Adaptive dual-polarization radio echo quality control method and system

Abstract

This invention relates to the field of meteorological data processing technology, specifically disclosing an adaptive dual-polarization radio echo quality control method and system. The method includes: acquiring dual-polarization radio volume scan base data; calculating reflectivity correlation, then calculating the local reflectivity field structure consistency index, and determining a local window effectiveness percentage threshold based on isolated anomaly markers; filtering the volume scan base data based on radial texture features and echo top height to obtain retained region data; calculating the reflectivity horizontal texture, differential reflectivity local variance, and correlation coefficient stability term based on the retained region data to generate mixed texture features; calculating the vertical gradient value; performing discrimination according to a distance database to obtain non-meteorological echo data; and performing fine processing and feature regeneration to obtain dual-polarization radio echo data. This invention significantly improves the balance between non-meteorological echo identification accuracy and retention rate by using the structure consistency index and the local window effectiveness percentage threshold.

Description

Technical Field

[0001] This invention relates to the field of meteorological data processing technology, and in particular to an adaptive dual-polarization radio echo quality control method and system. Background Technology

[0002] Dual-polarization radio sounding systems acquire multi-dimensional polarization parameters such as reflectivity factor, differential reflectivity, and correlation coefficient by alternately or simultaneously transmitting horizontally and vertically polarized electromagnetic waves. Compared to conventional single-polarization radio sounding systems, they can more precisely characterize the shape, phase, and spatial distribution of precipitation particles, and have become a core tool for monitoring, warning, and quantitative precipitation estimation of severe convective weather. However, dual-polarization radio sounding systems are often contaminated by various non-meteorological echoes in actual operational use, including sea clutter, electromagnetic interference, biological echoes (such as those from birds), and isolated noise points. Monitoring hazardous weather at sea is particularly crucial for maritime navigation safety.

[0003] Existing echo quality control methods for dual-polarization radio detection systems can be mainly divided into three categories: First, discrimination methods based on single polarization parameter texture features and fixed thresholds, such as using differential reflectivity horizontal textures or correlation coefficient textures to distinguish precipitation echoes from clutter. This type of method has high threshold sensitivity and a high misclassification rate in complex terrain or mixed echo scenarios. Second, multi-parameter fusion methods based on fuzzy logic, which score multiple parameters through membership functions and then make a comprehensive judgment. However, its rule design and weight allocation are highly dependent on human experience, and its generalization ability is limited. Third, deep learning methods that have emerged in recent years, which train semantic segmentation or classification models with labeled quality control results. However, these methods face problems such as scarce labeled samples, poor interpretability, and difficulty in cross-site model transfer.

[0004] The existing methods generally have obvious technical limitations: (1) The discrimination threshold is globally uniform and cannot be adaptively adjusted according to the spatial structure characteristics of the local echo field, making it difficult to balance the retention requirements of precipitation edge areas and the filtering requirements of isolated clutter areas; (2) The construction of texture features is generally limited to horizontal textures in the azimuth direction and does not make full use of high-resolution texture information in the radial distance direction. Therefore, it is urgent to develop a dual-polarization radio echo quality control method that can adapt to the local structure of the echo field, deeply integrate multi-dimensional physical features, and take into account computational efficiency. Summary of the Invention

[0005] This invention aims to solve the aforementioned problems. To this end, this invention provides an adaptive dual-polarization radio echo quality control method and system. Through a local reflectivity field structure consistency index and an adaptive local window effectiveness percentage threshold mechanism, it significantly improves the balance between non-meteorological echo identification accuracy and retention rate. Furthermore, a three-level progressive architecture is designed to balance processing efficiency and discrimination accuracy.

[0006] This invention provides an adaptive dual-polarization radio echo quality control method, the technical solution of which includes: S1: Acquire dual-polarization radio volume scan data and calculate echo top height; the volume scan data includes reflectivity, differential reflectivity, correlation coefficient, azimuth, elevation, range library number, beamwidth, pulse width, and atmospheric refraction model parameters; S2: Based on volume scan baseline data, within each target distance library and its neighborhood, the reflectivity correlation is calculated, and then the local reflectivity field structure consistency index is calculated. Based on the structure consistency index and isolated anomaly markers, the local window effectiveness percentage threshold is determined. S3: Calculate radial texture features based on reflectivity, and then use radial texture features, echo top height and local window validity percentage threshold to filter volume scan base data to obtain retained region data; S4: Based on the data of the preserved area, the reflectance horizontal texture, the local variance of differential reflectance, and the correlation coefficient are calculated using reflectance, differential reflectance, and correlation coefficient stability term to generate mixed texture features; S5: Based on the data of the preserved area, the vertical gradient value is calculated using the beamwidth, pulse width, and atmospheric refraction model parameters; S6: Using mixed texture features and vertical gradient values, discrimination is performed on a distance-by-distance basis to obtain non-meteorological echo data; S7: Fine-tune the non-meteorological echo data in the retained area data, and then perform feature regeneration to fill the misjudgment gaps to obtain dual-polarization radio echo data.

