Adaptive noise suppression based acoustic-optic radar wind data correction system

Abstract

The application discloses an adaptive noise suppression-based acoustic-optic radar wind data correction system and relates to the technical field of data correction.The dynamic adjustment module combines environmental parameters with acoustic-optic radar system parameters to dynamically adjust the signal-to-noise ratio threshold value, the echo signal quality analysis module inputs spectral feature data and the dynamically adjusted signal-to-noise ratio threshold value into an echo quality model to determine whether the current echo signal is valid, if not, a spatial smoothing interpolation method is adopted for interpolation, if yes, the main frequency of the echo signal is extracted and the Doppler shift is calculated, the radial wind speed is calculated according to the detection angle and the layout direction, the three-dimensional wind speed vector is inversely calculated and physical consistency correction is performed.The correction system combines real-time environmental parameters with system operation parameters by introducing a dynamic adjustment module to dynamically adjust the signal-to-noise ratio threshold value required for judgment, so that the judgment basis can be intelligently adapted to the measurement environment, and the accuracy and robustness of echo signal quality determination are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of data correction technology, and more specifically to an acoustic-optical radar wind measurement data correction system based on adaptive noise suppression. Background Technology

[0002] With the rapid development of atmospheric science, wind energy development and aerospace, the demand for high-precision, high spatiotemporal resolution wind field detection methods is increasing. As a non-contact, long-range advanced means of detecting wind speed and direction, acoustic-optical radar (SODAR, LIDAR, etc.) has the advantages of flexible deployment and high observation accuracy, and has been widely used in wind farm site selection, urban meteorological monitoring, boundary layer research and other scenarios.

[0003] However, in practical applications, due to the influence of various factors such as environmental background noise, electromagnetic interference, ground reflection and multipath effect, the raw data collected by acoustic-optical radar is often mixed with a lot of noise. This noise not only reduces the accuracy of wind measurement, but may even cause the omission or misjudgment of key wind field characteristics, affecting the reliability of wind measurement results. Traditional data filtering and denoising methods have problems such as poor adaptability, low processing efficiency and insufficient ability to identify outliers, making it difficult to cope with the challenge of multi-source noise superposition in complex environments.

[0004] The existing technology has the following drawbacks:

[0005] In traditional acoustic-optical radar wind measurement technology, the system often uses a preset fixed signal-to-noise ratio threshold to determine whether the echo signal is valid. However, due to the dynamic nature of the measurement environment, such as frequent changes in meteorological conditions, terrain structure, and background noise levels, the fixed threshold is difficult to accurately adapt to various actual scenarios. This leads to frequent misjudgments of valid signals as invalid signals under high noise or boundary conditions, resulting in missing or abnormal wind measurement data.

[0006] Based on this, the present invention proposes an acoustic-optical radar wind measurement data correction system based on adaptive noise suppression. By introducing a dynamic adjustment module, real-time environmental parameters are combined with system operating parameters to dynamically adjust the signal-to-noise ratio threshold required for judgment, thereby enabling the judgment criteria to intelligently adapt to the measurement environment, significantly improving the accuracy and robustness of echo signal quality judgment. Summary of the Invention

[0007] The purpose of this invention is to provide an acoustic-optical radar wind measurement data correction system based on adaptive noise suppression to address the shortcomings in the prior art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an acoustic-optical radar wind measurement data correction system based on adaptive noise suppression, comprising an acquisition module, a dynamic adjustment module, an echo signal quality analysis module, an interpolation module, and a wind speed analysis module;

[0009] Acquisition module: Acquires spectral characteristic data from the acoustic-optical radar;

[0010] Dynamic adjustment module: Combines environmental parameters with acoustic-optical radar system parameters to dynamically adjust the signal-to-noise ratio threshold;

[0011] Echo signal quality analysis module: Input the spectral characteristic data and the dynamically adjusted signal-to-noise ratio threshold into the echo quality model to determine whether the current echo signal is valid;

[0012] Interpolation module: If invalid, spatial smooth interpolation will be used for interpolation;

[0013] Wind speed analysis module: If effective, extract the main frequency of the echo signal and calculate the Doppler frequency shift. Based on the detection angle and deployment direction, calculate the radial wind speed and invert the three-dimensional wind speed vector and make physical consistency corrections.

[0014] In a preferred embodiment, the echo signal quality analysis module acquires spectral feature data and a dynamically adjusted signal-to-noise ratio threshold;

[0015] Calculate the instantaneous signal-to-noise ratio (SNR), and then use the calculated SNR... inst With dynamically adjusted signal-to-noise ratio threshold (SNR) th Compare, if SNR inst ≥SNR th If the dominant frequency signal in the current spectrum is significant, it is considered a valid echo signal; otherwise, it is considered an invalid signal.

