A method for inverting effective wave height using marine radar

Abstract

The invention discloses a method for inverting the effective wave height of sea waves by a marine radar, comprising: conducting observation tests offline, respectively counting the real effective wave height under different wave angle conditions and the original effective wave height under different wave angle conditions, counting the ratio of the real effective wave height to the original effective wave height under different wave angles, and fitting the relationship between the wave angle and the ratio using the least square method; calculating the original effective wave height, and calculating the original effective wave height using the relationship between the effective wave height and the root mean square wave steepness and the average zero-crossing period; correcting to obtain a standard effective wave height, and correcting the original effective wave height using the fitting relationship between the obtained wave angle and the ratio of the real effective wave height to the original effective wave height, and obtaining the standard effective wave height. The effective wave height of the invention does not change significantly with the change of the inversion area, significantly improves the inversion accuracy, and enhances the reliability and practicability of the algorithm in practical engineering applications.

Description

Technical Field

[0001] The invention belongs to the technical field of marine remote sensing, and relates to a method for inverting effective wave height of sea waves by a marine radar, in particular to a method for inverting effective wave height of sea waves by a marine radar based on wave angle correction. Background Art

[0002] Since its invention, X-band navigation radar has been widely used for ship navigation and obstacle avoidance. Due to its high temporal and spatial resolution, the sea clutter images obtained by X-band navigation radar contain rich ocean parameter information, and it has become an important instrument for ocean remote sensing. X-band navigation radar can be used to measure a variety of ocean parameters such as ocean internal waves, sea depth, sea surface wind, sea surface currents and waves. Among them, the research on X-band navigation radar inversion of wave parameters has become a scientific research hotspot.

[0003] In the process of being modulated by the waves, the electromagnetic waves of the marine radar are not only affected by factors such as the effective wave height and wave period, but also the selected inversion area and the different wind and wave directions will also affect the results of the wave inversion, thus limiting the application of the algorithm in engineering implementation. At present, the two main methods for inverting effective wave height using X-band marine radar images are spectral analysis and shadow method.

[0004] When using the 3D-FFT spectrum analysis type method to invert wave related parameters, it is necessary to select a square area in the radar image as the area to be processed. Compared with the traditional spectrum analysis method, the biggest advantage of the shadow method is that the calculation of the shadow method is completely based on theoretical derivation, and no buoy experiment is required to calibrate the relevant coefficients. In addition, the shadow method only requires one radar image for calculation, the algorithm has low time cost, and enhances the applicability of the algorithm. However, existing studies have not effectively solved the problem of large inversion errors of effective wave height caused by wave angle. Summary of the invention

[0005] In view of the problems existing in the above-mentioned prior art, the technical problem to be solved by the present invention is to provide a method for inverting the effective wave height of waves by marine radar based on wave angle correction, and to improve the traditional shadow method effective wave height inversion algorithm to accurately eliminate the influence of inversion area selection on the effective wave height result and improve the inversion accuracy.

[0006] In order to solve the above technical problems, a method for inverting the significant wave height of ocean waves by a marine radar of the present invention comprises the following steps:

[0007] Step 1: Conduct observation tests offline, select T radar images under different sea conditions, use buoys to obtain the real effective wave height, wave crest period and wave direction information corresponding to the radar images under different wave angle conditions, select fan-shaped study areas on each radar image, subtract the corresponding wave direction angle from the study area angle to obtain the wave angle, divide the study area into M sub-intervals according to the angle dimension and N sub-intervals radially, use the shadow method to obtain the original effective wave height under different wave angle conditions corresponding to the fan-shaped study area, count the ratio of the real effective wave height to the original effective wave height under different wave angles, and use the least squares method to fit the relationship between the wave angle and the ratio:

[0008] f(x)=a1*sin(b1*x+c1)

[0009] Among them, f(x) is the ratio of the true significant wave height to the original significant wave height, a1, b1, c1 are fitting coefficients, and x is the wave angle;

[0010] Step 2: According to step 1, select the sector-shaped study area on the radar image to be tested and divide it into sub-intervals. Then, according to step 1, use the shadow method to obtain the original effective wave height H of the sector-shaped study area. s2 ;

[0011] Step 3: Use the fitting relationship between the wave angle obtained in step 1 and the ratio of the true significant wave height to the original significant wave height to correct the original significant wave height and obtain the standard significant wave height H s1 :

[0012] H s1 =H s2 *a1*sin(b1*x+c1)

[0013] Furthermore, the shadow method is used to obtain the original effective wave heights under different wave angle conditions corresponding to the fan-shaped study area, including:

[0014] Step 1.1: Calculate the shadow grayscale threshold τ to distinguish between the occluded area and the non-occluded area in the fan-shaped study area S ;

[0015] Step 1.2: Calculate the corresponding occlusion ratio S for each partition sub-interval mn (γ n );

[0016] Step 1.3: Calculate the RMS steepness σ RMS ;

[0017] Step 1.4: Calculate the average zero-crossing period and peak period T M and the average zero-crossing period T m02 The empirical formula is:

[0018]

[0019] Step 1.5: Calculate the effective wave height H S :

[0020]

[0021] Where: g is the acceleration due to gravity.