[0007] Furthermore, in step S2, the formula for calculating the structural consistency index is: in, The local reflectivity field structure consistency index of the range library in the range direction x and azimuth direction y; The distance to the window size; This refers to the azimuth window size; For distance-to-neighborhood indexing; For directional neighborhood indexing; The target distance is the reflectivity of the library; The reflectivity of the neighborhood distance library; This is the reflectance correlation function.

[0008] Furthermore, the reflectance correlation function uses the Pearson linear correlation coefficient; the neighborhood window size is set to 10×10 distance libraries.

[0009] Furthermore, the formula for calculating the local window validity percentage threshold is as follows: in, This is the threshold for the percentage of effectiveness of a local window. The first filtering threshold; This is the second filtering threshold; This is a marker of isolated anomalies; The local reflectivity field structure consistency index of the range library in the range direction x and azimuth direction y; As a structural consistency threshold, the second filtering threshold is not lower than the first filtering threshold.

[0010] Furthermore, in step S3, the radial texture features and echo top height are normalized, and a weighted sum is calculated to obtain the screening score; If the screening score is greater than the local window validity percentage threshold, then the data is retained.

[0011] Furthermore, in step S3, the comparison screening score and local window validity percentage threshold are calculated for each distance library.

[0012] Furthermore, in step S4, reflectivity horizontal texture is calculated using reflectivity; local variance of differential reflectivity is calculated using differential reflectivity; and the complement of the mean of correlation coefficients within the neighborhood window is taken to obtain the stable term of correlation coefficients. The reflectivity horizontal texture, differential reflectivity local variance, and correlation coefficient stability term are scale-normalized and weighted sums are calculated to obtain the mixed texture features.

[0013] Furthermore, in step S6, if the mixed texture feature is less than the texture discrimination threshold and the vertical gradient value is less than the gradient discrimination threshold, it is determined to be precipitation echo data; otherwise, it is determined to be non-meteorological echo data.

[0014] Furthermore, in step S6, the distance library of precipitation echo data is labeled as precipitation echo; the distance library of non-meteorological echo data is labeled as non-meteorological echo. In step S7, the number of the distance reservoir marked as precipitation echo is output.

[0015] This invention also provides an adaptive dual-polarization radio echo quality control system, the technical solution of which is as follows: including: The data acquisition module is used to acquire dual-polarization radio volume scan data and calculate the echo top height; the volume scan data includes reflectivity, differential reflectivity, correlation coefficient, azimuth angle, elevation angle, range library number, beamwidth, pulse width and atmospheric refraction model parameters; The window validity threshold calculation module is used to calculate reflectance correlation within each target distance library and its neighborhood based on volume scan baseline data, then calculate the local reflectance field structure consistency index, and determine the local window validity percentage threshold based on the structure consistency index and isolated anomaly markers. The data filtering module is used to calculate radial texture features based on reflectivity, and then use the radial texture features, echo top height and local window validity percentage threshold to filter the volume scan base data to obtain the retained region data; The hybrid texture feature generation module is used to calculate the horizontal texture of reflectance, the local variance of differential reflectance, and the stable term of correlation coefficient based on the data of the preserved region using reflectance, differential reflectance, and correlation coefficient, and generate hybrid texture features. The vertical gradient calculation module is used to calculate the vertical gradient value based on the preserved region data, using beamwidth, pulse width, and atmospheric refraction model parameters. The echo discrimination module is used to perform discrimination based on a distance library using mixed texture features and vertical gradient values ​​to obtain non-meteorological echo data. The data processing module is used to perform fine processing on the non-meteorological echo data in the retained area data, and then perform feature regeneration to fill the misjudgment gaps to obtain dual-polarization radio echo data.

