A method for inverting ocean wave characteristics based on inverse echo measurement

Abstract

The present invention belongs to the field of ocean wave feature extraction, and specifically relates to a method for inverting sea wave features based on reverse echo measurement, which includes the following steps: obtaining the echo signal from the seabed to the ocean surface, performing adaptive time window smoothing processing on the echo signal, and performing outlier identification and missing data filling on the data after smoothing processing; using the improved Tukey method to remove outliers and using the piecewise cubic spline interpolation method to fill in the missing data; obtaining the sea wave echo signal with the low-frequency trend removed through high-pass filtering and detrending processing; selecting an appropriate time window to calculate the dispersion of the echo time, establishing a mathematical relationship between the dispersion and the significant wave height through a regression algorithm, and substituting the observed data into the equation to obtain the inversion result of the significant wave height. The present invention performs feature inversion on sea waves based on reverse echo measurement technology, effectively improving the accuracy of sea wave feature extraction, reducing the on-site measurement cost, and providing a stable and low-cost sea wave measurement method.

Description

Technical Field

[0001] The present invention belongs to the field of ocean wave feature extraction, and specifically relates to a method for inverting sea wave features based on reverse echo measurement. Background Art

[0002] Sea wave inversion is mainly used to study the characteristics, propagation, and energy distribution of ocean waves. Existing technical means and methods usually rely on wave observation devices (such as wave buoys, radars, satellite remote sensing, etc.) for sea wave inversion and monitoring, and analyze the measured data using inversion algorithms to deduce parameters such as the height of ocean waves.

[0003] The working principles of existing sea wave inversion methods generally include the following several types:

[0004] Wave buoy measurement: The buoy directly senses the undulation changes of waves through sensors installed on the sea surface and records the water surface height change information. Through spectral analysis, information such as the period and wave height of the waves is obtained.

[0005] Radar measurement: The radar emits electromagnetic waves and receives the echo signals, and analyzes the wave conditions on the water surface using the Doppler effect or reflection characteristics, which is particularly suitable for wave measurement near the coastline.

[0006] Satellite remote sensing: The satellite observes a large range of the ocean surface through an altimeter or synthetic aperture radar (SAR), and obtains information such as wave height and wave spectrum through processing and inversion.

[0007] Although significant progress has been made in sea wave inversion with existing technologies, there are still the following main problems:

[0008] Insufficient real-time performance: Existing methods, especially those based on satellite or buoy data processing, often have latency problems and are difficult to provide real-time and refined wave dynamic information. This poses limitations for some applications that require quick responses (such as vessel navigation, offshore operation safety, etc.).

[0009] Limited resolution: Although satellite remote sensing has a wide coverage, its spatial resolution is relatively low, and it is unable to accurately capture wave features in a small area, especially near complex terrains or coastlines. Existing inversion methods are restricted by the resolution of observation devices, the algorithm capabilities of data processing, and the complexity of the ocean environment, resulting in limited wave inversion accuracy.

[0010] Greatly affected by the environment, high cost, and difficult to maintain: Buoy devices are easily affected by extreme weather such as storms and strong waves, and the observation data of radar systems will also be interfered with in bad weather, thus affecting the accuracy and stability of the inversion results. The monitoring systems based on buoys and radars have high costs, and the devices are easily damaged in the marine environment, with high maintenance difficulty and costs.

[0011] Adverse effects caused by the above problems:

[0012] Lag in emergency response: Insufficient real-time performance and data latency result in the inability to obtain accurate wave information in a timely manner during offshore operations, increasing the difficulty and risk of emergency response.

[0013] Incomplete information: Low spatial resolution and limited wave inversion accuracy lead to the inability to capture the dynamic changes of small-scale ocean waves within the observation range, affecting high-precision ocean wave prediction and analysis.

[0014] Instability and unreliability: The equipment is vulnerable to environmental factors, resulting in insufficient data continuity and stability, and unable to provide stable and reliable wave data under complex sea conditions.

[0015] Regarding the main problems existing in existing ocean wave acquisition technologies (such as wave buoys, radars, satellite remote sensing), such as insufficient real-time performance, limited spatial resolution, low inversion accuracy, and high susceptibility to the environment. Summary of the Invention

[0016] The object of the present invention is to provide a method for inverting ocean wave characteristics based on reverse echo measurement, which can provide higher-precision ocean wave information in a complex ocean environment, has advantages in terms of real-time performance, stability, and economic benefits, and overcomes the defects of several existing ocean wave inversion methods mentioned in the above background technology.