[0022] Furthermore, step 1.1 calculates the shadow grayscale threshold τ for distinguishing the occluded area from the non-occluded area in the fan-shaped study area. S include:

[0023] Step 1.1.1: Apply the classic difference operator to the original radar echo image to obtain the difference image:

[0024]

[0025] In the above formula, I(r,θ) is the original radar echo image, H i (r, θ) represent the eight directional difference operators of the pixel, i = 1, 2, 3, ... 8;

[0026] The difference operators are:

[0027]

[0028]

[0029] I Ei (r,θ) is the difference image obtained after different difference operators are used;

[0030] Step 1.1.2: Take the pixels with the intensity values ​​in the first I% of the difference image as the boundary pixels between the occluded area and the non-occluded area, so as to obtain the sub-edge image I Ti (r,θ):

[0031]

[0032] Step 1.1.3: Superimpose the eight sub-edge images to obtain the original edge image I T (r,θ), and filter the original edge image to obtain the edge image I F (r,θ):

[0033]

[0034]

[0035] Step 1.1.4: For each isolated 1 in the edge image, find the pixel η of the original radar image at that position, and establish an image histogram function F through the correspondence between the edge image and the original radar image. H (η), for F H (η) Take the majority to get the shadow threshold τ S :

[0036] τ S =mode(F H (η)).

[0037] Furthermore, in step 1.2, the corresponding occlusion ratio S is calculated for each partition sub-interval. mn (γ n )include:

[0038] Step 1.2.1: Using the shadow threshold τ S Segment the original radar image into shadow image I s :

[0039] Where I(r,θ) is the original image;

[0040] Step 1.2.2: Calculate the corresponding grazing angle γ for each radial dimension block:

[0041] Obtain the grazing angle γ of N blocks according to the radial dimension n , where n is the serial number of the radial dimension block, n = 1, 2, 3, ..., N, and the grazing angle of the center point of each radial dimension block is taken as the grazing angle of each sub-block:

[0042]

[0043] in, H is the radar antenna height, E is the relative position calibration resolution, R is the resolution, and D n is the number of center points of the nth radial dimension block, K n is the relative number of starting points of the nth radial dimension block, L n is the relative number of points at the end of the nth radial dimension block;

[0044] Step 1.2.3: For the shadow image I obtained in step 1.2.1 s Each partition block in the mn (γ n ), the formula is as follows:

[0045]

[0046] Where P is the original shadow image I sThe number of pixels in a certain partitioned block whose value is 1, Q is the total number of pixels in the partitioned block, and m is the serial number of the angle-dimension partition, m=1, 2, 3, ..., M.

[0047] Furthermore, step 1.3 calculates the root mean square steepness σ RMS include:

[0048] Shadow ratio function S of the same radial dimension block area n (γ n ) and lighting ratio L n (γ n ) satisfy:

[0049] L n (γ n )=1-S n (γ n )

[0050] Wherein, n is the serial number of each radial dimension block, n=1, 2, ..., N;

[0051] Using the least squares method, the Smith function is used to fit the illumination ratio L of the corresponding N radial blocks in the angle dimension partition. n (γ n ) and the grazing angle γ n The relationship between the root mean square wave steepness σ RMS , the Smith function is:

[0052]

[0053] in

[0054]

[0055] μ n =tan(γ n )

[0056] erfc() is the Gaussian error function. The erfc() function is derived from the Gaussian surface probability density function. The expression of the Gaussian surface probability density function is:

[0057]

[0058] Furthermore, the marine radar is an X-band marine radar.

[0059] Beneficial effects of the present invention: The purpose of the present invention is to eliminate the influence of different wave angles on the performance of the shadow method when using marine radar to invert the effective wave height of ocean waves, and to improve the accuracy of the inversion result. The correction method proposed by the present invention obtains the fitting relationship between the wave angle and the ratio of the real effective wave height to the measured effective wave height by conducting an offline experiment, thereby eliminating the problem that the effective wave height is greatly affected by the wave angle in the original method. Compared with the original method. This method significantly improves the inversion accuracy and enhances the reliability and practicality of the algorithm in practical engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 The radar image corresponding to the data was selected in the experiment;

[0061] Figure 2 Schematic diagram of the study area selection;

[0062] Figure 3 Schematic diagram of the division of the study area;

[0063] Figure 4 Shadow images corresponding to the four sets of data;

[0064] Figure 5 Relationship diagram between wave angle and the ratio of true significant wave height to original significant wave height. DETAILED DESCRIPTION

[0065] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0066] The present invention can be specifically divided into the following steps. The first step is to determine the fitting relationship between the wave angle and the ratio between the real effective wave height and the original effective wave height: conduct observation tests offline, respectively count the real effective wave height under different wave angle conditions and the original effective wave height under different wave angle conditions, count the ratio of the real effective wave height to the original effective wave height under different wave angles, and use the least squares method to fit the relationship between the wave angle and the ratio; the second step is to calculate the original effective wave height: use the relationship between the effective wave height and the root mean square wave steepness and the average zero-crossing period to calculate the measured effective wave height. The third step is to correct and obtain the standard effective wave height: use the fitting relationship between the wave angle obtained in step 1 and the ratio between the real effective wave height and the original effective wave height to correct the original effective wave height to obtain the standard effective wave height.

