Vertical wind evaluation method based on multi-beam consistency score and sliding window processing

Abstract

The application discloses a vertical airflow evaluation method based on multi-beam consistency scoring and sliding window processing. The method comprises the following steps: S1, obtaining multi-beam radial velocity data of a laser Doppler radar under different pitch angle and azimuth angle conditions; S2, uniformly mapping data of observation distance layers corresponding to different pitch angles to a target height by using a spatial interpolation method; S3, based on the observation geometry relationship of each beam, a plurality of sets of vertical airflow estimation results are obtained through multi-beam combination inversion; S4, sliding window smoothing processing is introduced to each vertical airflow sequence to obtain a smoothed sequence; S5, based on the plurality of sets of smoothed vertical airflow results, the correlation degree between different results is calculated by using a correlation measurement function, and a consistency matrix is constructed; S6, the single model consistency score and the overall consistency score are calculated according to the consistency matrix, and quantitative evaluation of the consistency of the vertical airflow result is realized. The application belongs to the technical field of laser radar meteorological observation.

Description

Technical Field

[0001] This invention belongs to the field of lidar meteorological observation technology, specifically involving a vertical airflow assessment method based on multi-beam consistency scoring and sliding window processing. It is applicable to the consistency assessment of vertical airflow inversion results, spatial benchmark unification, and data smoothing processing in scenarios such as meteorological monitoring, low-altitude aviation safety, wind farm operation and maintenance, and atmospheric environment monitoring. Background Technology

[0002] Laser Doppler radar is a core piece of equipment for vertical airflow detection. Currently, it mainly uses two methods to obtain the vertical wind component: four-beam inversion and direct measurement with a fifth beam (vertically pointing). However, existing technologies have the following three major drawbacks: 1. Difficulty in determining the standard value of vertical wind component: The lack of high-precision, long-term stable reference measurement methods makes it impossible to directly obtain the true value of the vertical wind component. This results in a lack of a unified standard reference for the validity judgment of four-beam inversion results and fifth-beam measurement results, making accurate error analysis difficult.

[0003] 2. Mismatch between the range layer measured by the fifth beam and the target range layer: Due to the limitations of radar ranging resolution and beam geometry, the effective range layer directly measured by the fifth beam (vertically pointing) and the target range layer corresponding to the inversion of the fourth beam (non-vertically pointing) often have spatial deviations, which disrupts the spatial correspondence between different measurement results and makes data fusion difficult.

[0004] 3. Significant error components in direct measurement of vertical components, making it difficult to extract the true value: The actual value of the radar elevation angle may deviate from the set value (90°). The interference of random noise makes the vertical component directly measured by the fifth beam contain a large number of error components, making it difficult to effectively extract the true vertical airflow information.

[0005] Existing technologies typically treat four-beam inversion and fifth-beam measurement as independent data sources, without designing specific processing mechanisms to address the aforementioned issues, nor establishing a consistency evaluation system among multiple sets of observation results. Therefore, there is an urgent need for a vertical airflow assessment method that can unify spatial benchmarks, suppress noise interference, and establish a consistent evaluation approach. Summary of the Invention

[0006] The purpose of this invention is to overcome the aforementioned shortcomings in the prior art and provide a vertical airflow assessment method based on multi-beam consistency scoring and sliding window processing. This method aims to solve the assessment problem caused by the lack of standard values ​​for vertical wind components, eliminate the spatial mismatch between fifth-beam and four-beam observation results, and suppress random noise interference, thereby improving the reliability and practicality of vertical airflow observation data.

