show that this method significantly improves the reliability of remote sensing data and provides important data support for subsequent marine environmental monitoring and climate change research.

[0187] Based on the multi-scale sea surface temperature remote sensing data iterative correction method of the above embodiment, the embodiment of the present invention also provides a multi-scale sea surface temperature remote sensing data iterative correction device. Figure 5 As shown, the device includes an acquisition module 501, a correction module 502 and a processing module 503.

[0188] The acquisition module 501 is used to acquire the measured sea surface temperature data and the first remote sensing observed sea surface temperature data of the target sea area.

[0189] The correction module 502 is used to correct the first remote sensing observed sea surface temperature data by using the inverse distance weighted interpolation method based on the measured sea surface temperature data to generate the second remote sensing observed sea surface temperature data.

[0190] The correction module 502 is also used to perform time iterative correction on the second remotely sensed observed sea surface temperature data based on the time correction parameter library to generate third remotely sensed observed sea surface temperature data. The time correction parameter library is constructed based on the historical reanalysis sea surface temperature data of the target sea area and the first remotely sensed observed sea surface temperature data, and is used to record and extract deviation characteristics at different time scales in the first remotely sensed observed sea surface temperature data.

[0191] The correction module 502 is also used to perform spatial correction on the third remotely sensed observed sea surface temperature data based on the spatial correction parameter library to generate fourth remotely sensed observed sea surface temperature data. The spatial correction parameter library is constructed based on historical reanalysis sea surface temperature data and is used to store the covariance characteristics of the first remotely sensed observed sea surface temperature data between different spatial positions.

[0192] The processing module 503 is used to accumulate the third remote sensing observed sea surface temperature data and the fourth remote sensing observed sea surface temperature data to obtain the remote sensing observed sea surface temperature correction data corresponding to the target sea area.

[0193] In this embodiment, the acquisition module 501 is specifically used to acquire the initial measured sea surface temperature data and the initial remote sensing observed sea surface temperature data of the target sea area. Correspondingly, the processing module 503 is specifically used to perform denoising and outlier removal processing on the initial measured sea surface temperature data to generate the measured sea surface temperature data of the target sea area; perform denoising, outlier removal and resampling processing on the initial remote sensing observed sea surface temperature data to generate the first remote sensing observed sea surface temperature data.

[0194] In this embodiment, the correction module 502 is specifically used to determine the measured points corresponding to the multiple measured sea surface temperature data adjacent to the remote sensing observation point corresponding to the first remote sensing observation sea surface temperature data; determine the weight coefficients corresponding to the multiple measured points according to the distance between the remote sensing observation point and the multiple measured points; based on the weight coefficients, correct the first remote sensing observation sea surface temperature data by inverse distance weighted interpolation method to generate second remote sensing observation sea surface temperature data;

[0195] The weight coefficient is calculated based on the following formula (1):

[0196]

[0197] Among them, W m is the distance weight coefficient of the mth measured point, d m is the distance between the remote sensing observation point corresponding to each of the first remote sensing observed sea surface temperature data to be corrected and the mth measured point within the corresponding search radius R;

[0198] The second remote sensing observation sea surface temperature data is calculated based on the following formula (2):

[0199]

[0200] Among them, P m Search the observation value of the mth measured point within the radius R for the remote sensing observation point to be corrected.

[0201] In this embodiment, the processing module 503 is specifically used to resample the historical reanalysis sea surface temperature data and the first remote sensing observation sea surface temperature data to the same spatial resolution uniform grid matching alignment through a bilinear interpolation method, calculate the deviation of the historical reanalysis sea surface temperature data and the first remote sensing observation sea surface temperature data to obtain a deviation data set; process the deviation data set based on a wavelet transform method to extract climate scale periodic features, seasonal scale periodic features, and sub-seasonal scale periodic features in the deviation data set; based on the climate scale periodic features, seasonal scale periodic features, and sub-seasonal scale periodic features, perform bandpass filtering on the deviation data set to generate a standard climate scale time series, a standard seasonal scale time series, and a standard sub-seasonal scale time series, so as to generate a time correction parameter library according to the standard climate scale time series, the standard seasonal scale time series, and the standard sub-seasonal scale time series;