[0017] The technical solution adopted by the present invention to achieve the above object is: A method for inverting ocean wave characteristics based on reverse echo measurement, comprising the following steps:

[0018] S1: Deploy a reverse echo measurement device in the observation area and obtain the echo signal from the seabed to the ocean surface;

[0019] S2: Perform adaptive time window smoothing processing on the echo signal to eliminate noise interference;

[0020] S3: Identify outliers and fill in missing data for the data after smoothing processing;

[0021] S4: Remove outliers using the improved Tukey method and fill in missing data using the piecewise cubic spline interpolation method;

[0022] S5: Perform high-pass filtering and detrending processing on the time series of the preprocessed echo data to eliminate the influence of low-frequency trends;

[0023] S6: Calculate the dispersion of the echo time based on the detrended data; and establish an inversion equation based on the correlation between the dispersion and the significant wave height;

[0024] S7: Substitute the observed data into the inversion equation to inversely obtain the significant wave height data of the ocean waves with high time resolution.

[0025] In step S1, the observation frequency of the reverse echo measurement device is not less than 24 times per hour to ensure the time resolution and accuracy.

[0026] In step S2, the adaptive time window smoothing process is specifically as follows:

[0027] S21: Dynamically adjust the time window length according to the local variance of the echo signal, where the larger the local variance, the smaller the window length, and the smaller the local variance, the larger the window length;

[0028] S22: Calculate the mean value of the data within the window through a sliding window to replace the original data point.

[0029] The adjustment range of the time window length is set to 15 minutes to 1 hour, and the time window length is:

[0030]

[0031] where m is the window length, is the adjustment factor, is the local variance of the data.

[0032] Step S22 is specifically as follows:

[0033] Calculate the mean value in the window to replace the original data point:

[0034]

[0035] where m is the window size, is an index variable that traverses all data points within the window, and its value range is from the current data point position i to i + m - 1, is to replace the original data point with the mean value, is the original data value at the

[0036] In step S4, the improved Tukey method is adopted, including the following steps:

[0037] S41: Divide the time series data of the echo signal into multiple consecutive time periods according to a fixed period length; the data within each time period is processed independently, and the local interquartile range IQR is independently calculated within each period;

[0038] For the data within each time period, calculate the lower quartile Q1 and the upper quartile Q3 respectively, and obtain the IQR of this period:

[0039] IQR= Q3- Q1

[0040] S42: Dynamically set the outlier threshold based on the local interquartile range (IQR) of each time period, namely:

[0041] Lower limit = Q1 - k × IQR

[0042] Upper limit=Q3-k×IQR

[0043] Among them, k is the adjustment coefficient. If the data point exceeds the threshold range, it will be marked as an outlier;

[0044] S43: Use piecewise cubic spline interpolation to fill in outliers and missing data.

[0045] The step S43 is specifically as follows:

[0046] a. Divide the complete time series into several sub-segments based on the location of outliers or missing data;

[0047] b. Construct a cubic spline interpolation function for the valid data points in each sub-segment;

[0048] c. Use interpolation functions to smoothly fill in outliers and missing data to ensure continuity and differentiability between adjacent sub-segments;

[0049] The filling rules are:

[0050] If the proportion of outliers in a certain time period exceeds the preset threshold, the data of that period will be directly eliminated and replaced by the interpolation results of the previous and next time periods; for continuous missing data, it is necessary to verify whether it exceeds the physical meaning range after interpolation.

[0051] In step S5, the detrending process is specifically as follows:

[0052] S51: Trend extraction: For the high-pass filtered signal, the moving average method is used to extract the low-frequency trend component. The window length is set to 6 hours to filter out the low-frequency trend component with a period greater than 6 hours.

[0053] S52: Trend removal: subtract the extracted low-frequency trend component from the original signal;

[0054] S53: If the residual trend is still significant, further correction is performed using a quadratic polynomial fit to ensure that the signal baseline approaches zero.

[0055] In step S6, the time window length for discreteness calculation is adjustable, ranging from 1 hour to 6 hours, to meet the requirements of different time resolutions;

[0056] The discreteness of the echo time is:

[0057]

[0058] Among them, normalize() is the normalization function, n is the length of the dispersion time window, and x is the echo data after processing.