[0016] The above-described one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: 1. The local reflectivity field structure consistency index and adaptive local window effectiveness percentage threshold mechanism proposed in this invention can dynamically adjust the strictness of the judgment according to the spatial continuity characteristics of the echo field. In structurally continuous regions (such as the main body and edges of precipitation), a relaxed first filtering threshold is used to maximize the retention of real precipitation echoes, while a strict second filtering threshold is used in isolated anomalous regions to effectively filter out discrete clutter and noise. This achieves a fine balance between precipitation retention and non-meteorological echo filtering, overcoming the inherent defects of traditional fixed threshold methods that tend to compromise on one aspect while neglecting another.

[0017] 2. The hybrid texture feature constructed in this invention weighted and fused three parameters: horizontal reflectance texture, local variation of differential reflectance, and correlation coefficient stability term. At the same time, it introduced texture features along the radial distance direction, which broke through the limitations of existing methods that only rely on azimuth horizontal texture or independent discrimination of two parameters. It can comprehensively characterize target features from three dimensions: echo spatial heterogeneity, local variation of polarization information, and stability of polarization characteristics, which significantly improves the ability to distinguish different types of non-meteorological echoes such as ground clutter, sea clutter, and electromagnetic interference.

[0018] 3. This invention designs a hierarchical and progressive architecture of rapid screening (step S3), accurate discrimination (step S6), and fine processing and feature regeneration (step S7). In the rapid screening stage, radial texture and echo top height with low computational cost are used to quickly exclude obvious non-precipitation data and release the cache. In the accurate discrimination stage, high-resolution mixed texture features and vertical gradient values ​​are used for accurate classification. In the fine processing stage, suspected misjudged areas are re-evaluated with a larger neighborhood window. While ensuring accuracy, the overall computational cost is effectively controlled, making it suitable for real-time business processing of large-scale volume scan data.

[0019] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a flowchart of the method provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. The following embodiments are used to illustrate this invention but should not be used to limit the scope of this invention.

[0023] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0024] The following is combined Figure 1 The present invention will be further described in detail below, providing an adaptive dual-polarization radio echo quality control method and system: In this embodiment, as Figure 1 As shown, an adaptive dual-polarization radio echo quality control method is provided, comprising the following steps: S1: Acquire dual-polarization radio volume scan data and calculate echo top height; the volume scan data includes reflectivity, differential reflectivity, correlation coefficient, azimuth, elevation, range library number, beamwidth, pulse width, and atmospheric refraction model parameters.

[0025] This step receives volume scan data from the dual-polarization weather radio system via its operational processing link. This volume scan data includes reflectivity, differential reflectivity, correlation coefficient, azimuth, elevation, range library number, volume scan timestamp, beamwidth, pulse width, and atmospheric refraction model parameters. The echo top height is calculated from the original volume scan data using a standard operational algorithm, such as the highest elevation height interpolation method with a maximum reflectivity factor ≥18 dBZ.

[0026] This step also performs consistency verification on the received volume scan baseline data, including: verifying the continuity of the distance library numbers, the registration status of azimuth and elevation angles, the continuity of volume scan timestamps, the quality indicators of differential reflectivity and correlation coefficients, and the product validity indicator of echo top height. If any key parameter is missing, the volume scan baseline data acquired in this step cannot be used to update thresholds, weights, and lookup table compensation parameters in subsequent steps.

[0027] S2: Based on volume scan baseline data, calculate reflectivity correlation within each target distance library and its neighborhood, then calculate the local reflectivity field structure consistency index, and determine the local window effectiveness percentage threshold based on the structure consistency index and isolated anomaly markers.

[0028] A range and azimuth neighborhood window is formed centered on the target range library. The correlation between the reflectance of the target range library and the reflectance of the neighboring range libraries is calculated. Then, the structural consistency index of the local reflectance field is calculated. The higher the structural consistency index, the stronger the spatial continuity of the target range library and its neighborhood, and the more likely it is to belong to the edge or main body of continuous precipitation. The formula for calculating the structural consistency index is: in, The local reflectivity field structure consistency index of the range library in the range direction x and azimuth direction y; The distance to the window size; This refers to the azimuth window size; For distance-to-neighborhood indexing; For directional neighborhood indexing; The target distance is the reflectivity of the library; The reflectivity of the neighborhood distance library; The reflectance correlation function is used. In this embodiment, the reflectance correlation function adopts the Pearson linear correlation coefficient, and the neighborhood window size is set to 10 (distance direction) × 10 (azimuth direction) distance libraries by default.