[0202] The bandpass filtering process is calculated based on the following formula (3):

[0203] z i,j,n =a1(y i,j,n -y i,j,n-2 )-b1z i,j,n-1 -b2z i,j,n-2 (3)

[0204] Where a1, b1 and b2 are filter coefficients, Z i,j,n is the time series at space (i, j) and time n;

[0205] The deviation data set is processed based on the wavelet transform method to extract the climate scale periodic features, seasonal scale periodic features and sub-seasonal scale periodic features in the deviation data set, including:

[0206] Perform z-score normalization on each deviation value in the deviation data set to obtain a standardized deviation data set;

[0207] Based on the standardized deviation data set, the Morlet wavelet is used as the wavelet mother function to perform wavelet transform, and the decomposition result of the standardized deviation data set in the time and frequency domain is obtained. The decomposition result includes multiple time scale period information.

[0208] Based on multiple time scale period information, determine the climate scale period characteristics, seasonal scale period characteristics and sub-seasonal scale period characteristics;

[0209] The standardized deviation data set can be calculated based on the following formula (4):

[0210]

[0211] in, For x i,j,n The average value of (n=1,2,…,N), S i,j For x i,j,n Variance of (n=1,2,…,N);

[0212] The wavelet transform formula is shown in the following formula (5):

[0213]

[0214] Among them, ω0=6.2, a is the time scale transformation coefficient.

[0215] In this embodiment, the processing module 503 is specifically used to process the climate scale deviation data set, the seasonal scale deviation data set and the sub-seasonal scale deviation data set respectively through the spatial sliding window algorithm to generate a covariance matrix between each spatial position; based on the covariance matrix, determine the covariance value between different spatial positions to construct a spatial correction parameter library according to the covariance value;

[0216] The covariance matrix can be calculated based on the following formula (6):

[0217] Among them, z l,n 、z l+R,n are the spatial positions z l 、z l+R The sample value at time n, They are z l 、z l+R The sample mean of .

[0218] In this embodiment, the correction module 502 is specifically used to extract the remote sensing climate scale time series, the remote sensing seasonal scale time series and the remote sensing sub-seasonal scale time series from the second remote sensing observed sea surface temperature data based on a preset sliding window, and the length of the preset sliding window corresponds to the climate scale periodic characteristics, the seasonal scale periodic characteristics and the sub-seasonal scale periodic characteristics respectively; revise the time deviation of the remote sensing climate scale time series, the remote sensing seasonal scale time series and the remote sensing sub-seasonal scale time series through an iterative algorithm; when the correction value in the iterative process meets the preset convergence condition, stop the iteration to generate the third remote sensing observed sea surface temperature data at each time scale;

[0219] When the time deviation of the remote sensing climate scale time series, remote sensing seasonal scale time series and remote sensing sub-seasonal scale time series is revised by iterative algorithm, the deviation value can be calculated based on the following formula (7):

[0220]

[0221] Where ΔSST t is the deviation value at time t under time scale N, SST S t-n , The time scale is N = a i SST remote sensing observation value and SST reanalysis value at the next time tn.

[0222] In this embodiment, the correction module 502 is specifically used to preprocess the third remotely sensed sea surface temperature data at different time scales, so that the data format of the third remotely sensed sea surface temperature data is consistent with the spatial correction parameter library; according to the covariance value in the spatial correction parameter library, the deviation value of each remote sensing observation point in the third remotely sensed sea surface temperature data is corrected to generate fourth remotely sensed sea surface temperature data corresponding to each of the third remotely sensed sea surface temperature data at different time scales.

[0223] In this embodiment, the correction module 502 is specifically used to calculate the weighted sum corresponding to each remote sensing monitoring point based on the third remote sensing observed sea surface temperature data and the fourth remote sensing observed sea surface temperature data, respectively, to generate a weighted sum data set; perform spatial interpolation processing on the weighted sum data set to generate a remote sensing observed sea surface temperature correction result; perform outlier detection and removal on the remote sensing observed sea surface temperature correction result, and obtain the remote sensing observed sea surface temperature correction data with improved quality through smoothing and correction processing;

[0224] The correction result of remote sensing sea surface temperature can be calculated based on the following formula (8):

[0225]

[0226] Among them, SST S is the remote sensing observation value, B is the inverse distance weight correction coefficient, ΔSST k , A k , C l,k The time correction deviation, wavelet decomposition coefficient and spatial covariance correction coefficient at scale k are respectively used.