[0016] In a preferred embodiment, the wind speed analysis module calculates the Doppler frequency shift: Δf = f peak -f0, where Δf is the Doppler frequency shift, f pcak f0 is the current main frequency of the measuring point, and f0 is the transmission frequency of the acoustic-optical radar;

[0017] Convert the frequency shift into radial wind speed v r The calculation formula is: Where: v r denoted as radial wind speed, and c is the propagation speed of sound or light waves in the detection medium;

[0018] The radial wind speed v of each beam r1 ,v r2 ,…,v rn With the corresponding beam direction vectors d1, d2, ..., d n Perform joint inversion to construct a three-dimensional wind speed vector.

[0019] In a preferred embodiment, the interpolation module determines the spatial location of the current invalid measuring point in the detection grid and selects several measuring points with valid echo signals from the spatial neighborhood of the current measuring point.

[0020] For all nearby valid measurement points participating in the interpolation, the corresponding interpolation weights are calculated based on their spatial relative distances, and the dominant frequency value f of the nearby measurement points is used as the basis for the interpolation. i and corresponding weight w i Calculate the estimated dominant frequency f of the current invalid measurement point. interp : k represents the number of all neighboring valid measurement points involved in the interpolation.

[0021] In a preferred embodiment, the signal-to-noise ratio (SNR) inst The calculation formula is: Among them, P peak It is the peak power density corresponding to the dominant frequency in the spectrum, P noise It is a representative value of the average power within a frequency range on both sides of the main frequency.

[0022] In a preferred embodiment, the dynamic adjustment module dynamically adjusts the signal-to-noise ratio threshold based on the average ambient noise power, the degree of ambient noise fluctuation, and the system state offset factor.

[0023] In a preferred embodiment, the dynamically adjusted signal-to-noise ratio threshold output by the dynamic adjustment module is: In the formula, SNR th This is the dynamically adjusted signal-to-noise ratio threshold. σ represents the average ambient noise power. N δ represents the degree of environmental noise fluctuation. sys θ is the system state offset factor, α, β, γ are weighting coefficients, and θ is the initial signal-to-noise ratio threshold.

[0024] In a preferred embodiment, the average ambient noise power is: Where N(t) represents the background spectral power in the t-th second or sampling period, and T is the length of the statistical time window;

[0025] Standard deviation of ambient noise power: σ represents the average ambient noise power. N This refers to the degree of environmental noise fluctuation.

[0026] In a preferred embodiment, the acquisition module receives echo signals from the acoustic-optical radar system and preprocesses the echo signals.

[0027] Fast Fourier Transform is performed on the digitized echo signal to extract the spectral features related to wind speed.

[0028] The spectral feature data corresponding to the current time, current measurement direction, and altitude are packaged to form a standardized spectral data structure.

[0029] In a preferred embodiment, the acquisition module preprocesses the echo signal, including filtering, amplification, and analog-to-digital conversion of the original echo signal to convert the analog echo signal into a digital signal.

[0030] The spectral features related to wind speed are extracted, including the position of the main peak, the peak amplitude, the spectral bandwidth, and the background noise energy.

[0031] After forming a standardized spectrum data structure, timestamps, measurement point location information, and operating status information are added. The data structure includes multiple frequency components and their corresponding power values, background noise baseline, and a set of level change trends.

[0032] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0033] This invention uses a dynamic adjustment module to dynamically adjust the signal-to-noise ratio (SNR) threshold by combining environmental parameters with the parameters of the acoustic-optical radar system. An echo signal quality analysis module inputs spectral characteristic data and the dynamically adjusted SNR threshold into the echo quality model to determine the validity of the current echo signal. If invalid, spatial smoothing interpolation is used; if valid, the dominant frequency of the echo signal is extracted and the Doppler frequency shift is calculated. Based on the detection angle and deployment direction, the radial wind speed is calculated, and the three-dimensional wind speed vector is inverted and physical consistency corrections are performed. This correction system, by introducing a dynamic adjustment module that combines real-time environmental parameters with system operating parameters, dynamically adjusts the required SNR threshold for judgment, enabling the judgment criteria to intelligently adapt to the measurement environment, significantly improving the accuracy and robustness of echo signal quality determination.

[0034] This invention introduces an interpolation module into the system architecture. When a measurement point is determined to be invalid, it is not simply discarded. Instead, it is compensated and filled using the effective spectral characteristics of adjacent height layers or azimuth angles based on a spatial smoothing interpolation algorithm. This method not only preserves the continuity of the wind field structure but also suppresses the instability caused by short-term data loss, effectively enhancing the system's wind measurement capability in low-signal-quality areas.