[0007] The technical solution adopted in this invention is a vertical airflow assessment method based on multi-beam consistency scoring and sliding window processing, which includes the following steps: S1. Acquire multi-beam radial velocity data of laser Doppler radar under different elevation and azimuth angles to form a multi-angle observation dataset; S2. Spatial interpolation method is used to uniformly map the data of the observation distance layer corresponding to different pitch angles to the target height, so as to eliminate the spatial inconsistency caused by the difference in observation geometry. S3. Based on the observation geometry of each beam, multiple sets of vertical airflow estimation results are obtained through multi-beam combination inversion; S4. A sliding window smoothing process is introduced for each vertical airflow sequence to obtain a smoothed sequence, thereby reducing the influence of random noise. S5. Based on the smoothed vertical airflow results, the correlation between different results is calculated using a correlation metric function, and a consistency matrix is ​​constructed. S6. Calculate the single-model consistency score and the overall consistency score based on the consistency matrix to achieve a quantitative assessment of the consistency of the vertical airflow results.

[0008] Specifically, step S2 is as follows: S21. Set the target vertical detection range as a unified reference height; S22. Select two distance layers adjacent to the target height and use the two-point Lagrange interpolation method to calculate the radial velocity at the target distance layer.

[0009] Specifically, step S3 is as follows: S31. Under non-vertical pitch angle conditions, establish a radial velocity model based on the observation geometry. S32. Construct a multi-beam combination for inversion, wherein the multi-beam combination includes: a combination of a first beam and a third beam, a combination of a second beam and a fourth beam, and a four-beam combination; S33. Under the condition of vertical pitch angle, the radial velocity is averaged to obtain the measurement result of vertical direction.

[0010] Specifically, in step S4, the sliding window smoothing process for each vertical airflow sequence includes: S41. Set the sliding window length N ; S42. Calculate the moving average of the vertical airflow sequence over time. The calculation method is as follows: , in, t For time, This is the original vertical airflow sequence. This is the smoothed sequence.

[0011] Specifically, in step S5, calculating the degree of correlation between different results using a correlation measurement function includes: S51. Select the smoothed vertical airflow sequence and ; S52. Calculate its linear correlation coefficient. R ij As a measure of relevance, the formula is as follows: , in, R ij To determine the degree of correlation between different results, i, j For different vertical airflow sequences, For the first i, j The result of smoothing the combined vertical airflow sequence.

[0012] Specifically, in step S5, the construction of the consistency matrix is ​​as follows:

[0013] in, This represents the total number of types of multibeam combinations.

[0014] Specifically, in step S6, calculating the single-model consistency score based on the consistency matrix involves calculating the single-model consistency score according to the following formula: , in, S i For single-model consistency scoring.

[0015] Specifically, step S6 also includes calculating the average of all individual model consistency scores as the overall consistency score under the current observation conditions using the following formula: , in, S Score the overall consistency.

[0016] Specifically, the scanning process of the laser Doppler radar is as follows: under the first elevation angle condition, the four azimuth angles are scanned sequentially, and then under the second elevation angle condition, the same four azimuth angles are scanned to complete one cycle of multi-beam observation.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention achieves spatial benchmark unification between the fifth beam measurement range layer and the target range layer through a spatial interpolation algorithm, thus solving the problem of spatial mismatch between different observation results; 2. Sliding window processing is used to suppress random noise interference and reduce the noise component in direct measurement and inversion results; 3. Construct a multi-beam consistency scoring system to provide a consistency evaluation approach for vertical airflow observation results and make up for the evaluation difficulties caused by the lack of standard values ​​for vertical wind components; 4. This method provides a complete technical path for spatial unification, noise suppression, and consistency evaluation of vertical airflow observation by laser Doppler radar, which can effectively support the validity judgment of observation results and improve the reliability and practicality of vertical airflow observation data. Attached Figure Description

[0018] Figure 1 This is a block diagram illustrating the principle of the method of the present invention; Figure 2 This is a simplified flowchart of the method of the present invention. Detailed Implementation

[0019] like Figure 1 As shown, the method of this invention is based on multi-angle observation of three-dimensional wind fields by laser Doppler radar, using different elevation angles. With azimuth Acquiring multibeam radial velocity data under certain conditions This enables multi-view detection of three-dimensional wind fields.

[0020] By utilizing the observation geometry of each beam, the radial velocity is... With wind speed component Establish a mapping model Multiple sets of vertical airflow estimation results were obtained through multi-beam combination inversion. .