[0029] Based on the structural consistency index and isolated anomaly markers, a local window validity percentage threshold is determined. A first filtering threshold is applied to structurally continuous regions, while a second filtering threshold, not lower than the first, is applied to isolated anomaly regions. The formula for calculating the local window validity percentage threshold is as follows: in, This is the threshold for the percentage of effectiveness of the local window, with a value of [0,1]. The first filtering threshold; This is the second filtering threshold; This is a marker of isolated anomalies; It is the consistency index of the local reflectivity field structure; This is the structural consistency threshold.

[0030] The second filtering threshold is not lower than the first filtering threshold and is used to increase the filtering intensity of isolated outliers. In this embodiment, the second filtering threshold is determined by multiplying the first filtering threshold by an expansion factor, and the expansion factor ranges from [1.2, 2.0]. The first filtering threshold, the expansion factor, and the structural consistency threshold are all determined by a dataset composed of sea clutter samples, precipitation edge samples, and continuous precipitation samples.

[0031] Isolated anomaly flags for each distance database are generated according to the following rules: A 5×5 neighborhood window is taken centered on the target distance database, and the proportion of valid reflectance samples within the neighborhood is calculated; if this proportion is lower than the isolation determination threshold (e.g., 0.3), and the difference between the reflectance of the target distance database itself and the median reflectance of the neighborhood exceeds a preset threshold (e.g., 15 dBz), then the isolated anomaly flag is set to valid. =1), otherwise set to invalid ( =0).

[0032] The neighborhood window size of the reflectivity correlation function should be greater than or equal to the neighborhood window size of the isolated anomaly marker. In this embodiment, the ratio of the two is determined by the beamwidth and volume scan mode of the radio system. The ratio can be appropriately increased when the beam is narrower and the spatial resolution is higher, so that the physical scale covered by the neighborhood window size of the reflectivity correlation function remains basically constant.

[0033] S3: Calculate radial texture features based on reflectivity, and then use radial texture features, echo top height, and local window validity percentage threshold to filter volume scan base data to obtain retained region data.

[0034] In this embodiment, radial texture features and echo top height are normalized, a weighted sum is calculated to obtain a screening score, and combined with a local window validity percentage threshold, precipitation echo data is quickly screened to obtain retained area data. Specifically, radial texture is calculated along the distance direction using reflectivity and normalized (maximum and minimum normalization) to obtain normalized radial texture features. Simultaneously, echo top height is normalized to obtain normalized echo top height. Then, the screening score is calculated using the following formula: in, The score is set to [0,1] for filtering purposes; Radial texture weight; Normalized radial texture features; Weighting of echo tops; The echo top height is normalized. A higher screening score indicates a higher likelihood of precipitation echoes. Radial texture weights and echo top height weights are determined using the dataset described above.

[0035] The comparison and filtering scores for each distance library are compared against a local window validity percentage threshold. If the filtering score is greater than the local window validity percentage threshold, a reserved region flag is added to that distance library. After comparing all distance libraries, the reserved region data is obtained. The intermediate computation cache for non-reserved region data is released.

[0036] S4: Based on the data of the preserved area, using reflectivity, differential reflectivity and correlation coefficient, calculate the horizontal texture of reflectivity, the local variance of differential reflectivity and the stable term of correlation coefficient to generate mixed texture features.

[0037] For each distance library in the retained area data, the horizontal reflectance texture is calculated using reflectance within a neighborhood window, based on the azimuth and distance library number. The local local deviation of the differential reflectance is obtained by taking the standard deviation of the differential reflectance within the neighborhood window. The stable term of the correlation coefficient is obtained by taking the complement of the mean of the correlation coefficients within the neighborhood window. The default neighborhood window size is 5×5 distance libraries.

[0038] Scale normalization is performed on the horizontal reflectance texture, the local variance of differential reflectance, and the stability term of the correlation coefficient. The normalized components are then used to generate hybrid texture features, calculated using the following formula: in, It is a blended texture feature; For normalized reflectivity horizontal texture; The variance weights are the differential reflectance values. This is the local variation term of the normalized differential reflectance; Weights for the stable term of the correlation coefficient; This represents the stable term for the normalized correlation coefficient. In this embodiment, max-min normalization is used. The weights of the differential reflectance variance and the stable term for the correlation coefficient are also determined by the dataset.