[0059] In step S6, the inversion equation is established according to the correlation between the dispersion and the significant wave height, specifically as follows:

[0060] S61: Match the dispersion D of the current time window with the historical significant wave height SWH data of the corresponding time period to ensure consistent timestamps;

[0061] S62: Select the first 65% of the observation data as the training set, and the remaining 35% as the validation set;

[0062] S63: Assume that the dispersion D and the significant wave height SWH are linearly related, and use linear regression to fit the regression equation, that is:

[0063]

[0064] where a and b are regression coefficients, and D is the dispersion;

[0065] S64: Optimize the model:

[0066] Residual analysis: Calculate the residual between the predicted value and the actual value as:

[0067]

[0068] Verify whether the residuals follow a normal distribution; if the residuals show an obvious non-linear trend, introduce polynomial regression or a machine learning model to improve the accuracy.

[0069] The present invention has the following beneficial effects and advantages:

[0070] 1. The present invention obtains high-resolution and high-quality sea wave information: Compared with traditional buoys and radars, this method can flexibly deploy devices as needed, finely monitor sea waves in a large area, with higher data resolution. Especially in complex sea conditions, it can still maintain high-quality data collection, especially in coastal and complex terrain sea areas, which greatly improves the coverage and fineness of sea wave monitoring. It is of great significance for meteorological forecasting, ocean environment research, etc.

[0071] 2. The present invention has strong anti-interference ability: The reverse echo measurement avoids the direct influence of the surface ocean environment (such as storms, large waves, etc.) on the equipment. Compared with buoys that are easily damaged and satellite data that are easily affected by weather, the equipment of the present invention has better stability, and the data collection is continuous and reliable, especially suitable for long-term and complex sea condition sea wave monitoring tasks.

[0072] 3. The present invention has low cost and is simple to use: Compared with traditional buoys that require regular maintenance and high satellite remote sensing costs, the inverse echo measurement method for retrieving ocean waves has significant economic advantages in long-term monitoring projects. Its algorithm is easy to master, and the simplicity of operation enables this technology to be widely applied to various ocean monitoring scenarios without complex subsequent operations or a large amount of human input. Description of the Drawings

[0073] Figure 1 It is a flowchart of the method for retrieving ocean wave characteristics of the present invention;

[0074] Figure 2 It is the comparison result between the inversion example and historical data. Detailed Embodiment

[0075] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0076] As Figure 1 shown, it is a flowchart of the method for retrieving ocean wave characteristics of the present invention. In the actual application process, the method for retrieving ocean wave characteristics based on inverse echo measurement in this embodiment is as follows:

[0077] S1: Deploy an inverse echo measurement device in the ocean wave observation area and fix it through a deployment rack and an anchor system;

[0078] S2: Use the inverse echo measurement technology to obtain the echo signal from the seabed to the ocean surface. To ensure the time resolution and accuracy of ocean wave observation, the observation frequency is not less than 24 times per hour;

[0079] S3: Adaptively select the time window length according to the dynamic change characteristics of the echo signal for smoothing processing. The minimum boundary of the adaptive window length is 15 minutes, and the maximum boundary is 1 hour. Adjust the specific window size according to the local variance of the data. In the sliding window, the window size is automatically adjusted according to the local change of the data. When the data fluctuates greatly, the window size is reduced, and when the data fluctuates little, the window size is increased to ensure the smoothness of the data:

[0080]

[0081] In the formula, m is the window length, is the adjustment factor, is the local variance of the data.

[0082] Calculate the mean value in the window to replace the original data point:

[0083]

[0084] Among them, m is the window size, is an index variable that traverses all data points in the window, and its value range is from the current data point position i to i+m−1. To replace the original data point The mean of For the window The original data value of the location.