[0035] This invention also constructs a complete data closed-loop path. The spectral characteristics of the echo signal, the dynamic threshold, and the echo validity determination results are all input into the wind speed analysis module, ensuring that the main frequency extraction and Doppler frequency shift calculation are performed only under the premise that the signal quality is controllable. The spectral characteristics generated by the interpolation module also participate in the wind speed inversion in an equivalent manner, and combined with the physical consistency verification rules, a more reliable three-dimensional wind speed vector is finally output. This quality-based inversion logic solves the error accumulation problem caused by the "measure and calculate immediately without judgment" in traditional wind measurement systems. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0037] Figure 1 This is a system architecture diagram of the present invention.

[0038] Figure 2 This is a timing diagram of the correction coefficients of the present invention.

[0039] Figure 3 This is a flowchart of the method of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Example 1: Please refer to Figure 1 and Figure 2 As shown in the figure, the acoustic-optical radar wind measurement data correction system based on adaptive noise suppression described in this embodiment includes an acquisition module, a dynamic adjustment module, an echo signal quality analysis module, an interpolation module, and a wind speed analysis module;

[0042] Acquisition module: Acquires spectral characteristic data from the acoustic-optical radar and sends the spectral characteristic data to the echo signal quality analysis module;

[0043] Dynamic adjustment module: Combines environmental parameters with acoustic-optical radar system parameters to dynamically adjust the signal-to-noise ratio threshold, and sends the dynamically adjusted signal-to-noise ratio threshold to the echo signal quality analysis module;

[0044] Echo signal quality analysis module: Input the spectral characteristic data and the dynamically adjusted signal-to-noise ratio threshold into the echo quality model to determine whether the current echo signal is valid, and send the judgment result to the interpolation module and the wind speed analysis module;

[0045] Interpolation module: If invalid, spatial smoothing interpolation is used for interpolation, and the interpolated spectral characteristic data is sent to the wind speed analysis module;

[0046] Wind speed analysis module: If effective, extract the main frequency of the echo signal and calculate the Doppler frequency shift. Based on the detection angle and deployment direction, calculate the radial wind speed and invert the three-dimensional wind speed vector and make physical consistency corrections.

[0047] This application utilizes a dynamic adjustment module to dynamically adjust the signal-to-noise ratio (SNR) threshold by combining environmental parameters with the acoustic-optical radar system parameters. The echo signal quality analysis module inputs spectral characteristic data and the dynamically adjusted SNR threshold into the echo quality model to determine the validity of the current echo signal. If invalid, spatial smoothing interpolation is used; if valid, the dominant frequency of the echo signal is extracted and the Doppler frequency shift is calculated. Based on the detection angle and deployment direction, the radial wind speed is calculated, and the three-dimensional wind speed vector is inverted and physical consistency corrections are performed. This correction system, by introducing a dynamic adjustment module, combines real-time environmental parameters with system operating parameters to dynamically adjust the required SNR threshold for judgment. This allows the judgment criteria to intelligently adapt to the measurement environment, significantly improving the accuracy and robustness of echo signal quality determination.

[0048] Please see Figure 3 As shown, the specific workflow of the calibration system is as follows:

[0049] The calibration system collects spectral characteristic data from the acoustic-optical radar, dynamically adjusts the signal-to-noise ratio (SNR) threshold by combining environmental parameters with the SNR system parameters, and inputs the spectral characteristic data and the dynamically adjusted SNR threshold into the echo quality model to determine whether the current echo signal is valid. If invalid, spatial smoothing interpolation is used for interpolation; if valid, the dominant frequency of the echo signal is extracted and the Doppler frequency shift is calculated. Based on the detection angle and deployment direction, the radial wind speed is calculated, the three-dimensional wind speed vector is inverted, and physical consistency correction is performed.

[0050] Example 2: The acquisition module acquires the spectral characteristic data of the acoustic-optical radar, and the spectral characteristic data is sent to the echo signal quality analysis module.

[0051] The acquisition module, as the data input front-end of the entire system, is responsible for acquiring spectral characteristic data from the acoustic-optical radar equipment in real time. Its operation not only affects the accuracy of subsequent echo quality analysis but also directly impacts the overall response efficiency and stability of the system. The operation steps of this module can be broken down into the following aspects:

[0052] First, the acquisition module receives echo signals from the acoustic-optical radar system. These signals are energy information reflected back by aerosols, particles, or areas of temperature inhomogeneity in the atmosphere after the acoustic-optical radar emits an acoustic beam. To obtain useful spectral characteristics, the acquisition module needs to preprocess the echo signals, including filtering, amplifying, and performing analog-to-digital conversion on the original analog signals, thereby converting the analog echo signals into processable digital signals to prepare for frequency domain analysis.