[0039] S5: Based on the data of the preserved area, the vertical gradient value is calculated using the beamwidth, pulse width, and atmospheric refraction model parameters.

[0040] In this embodiment, for each distance library of the retained area data, the corresponding theoretical reflectivity vertical gradient expectation value is determined based on the beamwidth, pulse width, and atmospheric refraction model parameters, and then corrected to obtain the corrected vertical gradient value.

[0041] Based on the elevation angle, range station number, and range station length, combined with the antenna altitude of the radio system in the site calibration parameters and the equivalent Earth radius factor, the true altitude of the range station above sea level is calculated.

[0042] Based on beamwidth, pulse width, atmospheric refraction model parameters, and altitude, a standard vertical gradient lookup table is used to determine the corresponding expected value of the theoretical reflectivity vertical gradient. In this embodiment, a standard vertical gradient lookup table is pre-established based on beamwidth, pulse width, atmospheric refraction model parameters, and altitude layers. The construction method is as follows: Under standard atmospheric refraction conditions, for each altitude layer (step size 0.1 km, range 0-4 km), the expected value of the theoretical reflectivity vertical gradient corresponding to that altitude is calculated according to the equations of the radio system and the beam propagation model, and stored as a three-dimensional lookup table with (beamwidth, pulse width, atmospheric refraction model parameters, altitude) as the joint index. The gradient compensation coefficient is obtained by monthly fitting of historical radiosonde data or numerical model refractive index profiles, reflecting the systematic deviation of actual atmospheric refraction from standard atmospheric refraction. The altitude compensation function adopts a piecewise linear function form, and the piecewise nodes and slopes are determined by statistical regression of measured vertical profiles of precipitation echoes from historical volume scans. When the lookup exceeds the limit, the nearest table entry or boundary value is used.

[0043] The corrected formula for calculating the vertical gradient value is as follows: in, This is the corrected vertical gradient value; This represents the expected value of the vertical gradient of the theoretical reflectivity. Beamwidth; The pulse width; These are the parameters for the atmospheric refraction model; For height; These are the gradient compensation coefficients; This is the height compensation function.

[0044] S6: Using mixed texture features and vertical gradient values, discrimination is performed on a distance-by-distance basis to obtain non-meteorological echo data.

[0045] In this embodiment, mixed texture features and vertical gradient values ​​are used together to distinguish between precipitation echoes and non-meteorological echoes.

[0046] The specific discrimination rules are as follows: when the mixed texture feature is less than the texture discrimination threshold and the vertical gradient value is less than the gradient discrimination threshold, it is judged as precipitation echo data; when the mixed texture feature is greater than or equal to the texture discrimination threshold, or the vertical gradient value is greater than or equal to the gradient discrimination threshold, it is judged as non-meteorological echo data. Both thresholds (texture discrimination threshold and gradient discrimination threshold) are calibrated on the dataset with the objective of maximizing the Heidke skill score (HSS). Precipitation echo data is labeled as precipitation echo in the distance library; non-meteorological echo data is labeled as non-meteorological echo in the distance library.

[0047] S7: Fine-tune the non-meteorological echo data in the retained area data, and then perform feature regeneration to fill the misjudgment gaps to obtain dual-polarization radio echo data.

[0048] In this embodiment, for non-meteorological echo data in the retained area data, discrimination is performed on a distance-by-distance basis. When the echo top height exceeds the secondary processing height threshold (3km), the retained area label is valid, and the gradient quality score is not lower than the gradient quality threshold, fine processing is activated. The mixed texture features are recalculated with a larger neighborhood window (e.g., 15×15 distance databases), and discrimination is performed in conjunction with the vertical gradient limited by the height threshold. Distance databases misclassified as non-meteorological echoes are marked as precipitation echoes, resulting in the processed retained area data. Precipitation echo data in the retained area data is not processed and is part of the processed retained area data. The gradient quality score is determined by the lookup table hit status, interpolation residuals, and the validity of altitude data. The calculation method is as follows: (1) Determine the lookup table hit status score: When the beamwidth, pulse width, atmospheric refraction model parameters, and altitude corresponding to the distance library are accurately hit in the standard vertical gradient lookup table, a high score is given. When only adjacent entries are hit or boundary values ​​are used, the score is reduced according to the degree of deviation. (2) Calculate the interpolation residual score: When performing multidimensional linear interpolation between the four-dimensional grid nodes of the lookup table, the normalized residual between the interpolation result and the value of the nearest grid node is taken. The smaller the residual, the higher the score. (3) Evaluate the validity score of altitude data: It is determined by the quality indicators and measurement uncertainty of parameters such as elevation angle, distance library number, antenna altitude, and equivalent Earth radius factor involved in altitude calculation. The highest score is given when the altitude value comes from valid observations and all input parameters pass quality verification. The lookup table hit status score, interpolation residual score, and altitude data validity score are weighted and summed according to preset weights to obtain the gradient quality score of the distance library. The higher the score, the stronger the reliability of the vertical gradient value calculation of the distance library. The gradient quality threshold is determined by historical valid lookup table samples and misjudged void samples.