[0227] An embodiment of the present invention further provides an electronic device, comprising: at least one processor; and a memory in communication with the at least one processor. The memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, the electronic device executes the method of the embodiment of the present invention.

[0228] An embodiment of the present invention further provides a non-transitory machine-readable medium storing a computer program, wherein the computer program, when executed by a processor of a computer, is used to cause the computer to execute the method of the embodiment of the present invention.

[0229] An embodiment of the present invention further provides a computer program product, including a computer program, wherein the computer program, when executed by a processor of a computer, is used to enable the computer to execute the method of the embodiment of the present invention.

[0230] refer to Figure 6 , a block diagram of an electronic device that can be used as a server or client of an embodiment of the present invention will now be described, which is an example of a hardware device that can be applied to various aspects of the present invention. Electronic devices are intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0231] like Figure 6 As shown, the electronic device includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0232] Multiple components in the electronic device are connected to the I / O interface 605, including: an input unit 606, an output unit 607, a storage unit 608, and a communication unit 609. The input unit 606 can be any type of device that can input information to the electronic device, and the input unit 606 can receive input digital or character information, and generate key signal input related to user settings and / or function control of the electronic device. The output unit 607 can be any type of device that can present information, and can include but is not limited to a display, a speaker, a video / audio output terminal, a vibrator, and / or a printer. The storage unit 608 can include but is not limited to a disk, an optical disk. The communication unit 609 allows the electronic device to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks, and can include but is not limited to a modem, a network card, an infrared communication device, a wireless communication transceiver, and / or a chipset, such as a Bluetooth device, a WiFi device, a WiMax device, a cellular communication device, and / or the like.

[0233] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a CPU, a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above. For example, in some embodiments, the method embodiments of the present invention may be implemented as a computer program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on an electronic device via a ROM 602 and / or a communication unit 609. In some embodiments, the computing unit 601 may be configured to perform the above-described method in any other appropriate manner (e.g., by means of firmware).

[0234] The computer programs for implementing the methods of the embodiments of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer programs are executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer programs may be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0235] In the context of an embodiment of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable signal medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0236] It should be noted that the term "including" and its variations used in the embodiments of the present invention are open inclusions, that is, "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments". The modifications of "one" and "multiple" mentioned in the embodiments of the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless the context clearly indicates otherwise, it should be understood as "one or more".

[0237] 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, stored data, displayed data, etc.) involved in the embodiments of the present invention are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0238] The various steps described in the method implementation methods provided in the embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method implementation methods may include additional steps and / or omit the steps shown. The scope of protection of the present invention is not limited in this respect.

[0239] The term "embodiment" in this specification refers to specific features, structures or characteristics described in conjunction with the embodiment that can be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily mean the same embodiment, nor does it mean that it is mutually exclusive with other embodiments and is independent or optional. The various embodiments in this specification are described in a related manner, and the same or similar parts between the various embodiments refer to each other. In particular, for the device, equipment, and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts refer to the partial description of the method embodiment.

[0240] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of patent protection. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the attached claims.

Claims

1. A multi-scale sea surface temperature remote sensing data iterative correction method, characterized in that: include: Obtaining the measured sea surface temperature data and the first remote sensing observed sea surface temperature data of the target sea area; Based on the measured sea surface temperature data, the first remote sensing observed sea surface temperature data is corrected by an inverse distance weighted interpolation method to generate second remote sensing observed sea surface temperature data; performing time iterative correction on the second remotely sensed observed sea surface temperature data based on a time correction parameter library to generate third remotely sensed observed sea surface temperature data, wherein the time correction parameter library is constructed based on the historical reanalysis sea surface temperature data of the target sea area and the first remotely sensed observed sea surface temperature data and is used to record and extract deviation characteristics at different time scales in the first remotely sensed observed sea surface temperature data; Performing spatial correction on the third remotely sensed observed sea surface temperature data based on a spatial correction parameter library to generate fourth remotely sensed observed sea surface temperature data, wherein the spatial correction parameter library is constructed based on the historical reanalysis sea surface temperature data and is used to store covariance characteristics of the first remotely sensed observed sea surface temperature data between different spatial positions; The third remotely sensed sea surface temperature data and the fourth remotely sensed sea surface temperature data are cumulatively processed to obtain the remotely sensed sea surface temperature correction data corresponding to the target sea area.