[0053] Next, the acquisition module performs spectral analysis operations such as Fast Fourier Transform (FFT) on the digitized echo signal to extract key spectral features related to wind speed. These features mainly include multiple dimensions such as the position of the main peak, peak amplitude, spectral bandwidth, and background noise energy. They can reflect the frequency shift information, signal-to-noise ratio level, and energy distribution characteristics of the signal, and are important basic parameters for determining signal quality and calculating wind speed.

[0054] In an acoustic-optical radar wind measurement data correction system based on adaptive noise suppression, the acquisition module performs spectral analysis on the digitized echo signal to extract spectral features closely related to wind speed. The entire process includes the following detailed steps, supplemented with code examples illustrating its logical implementation:

[0055] The acquisition module first receives the echo signal after analog-to-digital conversion, typically represented as a time series. To improve the accuracy of signal analysis, the original signal undergoes DC offset removal processing, i.e., the average value of the signal is subtracted to suppress the interference of the DC component on the spectrum. The time-domain signal is then converted to the frequency domain using a Fast Fourier Transform (FFT) to obtain its spectral distribution. The FFT output is a complex vector; by taking its magnitude (amplitude spectrum), the energy distribution of the signal at different frequencies can be reflected, providing a foundation for subsequent dominant frequency identification and feature extraction.

[0056] The location of the main peak in the spectrum reflects the center of the frequency shift of the echo signal and is proportional to the radial velocity of the target object relative to the radar. The main peak can be obtained by finding the frequency corresponding to the point of maximum amplitude, which is the basis for calculating the Doppler frequency shift. The amplitude of the main peak represents the energy intensity of the signal and is an important indicator for evaluating signal effectiveness. The spectral bandwidth reflects the extent of the signal's spread in the frequency domain. A wider bandwidth usually means that the signal contains a wider velocity distribution, and may also indicate that the signal contains more noise or multipath components. After removing the influence of the main peak, the background noise level is estimated by statistically analyzing the average or median energy of the remaining frequency regions. This value is used for subsequent signal-to-noise ratio calculations and dynamic threshold adjustments, affecting signal quality assessment. The corresponding code example (Python) is as follows:

[0057]

[0058]

[0059]

[0060] Code explanation:

[0061] DC offset removal: Removes the average signal value to avoid the influence of DC on FFT analysis;

[0062] FFT analysis: The spectrum of the real signal is calculated using np.fft.rfft, and its amplitude spectrum is used for energy analysis;

[0063] Main frequency extraction: The frequency point corresponding to the maximum amplitude of the main frequency;

[0064] Bandwidth estimation: Identify the bandwidth of the spectrum that exceeds half the energy of the main peak;

[0065] Noise estimation: Avoiding the dominant frequency region, calculate the average energy of the remaining frequency bands as a background noise indicator.

[0066] After extracting the spectral features, the acquisition module packages the spectral feature data corresponding to the current time, current measurement direction, and altitude into a standardized spectral data structure according to a set format. It also adds necessary timestamps, measurement point location information, and system operating status information to facilitate accurate alignment and spatiotemporal analysis by subsequent modules. The data structure typically includes multiple frequency components and their corresponding power values, background noise baseline, and level variation trends, among other parameters.

[0067] Finally, the acquisition module sends these standardized spectral characteristic data to the echo signal quality analysis module via a high-speed bus or internal shared memory channel. To improve real-time performance and system robustness, this module is also equipped with a cache management mechanism and packet loss detection logic, ensuring that even in high-frequency measurement scenarios, the spectral data can be stably and completely transmitted to subsequent processing stages, avoiding data loss or delays from affecting the judgment results.

[0068] Through the orderly execution of the above steps, the acquisition module not only completed the task of converting the raw acoustic-optical radar signal into spectral characteristics, but also provided solid and controllable basic data support for subsequent signal quality assessment, wind speed inversion, and other tasks.

[0069] The dynamic adjustment module combines environmental parameters with the parameters of the acoustic-optical radar system to dynamically adjust the signal-to-noise ratio (SNR) threshold. The dynamically adjusted SNR threshold is then sent to the echo signal quality analysis module.