[0021] For different pitch angles The corresponding vertical airflow results are used to map different distance layers to the target distance layer using an interpolation algorithm. To eliminate spatial inconsistencies caused by differences in observation geometry, a sliding window algorithm is used to smooth the data, thereby improving the stability and comparability of the results.

[0022] Based on this, a consistency relationship is constructed among multiple sets of vertical airflow results. Combined vertical airflow Through correlation measurement function The degree of correlation between different outcomes is described to form a linear correlation matrix, and then the overall consistency score is calculated. It is used to characterize the overall consistency level of vertical airflow results under the current observation conditions.

[0023] The present invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0024] like Figure 2 As shown, this invention provides a vertical airflow assessment method based on multibeam consistency scoring and sliding window processing. The method includes the following steps: S1. Obtain laser Doppler radar data at different elevation angles. With azimuth Multibeam radial velocity data under certain conditions This forms a multi-angle observation dataset; S2. Use spatial interpolation to uniformly map the data of the observation distance layers corresponding to different pitch angles to the target height. To eliminate spatial inconsistencies caused by differences in observation geometry; for example, for adjacent distance layers. , Vertical airflow at the location , An estimate of the target height can be obtained through linear interpolation: ; After spatial unification is completed, the geometric relationship between radial velocity and wind speed components is established based on the pitch and azimuth angles of each beam: ; S3. Based on the observation geometry of each beam, multiple sets of vertical airflow estimation results are obtained through multi-beam combination inversion. ; S4. For each vertical airflow sequence By introducing a sliding window smoothing process, a smoothed sequence is obtained. To reduce the impact of random noise; its calculation method is as follows: , in, t For time, This is the original vertical airflow sequence. The smoothed sequence, This reduces the impact of random noise and improves data stability.

[0025] S5. Based on the smoothed multiple sets of vertical airflow results, a correlation measurement function is used. Calculate the degree of correlation between different results and construct a consistency matrix, where, For the first i, j The result of smoothing the combined vertical airflow sequence; S6. Calculate the single-model consistency score and the overall consistency score based on the consistency matrix to achieve a quantitative assessment of the consistency of the vertical airflow results. The consistency matrix is: , in This represents the total number of multi-beam combination types. Based on this, the single-model consistency score is calculated using the consistency matrix:

[0026] And overall consistency score:

[0027] The above scoring indicators enable a quantitative assessment of the consistency of vertical airflow results under different pitch angles and different inversion methods.

[0028] Specifically, step S2 is as follows: S21. Set the target vertical detection range as a unified reference height; S22. Select two distance layers adjacent to the target height and use the two-point Lagrange interpolation method to calculate the radial velocity at the target distance layer.

[0029] Specifically, step S3 is as follows: S31. Under non-vertical pitch angle conditions, establish a radial velocity model based on the observation geometry. S32. Construct a multi-beam combination for inversion, wherein the multi-beam combination includes: a combination of a first beam and a third beam, a combination of a second beam and a fourth beam, and a four-beam combination; S33. Under the condition of vertical pitch angle, the radial velocity is averaged to obtain the measurement result of vertical direction.

[0030] In step S4, the sliding window smoothing process for each vertical airflow sequence specifically includes: S41. Set the sliding window length N ; S42. Calculate the moving average of the vertical airflow sequence over time. The calculation method is as follows: , in, t For time, This is the original vertical airflow sequence. This is the smoothed sequence.

[0031] In step S5, calculating the degree of correlation between different results using a correlation measurement function specifically includes: S51. Select the smoothed vertical airflow sequence and ; S52. Calculate its linear correlation coefficient. R ij As a measure of relevance, the formula is as follows: , in,R ij To determine the degree of correlation between different results, i, j For different vertical airflow sequences, For the first i, j The result of smoothing the combined vertical airflow sequence.

[0032] Specifically, in step S5, the construction of the consistency matrix is ​​as follows:

[0033] in, This represents the total number of types of multibeam combinations.