[0067] Embodiment 1:

[0068] The present invention comprises the following steps:

[0069] Step 1: Conduct observation tests offline, select T radar images under different sea conditions, use buoys to obtain the real effective wave height, wave crest period and wave direction information corresponding to the radar images under different wave angle conditions, select fan-shaped study areas on each radar image, subtract the corresponding wave direction angle from the study area angle to obtain the wave angle, divide the study area into M sub-intervals according to the angle dimension and N sub-intervals radially, use the shadow method to obtain the original effective wave height under different wave angle conditions corresponding to the fan-shaped study area, count the ratio of the real effective wave height to the original effective wave height under different wave angles, and use the least squares method to fit the relationship between the wave angle and the ratio:

[0070] f(x)=a1*sin(b1*x+c1)

[0071] Among them, f(x) is the ratio of the true significant wave height to the original significant wave height, a1, b1, c1 are fitting coefficients, and x is the wave angle;

[0072] Step 2: According to step 1, select the sector-shaped study area on the radar image to be tested and divide it into sub-intervals. Then, according to step 1, use the shadow method to obtain the original effective wave height H of the sector-shaped study area. s2 ;

[0073] Step 3: Use the fitting relationship between the wave angle obtained in step 1 and the ratio of the true significant wave height to the original significant wave height to correct the original significant wave height and obtain the standard significant wave height H s1 :

[0074] H s1 =H s2 *a1*sin(b1*x+c1).

[0075] Embodiment 2:

[0076] The present invention comprises the following steps:

[0077] Step 1: Determine the fitting relationship between the wave angle and the ratio between the true effective wave height and the original effective wave height. Conduct observation tests offline, count the true effective wave height under different wave angle conditions and the original effective wave height under different wave angle conditions, count the ratio of the true effective wave height to the original effective wave height under different wave angles, and use the least squares method to fit the relationship between the wave angle and the ratio, including:

[0078] Step 1.1, conduct observation experiments offline, select T radar images under different sea conditions as the research objects of this example, and record the real effective wave height, wave crest period and wave direction information corresponding to the radar images.

[0079] Step 1.2, complete the selection of the study area for the radar image obtained in step 1.1. Record the angle range, radial distance range and the range of the angle between the study area and the wave direction of the selected study area, and subtract the wave direction angle from the selected area angle to obtain the wave angle. The formula is as follows:

[0080] x=α-β

[0081] Where: x---wave angle

[0082] α---Selection angle

[0083] β---wave direction angle

[0084] Step 1.3, divide the study area in step 1.2 into M sub-intervals according to the angular dimension and into N sub-intervals in the radial dimension.

[0085] Step 1.4, for the sector area in step 1.2, calculate the shadow grayscale threshold for distinguishing the occluded area from the non-occluded area, specifically:

[0086] Step 1.4.1, by applying the classic difference operator to the original image, the difference image is obtained. The specific formula is as follows:

[0087]

[0088] In the above formula, I(r,θ) is the original echo image, H i In (r,θ), i=1,2,3,…8, which represent the difference operators in eight directions of the pixel. The difference operators are

[0089]

[0090]

[0091] I Ei (r,θ) is the difference image obtained after different difference operators are used.

[0092] Step 1.4.2: Take the pixels with the intensity values ​​in the first I% of the difference image as the boundary pixels between the occluded area and the non-occluded area, so as to obtain the sub-edge image I Ti (r,θ). The specific formula is:

[0093]

[0094] Step 1.4.3, superimpose the eight sub-edge images to obtain the original edge image I T (r,θ), and filter the original edge image to obtain the edge image I F (r,θ), the specific formula is

[0095]

[0096]

[0097] Step 1.4.4, for each isolated 1 in the edge image, find the pixel η of the original radar image at that position, and establish an image histogram function F through the correspondence between the edge image and the original radar image H (η), for F H (η) Take the majority to get the shadow threshold τ S .

[0098] τ S =mode(F H (η))

[0099] Step 1.5, calculate the corresponding occlusion ratio S(γ) for each partitioned block obtained in step 1.3, specifically:

[0100] Step 1.5.1, use threshold to obtain image shadow: use the shadow threshold τ obtained in step 1.4.4 S Segment the original radar image into shadow image I s

[0101]

[0102] Where I(r,θ) is the original image.