[0070] The core function of the dynamic adjustment module is to dynamically adjust the signal-to-noise ratio (SNR) threshold in real time, based on external environmental parameters and acoustic-optical radar system parameters, under different measurement environments and system operating conditions. This adapts to changing noise backgrounds and signal conditions, improving the accuracy and robustness of echo signal quality assessment. The calculation expression for the dynamically adjusted SNR threshold is:

[0071] The signal-to-noise ratio (SNR) threshold output by the dynamic adjustment module th Determined by the following formula:

[0072] In the formula, SNR th This is the dynamically adjusted signal-to-noise ratio threshold. σ represents the average background noise power, indicating the average background noise level over a period of time. N The standard deviation of environmental noise power represents the degree of environmental noise fluctuation and reflects noise instability. δ sys The system state offset factor is a correction coefficient obtained based on the system operating state (such as transmit power, receive sensitivity, and temperature control status). α, β, and γ are weighting coefficients, which are empirical parameters or obtained through machine learning training, and determine the influence weight of each factor on the threshold. θ is the initial signal-to-noise ratio threshold.

[0073] Ambient average noise power: Where N(t) represents the background spectral power in the t-th second or sampling period, and T is the length of the statistical time window (e.g., 5 seconds, 10 seconds). This represents the dynamically estimated average environmental noise power.

[0074] Standard deviation of ambient noise power: This parameter indicates the degree of noise power fluctuation over a short period of time. The more violent the noise fluctuation, the more unstable the environment, and the signal-to-noise ratio threshold should be increased accordingly to prevent misjudgment.

[0075] System state offset factor:

[0076]

[0077] Where: P tx P represents the current transmit power. tx,nom Indicates the nominal transmit power, G rx Indicates the current receiver sensitivity, G rx,nom Indicates the nominal receiver sensitivity, T radar This indicates the current operating temperature of the device, T. nom This represents the nominal operating temperature. ω1, ω2, and ω3 represent the weighting parameters of the system offset term, configured according to system characteristics. When the transmit power decreases, the receive sensitivity decreases, or the equipment temperature is abnormal, the overall system signal quality deteriorates, and the threshold for judgment should be increased accordingly.

[0078] The above dynamic signal-to-noise ratio threshold formula achieves comprehensive perception and response to changes in real-time environmental noise characteristics and system operating status. It possesses the following functional characteristics:

[0079] Environmental perception capability: By calculating the average noise power and noise fluctuation level in real time, it effectively reflects the impact of external noise background on signal judgment.

[0080] System self-testing capability: By utilizing fluctuations in system operating parameters, the judgment strategy can be adjusted in a timely manner to prevent misjudgments caused by equipment aging or power drift.

[0081] Dynamic robustness adjustment: Under harsh environmental conditions or poor equipment condition, the signal-to-noise ratio threshold is increased to reduce erroneous reception; under ideal conditions, the threshold is appropriately lowered to ensure that the signal is not lost.

[0082] Adaptable to various measurement scenarios: Especially suitable for areas with significant day-night variations, complex terrain, and variable weather, improving the versatility and reliability of wind measurement radar systems.

[0083] Through the aforementioned mechanism, the dynamic adjustment module possesses the following functional characteristics: it can perceive environmental noise levels and fluctuations in real time, dynamically sense operational deviations of the radar system, and comprehensively generate the most suitable signal-to-noise ratio threshold, avoiding misjudgments or missed judgments caused by fixed thresholds. The final output dynamic threshold will be directly used by the echo signal quality analysis module to guide it in determining whether the currently measured spectral data has effective wind speed extraction value. This method effectively improves the accuracy and stability of wind measurement in complex environments.

[0084] The echo signal quality analysis module inputs the spectral characteristic data and the dynamically adjusted signal-to-noise ratio threshold into the echo quality model to determine whether the current echo signal is valid. The judgment result is then sent to the interpolation module and the wind speed analysis module.

[0085] The core task of the echo signal quality analysis module is to receive spectral characteristic data and dynamically adjusted signal-to-noise ratio thresholds, input this information into the echo quality model, and comprehensively determine whether the current echo signal has sufficient reliability for wind speed calculation. If the signal is determined to be invalid, the interpolation module will perform replacement processing; if it is determined to be valid, it will directly enter the wind speed analysis process. The specific judgment process can be broken down into the following detailed steps:

[0086] The echo signal quality analysis module uses two main components: spectral characteristic data measured by acoustic-optical radar and a real-time signal-to-noise ratio threshold dynamically adjusted by the module. Spectral characteristic data includes, but is not limited to, power spectrum distribution, peak frequency, and spectral width. During preprocessing, the module performs normalization, smoothing, and noise reduction on the spectral curve to eliminate unstable components caused by transient interference or system jitter, thus improving the stability of subsequent analysis.

[0087] In the preprocessed spectrum, the system will identify the power intensity P corresponding to the peak frequency. peak And calculate the average background noise power P in the non-dominant frequency region. noiseThe signal-to-noise ratio (SNR) calculation formula is as follows:

[0088] Among them, P peak It is the peak power density corresponding to the dominant frequency in the spectrum, usually selected as the highest single frequency point, while P noise It is a representative value of the average power within a frequency range on both sides of the main frequency, used to characterize the background noise level.