[0049] Feature regeneration is performed based on the processed retained area data. The scan proceeds layer by layer from high elevation angle to low elevation angle, filling only the voids connecting adjacent precipitation areas. The value of the distance library is restored to the value of the original volume scan baseline data, thus obtaining the feature regenerated data.

[0050] Finally, dual-polarization radio echo data is output, including processed retained area data and feature regeneration data. Precipitation echo markers are also output, specifically the distance library numbers tagged with precipitation echoes.

[0051] The processed data were examined, and the three-dimensional high-resolution data after suppression of non-meteorological clutter was saved as the results. The comparison results of the data before and after suppression are shown in Table 1.

[0052] Table 1

[0053] This embodiment also provides an adaptive dual-polarization radio echo quality control system, which employs the following technical solution: including: The data acquisition module is used to acquire dual-polarization radio volume scan data and calculate the echo top height; the volume scan data includes reflectivity, differential reflectivity, correlation coefficient, azimuth angle, elevation angle, range library number, beamwidth, pulse width and atmospheric refraction model parameters; The window validity threshold calculation module is used to calculate reflectance correlation within each target distance library and its neighborhood based on volume scan baseline data, then calculate the local reflectance field structure consistency index, and determine the local window validity percentage threshold based on the structure consistency index and isolated anomaly markers. The data filtering module is used to calculate radial texture features based on reflectivity, and then use the radial texture features, echo top height and local window validity percentage threshold to filter the volume scan base data to obtain the retained region data; The hybrid texture feature generation module is used to calculate the horizontal texture of reflectance, the local variance of differential reflectance, and the stable term of correlation coefficient based on the data of the preserved region using reflectance, differential reflectance, and correlation coefficient, and generate hybrid texture features. The vertical gradient calculation module is used to calculate the vertical gradient value based on the preserved region data, using beamwidth, pulse width, and atmospheric refraction model parameters. The echo discrimination module is used to perform discrimination based on a distance library using mixed texture features and vertical gradient values ​​to obtain non-meteorological echo data. The data processing module is used to perform fine processing on the non-meteorological echo data in the retained area data, and then perform feature regeneration to fill the misjudgment gaps to obtain dual-polarization radio echo data.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An adaptive dual-polarization radio echo quality control method, characterized in that, include: S1: Acquire dual-polarization radio volume scan data and calculate echo top height; the volume scan data includes reflectivity, differential reflectivity, correlation coefficient, azimuth, elevation, range library number, beamwidth, pulse width, and atmospheric refraction model parameters; S2: Based on volume scan baseline data, within each target distance library and its neighborhood, the reflectivity correlation is calculated, and then the local reflectivity field structure consistency index is calculated. Based on the structure consistency index and isolated anomaly markers, the local window effectiveness percentage threshold is determined. S3: Calculate radial texture features based on reflectivity, and then use radial texture features, echo top height and local window validity percentage threshold to filter volume scan base data to obtain retained region data; S4: Based on the data of the preserved area, the reflectance horizontal texture, the local variance of differential reflectance, and the correlation coefficient are calculated using reflectance, differential reflectance, and correlation coefficient stability term to generate mixed texture features; S5: Based on the data of the preserved area, the vertical gradient value is calculated using the beamwidth, pulse width, and atmospheric refraction model parameters; S6: Using mixed texture features and vertical gradient values, discrimination is performed on a distance-by-distance basis to obtain non-meteorological echo data; S7: Fine-tune the non-meteorological echo data in the retained area data, and then perform feature regeneration to fill the misjudgment gaps to obtain dual-polarization radio echo data.