[0034] In step S6, calculating the single-model consistency score based on the consistency matrix specifically involves calculating the single-model consistency score based on the consistency matrix using the following formula: , in, S i For single-model consistency scoring.

[0035] Step S6 also includes calculating the average of all individual model consistency scores as the overall consistency score under the current observation conditions using the following formula: , in, S Score the overall consistency.

[0036] The scanning process of the laser Doppler radar is as follows: under the first elevation angle condition, four azimuth angles are scanned sequentially, and then the same four azimuth angles are scanned under the second elevation angle condition to complete one cycle of multi-beam observation.

[0037] The spatial interpolation method is applied to map observation data at the vertical pitch angle (90°) to the target detection height corresponding to the non-vertical pitch angle, in order to solve the spatial mismatch problem between the fifth beam range measurement layer and the four beam inverted target range layer.

[0038] The present invention will now be described with reference to more specific embodiments.

[0039] This invention provides a vertical airflow assessment method based on multi-beam consistency scoring. It uses laser Doppler wind radar observation data for a specific time period as the implementation object. The original observation data all originate from a fixed distance layer. After constructing a unified spatial reference through interpolation, consistency analysis is performed. The specific steps are as follows: Step S1: Acquisition of multi-beam radial velocity data Step S11: Set the laser Doppler radar scanning strategy. Use a single 3D laser radar for periodic scanning. The scanning process is as follows: First, at the elevation angle... Scan the four azimuth angles sequentially under the condition Then at the pitch angle Under the same conditions, scan the same four azimuth angles to complete one cycle of multibeam observation; Step S12: Acquire radial velocity data corresponding to each beam during the above scanning process. The data are all from a distance layer of 480m; Step S13: Match and group the data according to timestamp and beam number to construct a multi-beam radial velocity data set under different pitch and azimuth angles.

[0040] Step S2, Spatial Interpolation Algorithm Step S21, due to pitch angle The observation direction is vertically upward, and its corresponding radial observation distance is... Due to differences in spatial location under different conditions, spatial uniform processing is required to achieve comparative analysis of observation results at different elevation angles in the same spatial location. Step S22: Using the target vertical detection distance of 464m as a unified reference height, a spatial interpolation method is introduced to adjust the pitch angle. The radial velocity data under the given conditions are reconstructed; Step S23: Select adjacent distance layers m、 m, the radial velocity at the target distance layer is calculated using the two-point Lagrange interpolation method:

[0041] Step S24: Perform the above interpolation operation on the radial velocity data under all time and beam conditions to obtain a unified radial velocity dataset corresponding to the 464 m distance layer, which will be used for subsequent vertical airflow inversion and consistency analysis.

[0042] Step S3: Multibeam Combination and Vertical Airflow Inversion Step S31, in Establish a radial velocity model based on observed geometric relationships under the following conditions: , Step S32, in Under the condition of constructing a multi-beam combination for inversion: Beam combination 1 and 3; Beam combination 2 and 4; Four-beam combination; Step S33, in Averaging the radial velocities under the given conditions yields:

[0043] Step S34: Form multiple vertical airflow sequences:

[0044] in, In order to be in Radial velocities directly measured in beams 1, 2, 3, and 4.

[0045] Step S4: Smoothing of sliding windows Step S41: Perform sliding window averaging on each vertical airflow sequence; Step S42, the sliding window calculation formula is: , Step S43: Select the sliding window length Perform smoothing processing.

[0046] Step S5: Consistency Matrix Construction Step S51: Based on the smoothed vertical airflow sequence, recalculate the correlation between each sequence: , Step S52: Calculate the linear correlation coefficients between the vertical airflow sequences under the condition of no sliding window, and construct the consistency matrix:

[0047] Step S53, after sliding window processing (with (For example) Construct a consistency matrix:

[0048] Step S6: Calculation of Consistency Score Step S61: Calculate univariate scores based on the consistency matrix:

[0049] Step S62: Calculate the overall consistency score:

[0050] Step S63, the scoring results without sliding window are shown in the table below:

[0051] Overall rating without sliding window .