[0089] The echo quality model uses the instantaneous signal-to-noise ratio (SNR) calculated above. inst The threshold SNR output by the dynamic adjustment module th Compare them. If SNR inst ≥SNR th If the signal is sufficiently significant, the dominant frequency signal in the current spectrum is considered a valid echo signal. Otherwise, it is deemed an invalid signal and will be processed by the interpolation module. This criterion ensures that the system dynamically adapts to different background noise and equipment conditions, improving overall robustness.

[0090] In certain critical measurement phases (such as scenarios with drastic boundary layer changes), the module also introduces two auxiliary judgment indicators to enhance the accuracy of discrimination: 1. Spectral peak sharpness index: used to analyze the steepness of the main peak of the spectrum and its adjacent regions, identifying whether it is a typical Doppler peak shape. 2. Spectrum Limitation indicators: These are used to determine whether the current spectral peak is due to excessively wide or unfocused interference, avoiding misidentifying broadband noise as the main frequency peak. The system will only ultimately determine it as valid if the main judgment condition is met (SNR is greater than the threshold) and the auxiliary indicators are within a reasonable range.

[0091] Once the validity of the echo signal is confirmed or denied, the module encodes the judgment result and sends it to two subsequent processing modules respectively:

[0092] If it is a valid signal, it will be sent to the wind speed analysis module, along with the main frequency of the spectrum and its related characteristics;

[0093] If the signal is invalid, it is sent to the interpolation module along with the geographic information of the current measurement point, the time index, and the spectral reference information of the neighboring points, so that the interpolation module can perform spatial smoothing substitution.

[0094] If the interpolation is invalid, the interpolation module will use spatial smoothing interpolation to perform interpolation, and the interpolated spectral feature data will be sent to the wind speed analysis module.

[0095] When the echo signal quality analysis module determines that the spectral signal at the current measurement point is invalid, the interpolation module will use spatial smoothing interpolation to estimate and complete the spectral characteristic data of that measurement point, ensuring that the wind speed analysis module can still obtain reasonable spectral data input at invalid measurement points. This spatial interpolation process not only considers the spatial distribution of neighboring measurement points but also combines the physical continuity of the spectral dominant frequency to ensure the consistency of the interpolation results in a physical sense. Specifically, it includes the following steps:

[0096] The interpolation module first determines the spatial location of the current invalid measurement point within the detection grid, typically indexed by height layer, azimuth, and range gate. Then, it selects several measurement points with valid echo signals from the spatial neighborhood of the current measurement point and extracts the corresponding spectral feature data, mainly including parameters such as dominant frequency, power spectrum shape, and spectral width.

[0097] The selection of the neighborhood should prioritize a three-dimensional spherical neighborhood. If there is significant data loss in a certain direction, it can be reduced to two-dimensional planar or linear interpolation. Common neighborhood scales include ±12 grid cells in the vertical direction.

[0098] For all nearby valid measurement points participating in the interpolation, the module calculates the corresponding interpolation weights based on their spatial relative distances. Typical weight calculation methods include Gaussian weighting functions or inverse distance weighting (IDW). Taking inverse distance weighting as an example: Among them, w i d represents the weight of the i-th neighboring measurement point. i It is the spatial Euclidean distance from the current invalid measurement point to the current invalid measurement point, where p is the weighting exponent, typically ranging from 1.5 to 2.5. All weights, after normalization, satisfy: This step ensures that closer effective measurement points have a greater impact on the interpolation results, which is consistent with the spatial continuity of the wind field.

[0099] Based on the dominant frequency value f of nearby measuring points i and corresponding weight w i The interpolation module calculates the estimated main frequency for the current invalid measurement point: Furthermore, to maintain the reasonableness of the spectral shape, the module also performs similar interpolation processing on the spectral width Δf and the power spectrum shape. For example, the spectral width interpolation value is:

[0100] Finally, based on the dominant frequency, spectral width, and amplitude obtained by interpolation, a simulated spectrum conforming to a typical Gaussian spectrum or empirical spectrum model is constructed as the reconstructed spectral characteristics of the current measurement point.

[0101] After the interpolated spectrum is initially generated, the module also introduces a spectrum shape correction mechanism to ensure that it maintains physical continuity with surrounding measurement points in terms of dominant frequency position, energy distribution, and spectral steepness. If there are unreasonable abrupt changes in the interpolated spectrum compared with the spectral shape of adjacent measurement points, such as frequency jumps, abnormal sharpness, or broadening, fine-tuning is performed by adjusting the spectral peak smoothing factor and power normalization constant to prevent it from being misjudged as an interference signal during the wind speed analysis stage.