2. The adaptive dual-polarization radio echo quality control method as described in claim 1, characterized in that, In step S2, the formula for calculating the structural consistency index is: in, The local reflectivity field structure consistency index of the range library in the range direction x and azimuth direction y; The distance to the window size; This refers to the azimuth window size; For distance-to-neighborhood indexing; For directional neighborhood indexing; The target distance is the reflectivity of the library; The reflectivity of the neighborhood distance library; This is the reflectance correlation function.

3. The adaptive dual-polarization radio echo quality control method as described in claim 2, characterized in that, The reflectance correlation function uses the Pearson linear correlation coefficient; the neighborhood window size is 10×10 distance libraries.

4. The adaptive dual-polarization radio echo quality control method as described in claim 1, characterized in that, The formula for calculating the local window validity percentage threshold is: in, This is the threshold for the percentage of effectiveness of a local window. The first filtering threshold; This is the second filtering threshold; This is a marker of isolated anomalies; The local reflectivity field structure consistency index of the range library in the range direction x and azimuth direction y; As a structural consistency threshold, the second filtering threshold is not lower than the first filtering threshold.

5. The adaptive dual-polarization radio echo quality control method as described in claim 1, characterized in that, In step S3, the radial texture features and echo top height are normalized, and a weighted sum is calculated to obtain the screening score; If the screening score is greater than the local window validity percentage threshold, then the data is retained.

6. The adaptive dual-polarization radio echo quality control method as described in claim 5, characterized in that, In step S3, the comparison screening score of the distance library and the local window validity percentage threshold are calculated.

7. The adaptive dual-polarization radio echo quality control method as described in claim 1, characterized in that, In step S4, reflectance horizontal texture is calculated using reflectance; local variance of differential reflectance is calculated using differential reflectance; the complement of the mean of correlation coefficients within the neighborhood window is taken to obtain the stable term of correlation coefficients. The reflectivity horizontal texture, differential reflectivity local variance, and correlation coefficient stability term are scale-normalized and weighted sums are calculated to obtain the mixed texture features.

8. The adaptive dual-polarization radio echo quality control method as described in claim 1, characterized in that, In step S6, if the mixed texture feature is less than the texture discrimination threshold and the vertical gradient value is less than the gradient discrimination threshold, it is determined to be precipitation echo data; otherwise, it is determined to be non-meteorological echo data.

9. The adaptive dual-polarization radio echo quality control method as described in claim 8, characterized in that, In step S6, the distance library of precipitation echo data is labeled as precipitation echo; the distance library of non-meteorological echo data is labeled as non-meteorological echo. In step S7, the number of the distance reservoir marked as precipitation echo is output.

10. An adaptive dual-polarization radio echo quality control system, characterized in that, An adaptive dual-polarization radio echo quality control method as described in any one of claims 1 to 9, comprising: The data acquisition module is used to acquire dual-polarization radio volume scan data and calculate the echo top height; the volume scan data includes reflectivity, differential reflectivity, correlation coefficient, azimuth angle, elevation angle, range library number, beamwidth, pulse width and atmospheric refraction model parameters; The window validity threshold calculation module is used to calculate reflectance correlation within each target distance library and its neighborhood based on volume scan baseline data, then calculate the local reflectance field structure consistency index, and determine the local window validity percentage threshold based on the structure consistency index and isolated anomaly markers. The data filtering module is used to calculate radial texture features based on reflectivity, and then use the radial texture features, echo top height and local window validity percentage threshold to filter the volume scan base data to obtain the retained region data; The hybrid texture feature generation module is used to calculate the horizontal texture of reflectance, the local variance of differential reflectance, and the stable term of correlation coefficient based on the data of the preserved region using reflectance, differential reflectance, and correlation coefficient, and generate hybrid texture features. The vertical gradient calculation module is used to calculate the vertical gradient value based on the preserved region data, using beamwidth, pulse width, and atmospheric refraction model parameters. The echo discrimination module is used to perform discrimination based on a distance library using mixed texture features and vertical gradient values ​​to obtain non-meteorological echo data. The data processing module is used to perform fine processing on the non-meteorological echo data in the retained area data, and then perform feature regeneration to fill the misjudgment gaps to obtain dual-polarization radio echo data.

Images