2. The method according to claim 1, characterized in that The step of obtaining the measured sea surface temperature data of the target sea area and the first remote sensing observed sea surface temperature data comprises: Acquiring initial measured sea surface temperature data and initial remote sensing observed sea surface temperature data of the target sea area; De-noising and outlier elimination processing is performed on the initial measured sea surface temperature data to generate the measured sea surface temperature data of the target sea area; The initial remotely sensed sea surface temperature data is subjected to denoising, outlier removal and resampling processing to generate the first remotely sensed sea surface temperature data.

3. The method according to claim 1, characterized in that The method of correcting the first remotely sensed observed sea surface temperature data by using an inverse distance weighted interpolation method based on the measured sea surface temperature data to generate second remotely sensed observed sea surface temperature data comprises: Determine a plurality of measured points corresponding to the measured sea surface temperature data adjacent to the remote sensing observation point corresponding to the first remote sensing observed sea surface temperature data; Determine the weight coefficients corresponding to the plurality of measured points respectively according to the distances between the remote sensing observation point and the plurality of measured points; Based on the weight coefficient, the first remote sensing observation sea surface temperature data is corrected by an inverse distance weighted interpolation method to generate the second remote sensing observation sea surface temperature data; The weight coefficient is calculated based on the following formula (1): Among them, W m is the distance weight coefficient of the mth measured point, d m is the distance between the remote sensing observation point corresponding to each of the first remote sensing observed sea surface temperature data to be corrected and the mth measured point within the corresponding search radius R; The second remote sensing observation sea surface temperature data is calculated based on the following formula (2): Among them, P m The observation value of the mth measured point within the search radius R for the remote sensing observation point to be corrected.

4. The method according to claim 1, characterized in that: The time correction parameter library is constructed based on the historical reanalyzed sea surface temperature data of the target sea area and the first remote sensing sea surface temperature data, including: Resampling the historical reanalysis sea surface temperature data and the first remote sensing observation sea surface temperature data to the same spatial resolution uniform grid matching alignment by a bilinear interpolation method, and calculating the deviation between the historical reanalysis sea surface temperature data and the first remote sensing observation sea surface temperature data to obtain a deviation data set; Processing the deviation data set based on a wavelet transform method to extract climate scale periodic features, seasonal scale periodic features and sub-seasonal scale periodic features in the deviation data set; Based on the climate scale periodic characteristics, the seasonal scale periodic characteristics and the sub-seasonal scale periodic characteristics, the deviation data set is subjected to band-pass filtering to generate a standard climate scale time series, a standard seasonal scale time series and a standard sub-seasonal scale time series, so as to generate the time correction parameter library according to the standard climate scale time series, the standard seasonal scale time series and the standard sub-seasonal scale time series; The bandpass filtering process is calculated based on the following formula (3): z i,j,n =a1(y i,j,n -y i,j,n-2 )-b1z i,j,n-1 -b2z i,j,n-2 (3) Where a1, b1 and b2 are filter coefficients, Z i,j,n is the time series at space (i, j) and time n; The wavelet transform-based method processes the deviation data set to extract climate scale periodic features, seasonal scale periodic features, and sub-seasonal scale periodic features in the deviation data set, including: Performing z-score normalization processing on each deviation value in the deviation data set to obtain a normalized deviation data set; Based on the standardized deviation data set, using Morlet wavelet as a wavelet mother function, performing wavelet transform to obtain a decomposition result of the standardized deviation data set in the time-frequency domain, wherein the decomposition result includes multiple time scale period information; Based on the multiple time scale period information, determining the climate scale period characteristics, the seasonal scale period characteristics and the sub-seasonal scale period characteristics; The standardized deviation data set can be calculated based on the following formula (4): in, For x i,j,n The average value of (n=1,2,…,N), S i,j For x i,j,n Variance of (n=1,2,…,N); The wavelet transform formula is shown in the following formula (5): Among them, ω0=6.2, a is the time scale transformation coefficient.