[0102] After interpolation is completed, the module sends the reconstructed spectral feature data to the wind speed analysis module. The interpolation result is structurally consistent with the original effective spectrum, including information such as the dominant frequency position, spectral width, and power density sequence, and is marked as an interpolation result. The wind speed analysis module will extract the dominant frequency according to the conventional process for Doppler frequency shift calculation and wind vector inversion, and can also mark the interpolated data for subsequent quality control or statistical processing.

[0103] The interpolation module achieves high-fidelity compensation for invalid echo measurement points, which not only maintains the integrity of the three-dimensional wind field in terms of data continuity, but also improves the reliability of the overall wind speed inversion results in boundary areas and areas with missing data, ensuring that the system has anti-interference and anti-blocking capabilities.

[0104] If effective, the wind speed analysis module extracts the dominant frequency of the echo signal and calculates the Doppler frequency shift. Based on the detection angle and deployment direction, it calculates the radial wind speed, inverts the three-dimensional wind speed vector, and performs physical consistency correction.

[0105] When the echo signal quality analysis module determines that the echo signal from the measuring point is valid, the wind speed analysis module uses the spectral characteristics of that measuring point as input and sequentially performs main frequency extraction, Doppler frequency shift calculation, radial wind speed calculation, three-dimensional wind speed vector inversion, and physical consistency correction to ensure that the output wind field information is complete and accurate in terms of spatial structure and physical laws. The specific analysis process includes the following detailed steps:

[0106] The wind speed analysis module first extracts the dominant frequency from the input spectral feature data. This process typically involves finding the spectral peak location based on the power spectrum curve, i.e., identifying the frequency point with the highest power. If multiple local peaks appear in the power spectrum, the system will combine the peak height, peak width, and the ratio of peak height to background noise to ensure that the extracted frequency component corresponds to the main reflected wave, rather than secondary interference or echo ghosting.

[0107] After extracting the main frequency, the system calculates the Doppler frequency shift according to the following formula: Δf = f peak -f0, where Δf is the Doppler frequency shift, f pcak f0 is the dominant frequency at the current measuring point, and f0 is the transmission frequency of the acoustic-optical radar. This frequency shift value directly reflects the relative velocity projection of the target air mass with respect to the radar system, and is the core physical quantity for subsequent wind speed calculation.

[0108] Based on the classical Doppler effect principle, the system converts the frequency shift into radial wind speed v. r The calculation formula is as follows: Where: v r Where is the radial wind speed, and c is the propagation speed of sound or light waves in the detection medium (e.g., sound waves in air are approximately 343 m / s, and light waves are approximately 3 × 10⁻⁶ m / s). 8 (m / s), other parameters are the same as above. This step outputs the projected value of the wind speed in the direction of the radar beam, that is, the component velocity of the target air mass on the detection path.

[0109] Acoustic-optical radar systems typically employ a multi-angle, multi-beam deployment, scanning the same volume area from different azimuths and elevation angles to obtain radial wind speeds in multiple directions. The wind speed analysis module then calculates the radial wind speed v from each beam. r1 ,v r2 ,…,v rn With the corresponding beam direction vectors d1, d2, ..., d n Perform joint inversion to construct a three-dimensional wind speed vector. Specifically, the least squares method is used to solve the following system of equations: Where: u, v, w are the east-west, north-south, and vertical wind speed components, respectively, and d i The unit direction vector representing the i-th beam, and n representing the number of beams, are typically no less than 3. This step outputs complete three-dimensional vector information of the wind speed, used to describe the actual movement of the air mass in space.

[0110] Below is a code example (written in Python) that implements the following five steps: "If effective, the wind speed analysis module extracts the dominant frequency of the echo signal and calculates the Doppler frequency shift; based on the detection angle and deployment direction, it calculates the radial wind speed, inverts the three-dimensional wind speed vector, and performs physical consistency correction." Detailed instructions are also provided.

[0111]

[0112] Extracting the main frequency

[0113] peaks,_=find_peaks(spectrum_power)

[0114] iflen(peaks) == 0:

[0115] return None

[0116] main_peak=peaks[np.argmax(spectrum_power[peaks])]

[0117] return spectrum_freq[main_peak]

[0118] def compute_doppler_shift(f_peak,f0):

[0119] "Calculating Doppler frequency shift"

[0120] return f_peak-f0

[0121] def compute_radial_velocity(delta_f,f0,c):

[0122] Calculating radial wind speed

[0123] return(c*delta_f) / (2*f0)

[0124] definvert_3d_wind(radial_velocities,beam_dirs):

[0125] "Inverting the 3D Wind Speed ​​Vector"