[0103] Step 1.5.2, calculate the corresponding grazing angle γ for each radial dimension block.

[0104] Obtain the grazing angle γ of N blocks according to the radial dimension n , where n is the serial number of the radial dimension block, n = 1, 2, 3, ..., N. The grazing angle of the center point of each radial dimension block is taken as the grazing angle of each partition block, and the expression is as follows:

[0105]

[0106] in,

[0107] Where: H---Height of radar antenna

[0108] E---Relative position calibration resolution

[0109] R---resolution

[0110] D n ---The number of center points of the nth radial dimension block

[0111] K n ---Relative starting point number of the nth radial dimension block

[0112] L n ---Relative number of points at the end of the nth radial dimension block

[0113] Step 1.5.3, the original shadow image I obtained in step 1.5.1 s Each partition block in the mn (γ n ), the formula is as follows:

[0114]

[0115] Where P is the original shadow image I s The number of pixels in a certain partitioned block whose value is 1, Q is the total number of pixels in the partitioned block, and m is the serial number of the angle-dimension partition, m=1, 2, 3, ..., M.

[0116] Step 1.6, calculate the root mean square steepness, specifically:

[0117] Shadow ratio function S of the same area n (γ n ) and lighting ratio L n (γ n ) has the following relationship:

[0118] L n (γ n )=1-S n (γ n )

[0119] Wherein, n is the serial number of each radial dimension block, n=1, 2, ..., N.

[0120] Using the least squares method, the Smith function is used to fit the illumination ratio L of the corresponding N radial blocks in the angle dimension partition. n (γ n ) and the grazing angle γ n The relationship between the root mean square wave steepness σ RMS , the Smith function is

[0121]

[0122] in

[0123]

[0124] μ n =tan(γ n )

[0125] erfc() is the Gaussian error function. The erfc() function is derived from the Gaussian surface probability density function. The expression of the Gaussian surface probability density function is:

[0126]

[0127] Step 1.7, calculate the average zero-crossing period. Calculate the average zero-crossing period: peak period T M and the average zero-crossing period T m02 The empirical formula is:

[0128]

[0129] Among them, T M Given by a buoy.

[0130] Step 1.8, calculate the original significant wave height: Use the significant wave height H s RMS steepness σ RMS , average zero-crossing period T m02 The effective wave height is calculated by the relationship between:

[0131]

[0132] Where: g is the acceleration due to gravity, H is S is the effective wave height, σ RMS is the root mean square wave steepness, T m02 is the average zero-crossing period.

[0133] Step 1.9, the relationship between the fitted wave angle and the ratio of the true significant wave height to the original significant wave height is:

[0134] Step 1.9.1, calculate the ratio parameter of the true significant wave height to the original significant wave height.

[0135] Step 1.9.2, with the wave angle as the horizontal coordinate and the ratio of the true significant wave height to the original significant wave height as the vertical coordinate, use the least squares method to perform nonlinear fitting of the wave angle and the ratio of the true significant wave height to the original significant wave height in the form of f(x)=a1*sin(b1*x+c1), where a1, b1, and c1 are fitting coefficients, and the relationship between the wave angle and the ratio of the true significant wave height to the original significant wave height is obtained.

[0136] Step 2: Calculate the original effective wave height. The original effective wave height is calculated using the relationship between the effective wave height and the root mean square wave steepness and the average zero-crossing period, specifically including:

[0137] Step 2.1, select the tth (t∈T) radar image and the mth (m∈M) study area. Record the buoy wave height and buoy wave period at this time and calculate the value of the wave angle.

[0138] Step 2.2, divide the study area in step 2.1 into N intervals according to the radial dimension.

[0139] Step 2.3, using step 1.4, obtain the threshold for distinguishing the occluded area from the non-occluded area in the study area in step 2.1, specifically:

[0140] Step 2.3.1, use step 1.4.1 to obtain the difference image.

[0141] Step 2.3.2, using step 1.4.2, obtain the sub-edge image I Ti (r,θ).

[0142] Step 2.3.3, using step 1.4.3, the eight sub-edge images I obtained in step 2.3.2 are Ti (r,θ) is superimposed to obtain the original edge image I T (r,θ), and filter the original edge image to obtain the edge image I F (r,θ).

[0143] Step 2.3.4, using step 1.4.4, count the edge image I obtained in step 2.3.3 F The intensity distribution of pixels with a value of 1 in (r,θ) in the original image is used to obtain the edge histogram F H (η), thereby defining the threshold τ for distinguishing between occluded and non-occluded areas S is the mode in the histogram.

[0144] Step 2.4: For each partitioned block obtained in step 2.2, use step 1.5 to calculate the corresponding occlusion ratio S n (γ n ), specifically:

[0145] Step 2.4.1, use the shadow threshold τ obtained in step 2.3.4 S Segment the original radar image into shadow image I s .