5. The method according to claim 4, characterized in that The spatial correction parameter library is constructed based on the historical reanalysis sea surface temperature data, including: The climate scale deviation dataset, the seasonal scale deviation dataset and the sub-seasonal scale deviation dataset are processed respectively by a spatial sliding window algorithm to generate a covariance matrix between each spatial position; Based on the covariance matrix, determining covariance values ​​between different spatial positions, so as to construct the spatial correction parameter library according to the covariance values; The covariance matrix can be calculated based on the following formula (6): Among them, z l,n 、z l+R,n are the spatial positions z l 、z l+R The sample value at time n, They are z l 、z l+R The sample mean of .

6. The method according to claim 4, characterized in that The step of performing time iteration correction on the second remotely sensed observed sea surface temperature data based on the time correction parameter library to generate third remotely sensed observed sea surface temperature data comprises: Extracting a remote sensing climate scale time series, a remote sensing seasonal scale time series and a remote sensing sub-seasonal scale time series from the second remotely sensed sea surface temperature data based on a preset sliding window, wherein the lengths of the preset sliding window correspond to the climate scale periodic characteristics, the seasonal scale periodic characteristics and the sub-seasonal scale periodic characteristics, respectively; Revising the time deviations of the remote sensing climate scale time series, the remote sensing seasonal scale time series and the remote sensing sub-seasonal scale time series through an iterative algorithm; When the correction value in the iterative process meets the preset convergence condition, the iteration is stopped to generate the third remote sensing observation sea surface temperature data at each time scale; Wherein, when the time deviation of the remote sensing climate scale time series, the remote sensing seasonal scale time series and the remote sensing sub-seasonal scale time series is revised by the iterative algorithm, the deviation value can be calculated based on the following formula (7): Where ΔSST t is the deviation value at time t under time scale N, SST S t-n , The time scale is N = a i SST remote sensing observation value and SST reanalysis value at the next time tn.

7. The method according to claim 1, characterized in that The third remote sensing observed sea surface temperature data is spatially corrected based on the spatial correction parameter library to generate fourth remote sensing observed sea surface temperature data, including: Preprocessing the third remote sensing observed sea surface temperature data at different time scales respectively, so that the data format of the third remote sensing observed sea surface temperature data is consistent with the spatial correction parameter library; According to the covariance value in the spatial correction parameter library, the deviation value of each remote sensing observation point in the third remote sensing observation sea surface temperature data is corrected to generate the fourth remote sensing observation sea surface temperature data corresponding to each of the third remote sensing observation sea surface temperature data at different time scales.

8. The method according to claim 1, characterized in that The step of accumulating the third remotely sensed sea surface temperature data and the fourth remotely sensed sea surface temperature data to obtain the remotely sensed sea surface temperature correction data corresponding to the target sea area comprises: Based on the third remote sensing observation sea surface temperature data and the fourth remote sensing observation sea surface temperature data, respectively calculating the weighted sum corresponding to each remote sensing monitoring point to generate a weighted sum data set; Performing spatial interpolation processing on the weighted sum data set to generate a correction result of the remote sensing observed sea surface temperature; Performing outlier detection and removal on the remote sensing observation sea surface temperature correction result, and obtaining the remote sensing observation sea surface temperature correction data with improved quality through smoothing and correction processing; The correction result of the remote sensing observation sea surface temperature can be calculated based on the following formula (8): Among them, SST S is the remote sensing observation value, B is the inverse distance weight correction coefficient, ΔSST k , A k , C l,k The time correction deviation, wavelet decomposition coefficient and spatial covariance correction coefficient at scale k are respectively used.

9. An electronic device, comprising: A processor, and a memory storing a program, wherein the program comprises instructions, which, when executed by the processor, cause the processor to perform the method according to any one of claims 1 to 8.

10. A non-transitory machine-readable medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 8.

Images