[0085] S4: The modified Tukey method was used to remove outliers, namely the local interquartile range (IQR) test method, specifically:

[0086] S41: Divide the time series data of the echo signal into multiple continuous time periods according to a fixed time period length; process the data in each time period independently, and calculate the local interquartile range (IQR) independently in each time period;

[0087] For the data in each time period, calculate the lower quartile Q1 and the upper quartile Q3 respectively, and get the IQR of the period:

[0088] IQR= Q3- Q1

[0089] S42: Dynamically set the outlier threshold based on the local interquartile range (IQR) of each time period, namely:

[0090] Lower limit = Q1 - k × IQR

[0091] Upper limit=Q3-k×IQR

[0092] Among them, k is the adjustment coefficient. If the data point exceeds the threshold range, it will be marked as an outlier;

[0093] S43: The outlier identification period is set to 1 hour, and the outliers and missing data are filled using the piecewise cubic spline interpolation method, specifically:

[0094] a. Divide the complete time series into several sub-segments based on the location of outliers or missing data;

[0095] b. Construct a cubic spline interpolation function for the valid data points in each sub-segment;

[0096] c. Use interpolation functions to smoothly fill in outliers and missing data to ensure continuity and differentiability between adjacent sub-segments;

[0097] The filling rules are:

[0098] If the proportion of outliers in a certain time period exceeds the preset threshold, the data of that period will be directly eliminated and replaced with the interpolation results of the previous and next periods;

[0099] For continuous missing data, it is necessary to verify whether it exceeds the physical significance range after interpolation.

[0100] The local IQR is adjusted according to the distribution changes of the data in different time periods, so as to achieve more accurate outlier identification.

[0101] S5: Perform high-pass filtering and detrending on the time series of the preprocessed echo data to eliminate the influence of low-frequency trends;

[0102] S51: Trend extraction: For the signal after high-pass filtering, use the moving average method to extract the low-frequency trend component, and set the window length to 6 hours to filter out the low-frequency trend component with a period greater than 6 hours.

[0103] S52: Trend removal: Subtract the extracted low-frequency trend component from the original signal;

[0104] S53: If the remaining trend is still significant, use quadratic polynomial fitting for further correction to ensure that the signal baseline approaches zero.

[0105] S6: Calculate the dispersion of the detrended data, and set the calculation time window length to 6 hours, which is calculated by the following formula:

[0106]

[0107] In the formula, normalize is the normalization function, n is the dispersion time window length, and x is the echo data processed through the steps S2 to S5.

[0108] S7: Take the first 65% of the data to establish an inversion relationship with the historical effective wave height data. Considering the high linear correlation between the echo time dispersion and the effective wave height, use a linear regression model to establish an inversion equation:

[0109]

[0110] In the formula, a and b are the coefficients of the regression model, and D is the dispersion.

[0111] Optimize the model:

[0112] Residual analysis: Calculate the residual between the predicted value and the actual value as:

[0113]

[0114] Verify whether the residual follows a normal distribution; if the residual shows an obvious non-linear trend, introduce polynomial regression or machine learning models to improve the accuracy.

[0115] S8: Substitute the observed data into the inversion equation to inversely obtain the significant wave height data of the sea waves with high time resolution. The final inversion result is as Figure 2 shown.

[0116] Generally, the time resolution of the significant wave height data of the sea waves is 1 hour at a relatively high level. The time resolution can be adjusted by adjusting the size of the dispersion time window in step S6. To sum up, the improvement of the present invention overcomes the problems of the existing sea wave measurement methods, such as being vulnerable to sea conditions and human factors interference and having a short measurement period, and provides a new idea for the measurement technology of the significant wave height of the sea waves.

[0117] In this specification, the present invention has been described with reference to its specific embodiments. The above embodiments are the preferred embodiments of this patent and are not used to limit the scope of implementation of the present invention. It should be noted that the present invention is not limited to the above specific implementation manners. Those skilled in the art, without departing from the principle of the present invention, the improvements, changes, combinations, substitutions, etc. made are all within the scope of protection required by the claims of the present invention.

[0118] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the present invention. Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims

1. An inversion method for ocean wave characteristics based on inverse echo measurement, characterized in that, It includes the following steps: S1: Deploy a reverse echo measurement device in the observation area and obtain the echo signal from the seabed to the ocean surface; S2: Perform adaptive time window smoothing on the echo signal to eliminate noise interference; S3: Identify outliers and fill in missing data for the data after smoothing processing; S4: Remove outliers using the improved Tukey method and fill in missing data using the piecewise cubic spline interpolation method; S5: Perform high-pass filtering and detrending on the time series of the preprocessed echo data to eliminate the influence of low-frequency trends; The detrending process is specifically as follows: S51: Trend extraction: For the signal after high-pass filtering, use the moving average method to extract the low-frequency trend component, and set the window length to 6 hours to filter out the low-frequency trend component with a period greater than 6 hours; S52: Trend removal: Subtract the extracted low-frequency trend component from the original signal; S53: If the remaining trend is still significant, use quadratic polynomial fitting for further correction to ensure that the signal baseline approaches zero; S6: Calculate the dispersion of the echo time based on the detrended data; And establish an inversion equation based on the correlation between the dispersion and the significant wave height; The time window length for calculating the dispersion is adjustable, and the adjustment range is from 1 hour to 6 hours to adapt to the requirements of different time resolutions; The dispersion of the echo time is: ； where normalize() is a normalization function, n is the dispersion time window length, and x is the processed echo data; The establishment of the inversion equation based on the correlation between the dispersion and the significant wave height is specifically as follows: S61: Match the dispersion D of the current time window with the historical significant wave height SWH data of the corresponding time period to ensure consistent timestamps; S62: Select the first 65% of the observation data as the training set and the remaining 35% as the validation set; S63: Assume that the dispersion D and the significant wave height SWH are linearly related, and use linear regression to fit the regression equation, that is: ； where a and b are regression coefficients and D is the dispersion; S64: Optimize the model: Residual analysis: Calculate the residual between the predicted value and the actual value as: ； Verify whether the residual follows a normal distribution; if the residual shows an obvious non-linear trend, introduce polynomial regression or a machine learning model to improve the accuracy; S7: Substitute the observation data into the inversion equation to invert the significant wave height data of the sea waves with high time resolution.

2. The method for inverting ocean wave characteristics based on inverse echo measurement according to claim 1, wherein In step S1, the observation frequency of the reverse echo measurement device is not less than 24 times per hour to ensure the time resolution and accuracy.

3. The method for inverting ocean wave characteristics based on reverse echo measurement according to claim 1, characterized in that In step S2, the adaptive time window smoothing process is specifically as follows: S21: Dynamically adjust the time window length according to the local variance of the echo signal, where the larger the local variance, the smaller the window length, and the smaller the local variance, the larger the window length; S22: Calculate the mean value within the window through a sliding window to replace the original data point.

4. The method for inverting ocean wave characteristics based on reverse echo measurement according to claim 3, wherein, The adjustment range of the time window length is set to 15 minutes to 1 hour, and the time window length is: ； where m is the window length, is the adjustment factor, is the local variance of the data.

5. A method for retrieving ocean wave characteristics based on inverse echo measurement according to claim 3, characterized in that, Step S22 is specifically as follows: Calculate the mean value in the window to replace the original data point: ； where m is the window size, is an index variable that traverses all data points within the window, with a value range from the current data point position i to i + m - 1, is the mean value to replace the original data point and is the original data value at the th position within the window.

6. The method for inverting ocean wave characteristics based on inverse echo measurement according to claim 1, characterized in that In step S4, the improved Tukey method includes the following steps: S41: Divide the time series data of the echo signal into multiple consecutive time periods according to a fixed time period length; the data within each time period is processed independently, and the local interquartile range IQR is calculated independently within each time period. For the data within each time period, calculate the lower quartile Q1 and the upper quartile Q3 respectively, and obtain the IQR of this time period: IQR = Q3 - Q1; S42: Dynamically set the outlier determination threshold according to the local interquartile range IQR of each time period, that is: Lower limit = Q1 - k × IQR; Upper limit = Q3 - k × IQR; where k is an adjustment coefficient, and the data points outside the threshold range are marked as outliers. S43: Use piecewise cubic spline interpolation method to fill in the outliers and missing data.

7. The method for retrieving ocean wave characteristics based on inverse echo measurement according to claim 6, characterized in that The specific steps of step S43 are as follows: a. Divide the complete time series into several sub-segments according to the positions of the outliers or missing data. b. Construct a cubic spline interpolation function for the valid data points within each sub-segment. c. Use the interpolation function to smoothly fill in the outliers and missing data to ensure the continuity and differentiability between adjacent sub-segments. Among them, the filling rule is: If the outlier ratio within a certain time period exceeds the preset threshold, the data of this time period is directly removed, and the interpolation results of the previous and subsequent time periods are used for replacement. For continuous missing data, it is necessary to verify whether it exceeds the physical meaning range after interpolation.

Images