[0146] Step 2.4.2, using step 1.5.2, calculate the corresponding grazing angle γ for each radial dimension block n .

[0147] Step 2.4.3, by calculating the proportion of pixels whose intensity values ​​are less than the threshold, the shadow ratio S (γ ii ):

[0148]

[0149] Where ii is the number of partition blocks, ii = 1, 2, 3, ... 49.ii is the original shadow image I s The number of pixels with a value of 1 in the ii-th partition, Q ii is the total number of pixels in the partitioned block.

[0150] Step 2.5, calculate the root mean square steepness σ using step 1.6 RMS .

[0151] Step 2.6, calculate the average zero-crossing period T using step 1.7 m02 .

[0152] Step 2.7, calculate the original significant wave height H using step 1.8 s2 .

[0153] Step 3: Correction to obtain the standard significant wave height. The original significant wave height is corrected using the fitting relationship between the wave angle obtained in step 1 and the ratio of the true significant wave height to the original significant wave height to obtain the standard significant wave height.

[0154] The following is an example with specific parameters:

[0155] The marine radar used in the implementation of the present invention is an X-band navigation radar, which works in short pulse mode. The echo data is digitized and stored in polar coordinates by line. The time interval between two adjacent stored lines is less than 1ms. The radar antenna scans one circle in about 2.5s. The total number of lines in a radar image is 2048, and there are 2048 pixel points on each line. The azimuth resolution is about 0.1° and the radial resolution is about 2.5m.

[0156] The main technical parameters of the marine radar are shown in Table 1:

[0157] Table 1 Technical parameters of marine radar

[0158]

[0159] Combined with Figure 1-5 , the specific experimental steps of the present invention are:

[0160] The first step is to determine the fitting relationship between the wave angle and the ratio between the true effective wave height and the original effective wave height: conduct observation tests offline, respectively count the true effective wave height under different wave angle conditions and the original effective wave height under different wave angle conditions, count the ratio of the true effective wave height to the original effective wave height under different wave angles, and use the least squares method to fit the relationship between the wave angle and the ratio, including the following steps:

[0161] Step 1.1, conduct observation test offline, select four radar images with effective wave height ranging from 1.5m to 3.5m under different sea conditions as the research objects of this example. The buoy obtains the real effective wave height, wave crest period and wave direction information corresponding to the radar image. Figure 1 As shown, Figure 1 (a), (b), (c), and (d) are four radar images. The experimental location is Pingtan, Fujian. The acquisition time is from 18:00 on November 16, 2014 to 0:00 on November 19, 2014. The radar works in single pulse mode.

[0162] The relevant information of the experimental data is shown in Table 2:

[0163] Table 2 Test data summary

[0164]

[0165]

[0166] Step 1.2: Select the research area based on the four radar images obtained offline. Figure 2 The area shown in the middle sector is the research area, the corresponding angle is 120°~220°northeast, and the radial distance is 200m~2500m. The angle with the wave direction in the figure ranges from 45° to 130°. The wave angle is obtained by subtracting the wave direction angle from the selected area angle, and the formula is as follows:

[0167] x=α-β

[0168] Where: x---wave angle

[0169] α---Selection angle

[0170] β---wave direction angle

[0171] In this example, the selection angle refers to step 1.3. The wave direction angle is given by the buoy. In this example, 64° is taken, and 18 groups of wave angles are obtained. The wave angle parameters are shown in Table 3:

[0172] Table 3 Wave Angle Parameters

[0173]

[0174]

[0175] Step 1.3, the study area is divided into 18 sub-intervals according to the angle dimension. The coverage of each sub-interval is 15°, and the angle interval between two adjacent sub-intervals is 5°. The obtained selection angles are shown in Table 4. The radial division is divided into 49 intervals, and the distance between each interval is 50m, which is divided into 882 sub-blocks in total. Figure 3As shown, the horizontal axis in the figure is the angle and the vertical axis is the radial distance.

[0176] Table 4 Selection Angle Parameters

[0177]

[0178] Step 1.4, calculating the shadow grayscale threshold for distinguishing the occluded area from the non-occluded area for the sector area in step 1.2.

[0179] Specifically include:

[0180] Step 1.4.1, by applying the classic difference operator to the original image, the difference image is obtained. The specific formula is as follows:

[0181]

[0182] In the above formula, I Ei (r,θ) is the edge gradient, r is the distance, θ is the azimuth, I(r,θ) is the original echo image, represents convolution, H i In (r,θ), i=1,2,3,…8, which represent the difference operators in eight directions of the pixel. The difference operators are

[0183] I Ei (r,θ) is the difference image obtained after different difference operators are used.