[0052] Step S64, Scoring after sliding window processing: The ratings for different lengths of sliding windows are as follows, with the subscript indicating the length of the sliding window: .

[0053] In this embodiment of the invention, the overall consistency score is significantly improved after spatial unification and sliding window smoothing. In particular, the consistency score of the vertical beam measurement results (W90_mean) is the highest (0.89), verifying the effectiveness of the method in eliminating spatial mismatches and suppressing noise. This invention provides a reliable quality assessment index for vertical airflow observation.

[0054] Finally, it should be emphasized that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A vertical airflow assessment method based on multi-beam consistency scoring and sliding window processing, characterized in that, The method includes the following steps: S1. Acquire multi-beam radial velocity data of laser Doppler radar under different elevation and azimuth angles to form a multi-angle observation dataset; S2. Spatial interpolation method is used to uniformly map the data of the observation distance layer corresponding to different pitch angles to the target height, so as to eliminate the spatial inconsistency caused by the difference in observation geometry. S3. Based on the observation geometry of each beam, multiple sets of vertical airflow estimation results are obtained through multi-beam combination inversion; S4. A sliding window smoothing process is introduced for each vertical airflow sequence to obtain a smoothed sequence, thereby reducing the influence of random noise. S5. Based on the smoothed vertical airflow results, the correlation between different results is calculated using a correlation metric function, and a consistency matrix is ​​constructed. S6. Calculate the single-model consistency score and the overall consistency score based on the consistency matrix to achieve a quantitative assessment of the consistency of the vertical airflow results; The specific steps in S2 are as follows: S21. Set the target vertical detection range as a unified reference height; S22. Select two distance layers adjacent to the target height and use the two-point Lagrange interpolation method to calculate the radial velocity at the target distance layer; In step S4, the sliding window smoothing process for each vertical airflow sequence specifically includes: S41. Set the sliding window length N ; S42. Calculate the moving average of the vertical airflow sequence over time. The calculation method is as follows: ， in, t For time, This is the original vertical airflow sequence. The sequence is smoothed. In step S5, calculating the degree of correlation between different results using a correlation measurement function specifically includes: S51. Select the smoothed vertical airflow sequence and ; S52. Calculate its linear correlation coefficient. R ij As a measure of relevance, the formula is as follows: ， in, R ij To determine the degree of correlation between different results, i, j For different vertical airflow sequences, For the first i, j The result of smoothing the combined vertical airflow sequence; In step S5, the construction of the consistency matrix is ​​specifically as follows: ， in, The total number of types of multibeam combinations; In step S6, calculating the single-model consistency score based on the consistency matrix specifically involves calculating the single-model consistency score based on the consistency matrix using the following formula: ， in, S i For single-model consistency scoring.

2. The vertical airflow assessment method based on multi-beam consistency scoring and sliding window processing according to claim 1, characterized in that, The S3 steps are as follows: S31. Under non-vertical pitch angle conditions, establish a radial velocity model based on the observation geometry. S32. Construct a multi-beam combination for inversion, wherein the multi-beam combination includes: a combination of a first beam and a third beam, a combination of a second beam and a fourth beam, and a four-beam combination; S33. Under the condition of vertical pitch angle, the radial velocity is averaged to obtain the measurement result of vertical direction.

3. The vertical airflow assessment method based on multi-beam consistency scoring and sliding window processing according to claim 1, characterized in that, Step S6 also includes calculating the average of all individual model consistency scores as the overall consistency score under the current observation conditions using the following formula: ， in, S Score the overall consistency.

4. The vertical airflow assessment method based on multi-beam consistency scoring and sliding window processing according to any one of claims 1 to 3, characterized in that, The scanning process of the laser Doppler radar is as follows: under the first elevation angle condition, four azimuth angles are scanned sequentially, and then the same four azimuth angles are scanned under the second elevation angle condition to complete one cycle of multi-beam observation.

Images