[0184] Step 1.4.2: Take the pixels with the intensity values ​​in the top 10% of the difference image as the boundary pixels between the occluded area and the non-occluded area, so as to obtain the sub-edge image I Ti (r,θ). The specific formula is

[0185]

[0186] Step 1.4.3, superimpose the eight sub-edge images to obtain the original edge image I T (r,θ), and filter the original edge image to obtain the edge image I F (r,θ), the specific formula is

[0187]

[0188]

[0189] Among them, τ F The value is 5.

[0190] Step 1.4.4, for each isolated 1 in the edge image, find the pixel η of the original radar image at that position, and establish an image histogram function F through the correspondence between the edge image and the original radar image H (η), for F H (η) Take the majority to get the shadow threshold τ S .

[0191] τ S =mode(F H (η))

[0192] Where: η-----pixel.

[0193] Step 1.5, calculate the corresponding occlusion ratio S(γ) for each partitioned block obtained in step 1.3.

[0194] Step 1.5.1, use threshold to obtain image shadow:

[0195] Use the shadow threshold τ obtained in step 1.4.4 S Segment the original radar image into shadow image I s

[0196]

[0197] Where I(r,θ) is the original image. The resulting shadow image is as follows Figure 4 As shown, Figure 4 (a), (b), (c), and (d) are the shadow images of the four radar images respectively.

[0198] Step 1.5.2, calculate the corresponding grazing angle γ for each radial dimension block.

[0199] Obtain the grazing angle γ of 49 blocks according to the radial dimension n , where n is the serial number of the radial dimension block, n = 1, 2, 3, ..., 49. The height of the radar antenna is 45m, and the grazing angle of the center point of each radial dimension block is taken as the grazing angle of each partition block, and the expression is as follows:

[0200]

[0201] Where, the radar antenna height H is 45m, the relative position calibration E is 25, the resolution R is 7.5, and D n is the number of center points of the nth radial dimension block, and the expression is as follows:

[0202]

[0203] In the formula, K n ---Relative starting point number of the nth radial dimension block

[0204] L n ---Relative number of points at the end of the nth radial dimension block

[0205] Step 1.5.3, the original shadow image I obtained in step 1.5.1 s Each partition block in the mn (γ n ), the formula is as follows:

[0206]

[0207] Where P is the original shadow image I s The number of pixels in a certain partitioned block whose value is 1, Q is the total number of pixels in the partitioned block, and m is the serial number of the angle-dimension partition, m=1, 2, 3, ..., 18.

[0208] Step 1.6, calculate the root mean square wave steepness. Specifically include:

[0209] The shadow ratio function S(γ) and the lighting ratio L(γ) of the same area have the following relationship:

[0210] L mn (γ n )=1-S mn (γ n )

[0211] Using the least squares method, for the mth angular region, the Smith function is used to fit the illumination ratio (M)L of the corresponding 49 (N) radial blocks in the angular partition. mn (γ n ) and the grazing angle γ n The relationship between the root mean square wave steepness of different angle dimensions is obtained. The Smith function is

[0212]

[0213] in

[0214]

[0215] μ n =tan(γ n )

[0216] erfc() is the Gaussian error function. The erfc() function is derived from the Gaussian surface probability density function. The expression of the Gaussian surface probability density function is:

[0217]

[0218] Where ξ is the height of the Gaussian sea level from the horizontal plane, and υ is the root mean square of the height deviation.

[0219] The partial wave steepness obtained in different areas is shown in Table 5:

[0220] Table 5 Some RMS steepness parameters in different regions

[0221]

[0222] Step 1.7, calculate the average zero-crossing period: peak period T M and the average zero-crossing period T m02 The empirical formula is:

[0223]

[0224] Among them, T M Given by the buoy. The average zero-crossing period is shown in Table 6:

[0225] Table 6 Average zero-crossing period parameter table

[0226]

[0227] Step 1.8, calculate the original significant wave height: use the significant wave height Hs and the root mean square wave steepness σ RMS , average zero-crossing period T m02 The effective wave height is calculated by the relationship between:

[0228]

[0229] Where: g is the acceleration due to gravity, H is S is the effective wave height, σ RMS is the root mean square wave steepness, T m02 is the average zero-crossing period.

[0230] The original significant wave heights of some regions are shown in Table 7:

[0231] Table 7. Some original significant wave height parameters in different regions

[0232]

[0233] Step 1.9, fit the relationship between the wave angle and the ratio of the true significant wave height to the original significant wave height.

[0234] Step 1.9.1, the ratio parameters of the real significant wave height to the original significant wave height are shown in Table 8:

[0235] Table 8 Parameters of the ratio of the real significant wave height to the original significant wave height

[0236]

[0237]

[0238] Step 1.9.2, with the wave angle as the horizontal coordinate and the ratio of the true significant wave height to the original significant wave height as the vertical coordinate, the wave angle and the ratio of the true significant wave height to the original significant wave height are subjected to nonlinear fitting in the form of f(x)=a1*sin(b1*x+c1) using the least squares method, where a1, b1, c1 are fitting coefficients. The fitting curve is shown in the attached figure. Figure 5 , the relationship between the wave angle and the ratio of the true effective wave height to the original effective wave height is obtained as follows:

[0239] f(x)=0.9761*sin(0.01601x-0.03657)

[0240] In the formula: x-----wave angle

[0241] f-----The ratio of the real effective wave height to the original effective wave height

[0242] The second step is to calculate the original effective wave height.

[0243] Step 2.1, select the fourth radar image, the time is 20:00 on November 17, and the study area is 135°-150°. At this time, the buoy wave height is 3.3m, the buoy wave period is 8.9, and the wave angle is 142.5°.

[0244] Step 2.2: Divide the study area in step 2.1 into 49 intervals according to the radial dimension. The distance between each interval is 50m, resulting in a total of 49 partition blocks.

[0245] Step 2.3, use step 1.4 to obtain the shadow grayscale threshold for distinguishing the occluded area from the non-occluded area in the study area in step 2.1.

[0246] Step 2.3.1, use step 1.4.1 to obtain the difference image.

[0247] Step 2.3.2, using step 1.4.2, obtain the sub-edge image I Ti (r,θ).

[0248] Step 2.3.3, using step 1.4.3, the eight sub-edge images I obtained in step 2.3.2 are Ti (r,θ) is superimposed to obtain the original edge image I T (r,θ), and filter the original edge image to obtain the edge image I F (r,θ).

[0249] Step 2.3.4, using step 1.4.4, count the edge image I obtained in step 2.3.3 F The intensity distribution of pixels with a value of 1 in (r,θ) in the original image is used to obtain the edge histogram F H (η), thereby defining the threshold τ for distinguishing between occluded and non-occluded areas S is the mode in the histogram.

[0250] Step 2.4: For each partitioned block obtained in step 2.2, use step 1.5 to calculate the corresponding occlusion ratio S n (γ n ).

[0251] Step 2.4.1, use the shadow threshold τ obtained in step 2.3.4 S Segment the original radar image into shadow image I s .

[0252] Step 2.4.2, using step 1.5.2, calculate the corresponding grazing angle γ for each radial dimension block n .

[0253] Step 2.4.3, by calculating the proportion of pixels whose intensity values ​​are less than the threshold, the shadow ratio S (γ ii ).

[0254]

[0255] Where ii is the number of partition blocks, ii = 1, 2, 3, ... 49. ii is the original shadow image I s The number of pixels with a value of 1 in the ii-th partition, Q ii is the total number of pixels in the partitioned block.

[0256] Step 2.5, calculate the root mean square steepness σ using step 1.6 RMS ,

[0257] Shadow ratio function S of the same area n (γ n ) and lighting ratio L n (γ n ) has the following relationship:

[0258] L n (γ n )=1-S n (γ n )

[0259] Wherein, n is the serial number of each radial dimension block, n=1, 2, ..., 49.

[0260] Using the least squares method, the Smith function is used to fit the illumination ratio L of the corresponding 49 radial blocks in the angle dimension partition. n (γ n ) and the grazing angle γ n The relationship between the root mean square wave steepness σ RMS , the Smith function is

[0261]

[0262] in

[0263]

[0264] μ n =tan(γ n )

[0265] erfc() is the Gaussian error function. The erfc() function is derived from the Gaussian surface probability density function. The expression of the Gaussian surface probability density function is:

[0266]

[0267] The root mean square wave steepness σ RMS The value of is 0.039.

[0268] Step 2.6, calculate the average zero-crossing period T using step 1.7 m02 The average zero-crossing period is 6.32.

[0269] Step 2.7, calculate the original significant wave height H using step 1.8 s2 , the original effective wave height is 3.46.

[0270] In the third step, the original significant wave height is corrected using the fitting relationship between the wave angle obtained by fitting and the ratio of the true significant wave height to the original significant wave height to obtain the standard significant wave height.

[0271] The correction formula is:

[0272] H s1 =H s2 *0.9761*sin(0.01601x-0.03657)

[0273] Where: H s1 -----Standard effective wave height

[0274] H s2 -----Original effective wave height

[0275] x-----wave angle

[0276] The standard significant wave height obtained by correcting the first set of data in Table 7 using the above correction formula is shown in Table 9:

[0277] Table 9 Part of the standard effective wave height parameters

[0278]

Claims

1. A method for inverting effective wave height of ocean waves using marine radar, characterized in that: The following steps are involved: Step 1: Conduct observation tests offline, select T radar images under different sea conditions, use buoys to obtain the real effective wave height, wave crest period and wave direction information corresponding to the radar images under different wave angle conditions, select fan-shaped study areas on each radar image, subtract the corresponding wave direction angle from the study area angle to obtain the wave angle, divide the study area into M sub-intervals according to the angle dimension and N sub-intervals radially, use the shadow method to obtain the original effective wave height under different wave angle conditions corresponding to the fan-shaped study area, count the ratio of the real effective wave height to the original effective wave height under different wave angles, and use the least squares method to fit the relationship between the wave angle and the ratio: f(x)=a1*sin(b1*x+c1) Among them, f(x) is the ratio of the true significant wave height to the original significant wave height, a1, b1, c1 are fitting coefficients, and x is the wave angle; Step 2: According to step 1, select the sector-shaped study area on the radar image to be tested and divide it into sub-intervals. Then, according to step 1, use the shadow method to obtain the original effective wave height H of the sector-shaped study area. s2 ; Step 3: Use the fitting relationship between the wave angle obtained in step 1 and the ratio of the true significant wave height to the original significant wave height to correct the original significant wave height and obtain the standard significant wave height H s1 : H s1 =H s2 *a1*sin(b1*x+c1)。 2. The method for inverting significant wave height of ocean waves by marine radar according to claim 1, characterized in that: The original effective wave height under different wave angle conditions corresponding to the fan-shaped study area obtained by the shadow method includes: Step 1.1: Calculate the shadow grayscale threshold τ to distinguish between the occluded area and the non-occluded area in the fan-shaped study area S ; Step 1.2: Calculate the corresponding occlusion ratio S for each partition sub-interval mn (γ n ); Step 1.3: Calculate the RMS steepness σ RMS ; Step 1.4: Calculate the average zero-crossing period and peak period T M and the average zero-crossing period T m02 The empirical formula is: Step 1.5: Calculate the effective wave height H S : Where: g is the acceleration due to gravity.

3. The method for inverting significant wave height of ocean waves by marine radar according to claim 2, characterized in that: Step 1.1 calculates the shadow grayscale threshold τ in the fan-shaped study area to distinguish between the occluded area and the non-occluded area S include: Step 1.1.1: Apply the classic difference operator to the original radar echo image to obtain the difference image: In the above formula, I(r,θ) is the original radar echo image, H i (r,θ) represent the eight directional differential operators of the pixel points respectively, i＝1,2,3,…8； The difference operators are: I Ei (r,θ) is the difference image obtained after different difference operators are used; Step 1.1.2: Take the pixels with the intensity values ​​in the top 10% in the differential image as the boundary pixels between the occluded area and the non-occluded area, thus obtaining the sub-edge image I Ti (r,θ): Step 1.1.3: Superimpose the eight sub-edge images to obtain the original edge image I T (r,θ), and filter the original edge image to obtain the edge image I F (r,θ): Step 1.1.4: For each isolated 1 in the edge image, find the pixel η of the original radar image at that position, and establish an image histogram function F through the correspondence between the edge image and the original radar image. H (η), for F H (η) Take the majority to get the shadow threshold τ S : t S =mode(F H (or)).

4. The method for inverting significant wave height of ocean waves by marine radar according to claim 2, characterized in that: Step 1.2: Calculate the corresponding occlusion ratio S for each partition sub-interval mn (γ n )include: Step 1.2.1: Using the shadow threshold τ S Segment the original radar image into shadow image I s : Where I(r,θ) is the original image; Step 1.2.2: Calculate the corresponding grazing angle γ for each radial dimension block: Obtain the grazing angle γ of N blocks according to the radial dimension n , where n is the serial number of the radial dimension block, n = 1, 2, 3, ..., N, and the grazing angle of the center point of each radial dimension block is taken as the grazing angle of each sub-block: in, H is the radar antenna height, E is the relative position calibration resolution, R is the resolution, and D n is the number of center points of the nth radial dimension block, K n is the relative number of starting points of the nth radial dimension block, L n is the relative number of points at the end of the nth radial dimension block; Step 1.2.3: For the shadow image I obtained in step 1.2.1 s Each partition block in the mn (γ n ), the formula is as follows: Where P is the original shadow image I s The number of pixels in a certain partitioned block whose value is 1, Q is the total number of pixels in the partitioned block, and m is the serial number of the angle-dimension partition, m=1, 2, 3, ..., M.

5. The method for inverting significant wave height of ocean waves by marine radar according to claim 2, characterized in that: Calculate the root mean square steepness σ as described in step 1.3 RMS include: Shadow ratio function S of the same radial dimension block area n (γ n ) and lighting ratio L n (γ n ) satisfy: L n (c n )=1-S n (c n ) Wherein, n is the serial number of each radial dimension block, n=1, 2, ..., N; Using the least squares method, the Smith function is used to fit the illumination ratio L of the corresponding N radial blocks in the angle dimension partition. n (γ n ) and the grazing angle γ n The relationship between the root mean square wave steepness σ RMS , the Smith function is: in m n =tan(γ n ) erfc() is the Gaussian error function. The erfc() function is derived from the Gaussian surface probability density function. The expression of the Gaussian surface probability density function is:

6. The method for inverting significant wave height of ocean waves by marine radar according to claim 1, characterized in that: The navigation radar is an X-band navigation radar.

Images