[0126] D = np.array(beam_dirs) # Beam direction matrix

[0127] v_r=np.array(radial_velocities).reshape(-1,1)

[0128] V,_,_,_=lstsq(D,v_r)

[0129] return V.flatten()#[u,v,w]

[0130] def smooth_velocity_field(current_vector,neighbor_vectors,weight=0.5):

[0131] "Smoothing Correction: Weighted average with surrounding wind speed vector"

[0132] avg_neighbor=np.mean(neighbor_vectors,axis=0)

[0133] return weight*current_vector+(1-weight)*avg_neighbor

[0134] #Example data (provided by spectrum analysis and radar systems in practice)

[0135]

[0136] print("Original wind speed vector:", wind_vector)

[0137] print("Corrected wind speed vector:", corrected_wind_vector)

[0138] else:

[0139] print("Failed to identify a valid CPU frequency")

[0140] Code explanation:

[0141] Step 1: Extract the main frequency

[0142] Use scipy.signal.find_peaks to find all peaks in the power spectrum; select the frequency corresponding to the peak with the largest power value as the main frequency f_peak.

[0143] Step 2: Calculate the Doppler frequency shift

[0144] The Doppler frequency shift Δf is the difference between the main frequency and the transmission frequency: Δf = f_peak - f0; if the target is approaching the radar, the frequency shift is positive; if it is moving away, it is negative.

[0145] Step 3: Calculate radial wind speed

[0146] Using the standard Doppler formula v_r=(c*Δf) / (2*f); this is the velocity of the target along the radar beam direction.

[0147] Step 4: Invert the 3D wind speed vector

[0148] A system of linear equations is established for the radial wind speed measured for each beam; the three-dimensional wind speed [u,v,w], i.e., each component of the three-dimensional wind speed, is solved using the least squares method.

[0149] Step 5: Physical Consistency Correction

[0150] The example uses the nearest-neighbor average smoothing method to correct wind speed and avoid abrupt changes; it can be extended to more complex physical constraint models such as boundary constraints and continuity corrections.

[0151] The initially derived 3D wind speed vector may exhibit physical inconsistencies due to local measurement errors or an incompatible inversion algorithm, such as sudden wind direction reversals, abrupt changes in wind speed, or abnormal vertical velocity. Therefore, the wind speed analysis module also includes the following physical consistency correction steps:

[0152] Smoothing correction: Spatially smooth the wind speed at the current measuring point with that at neighboring measuring points to reduce isolated outliers;

[0153] Boundary constraints: Adjust the boundary shape of the wind speed vector by combining meteorological models or boundary conditions (such as near-surface vertical wind speed approaching zero);

[0154] Continuity constraints: Ensure that wind field changes satisfy mass conservation and avoid passive or sinkless airflow abrupt changes at the radar shear layer;

[0155] Historical trend constraint: In continuous observation mode, the system will also use a time series trend model to evaluate whether the current wind speed is consistent with the wind field trend at the previous moment. If the deviation is too large, a correction will be triggered.

[0156] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0157] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An acoustic-optical radar wind measurement data correction system based on adaptive noise suppression, characterized in that: It includes an acquisition module, a dynamic adjustment module, an echo signal quality analysis module, an interpolation module, and a wind speed analysis module; Acquisition module: Acquires spectral characteristic data from the acoustic-optical radar; Dynamic adjustment module: Combines environmental parameters with acoustic-optical radar system parameters to dynamically adjust the signal-to-noise ratio threshold; Echo signal quality analysis module: Input the spectral characteristic data and the dynamically adjusted signal-to-noise ratio threshold into the echo quality model to determine whether the current echo signal is valid; Interpolation module: If invalid, spatial smooth interpolation will be used for interpolation; Wind speed analysis module: If effective, extract the main frequency of the echo signal and calculate the Doppler frequency shift. Based on the detection angle and deployment direction, calculate the radial wind speed and invert the three-dimensional wind speed vector and make physical consistency corrections. The echo signal quality analysis module acquires spectral feature data and dynamically adjusted signal-to-noise ratio thresholds. Calculate the instantaneous signal-to-noise ratio, and then use the calculated instantaneous signal-to-noise ratio... With dynamically adjusted signal-to-noise ratio threshold If a comparison is made, If the dominant frequency signal in the current spectrum is significant, it is considered a valid echo signal; otherwise, it is considered an invalid signal. The dynamic adjustment module dynamically adjusts the signal-to-noise ratio threshold based on the average environmental noise power, the degree of environmental noise fluctuation, and the system state offset factor.

2. The acoustic-optical radar wind measurement data correction system based on adaptive noise suppression according to claim 1, characterized in that: The wind speed analysis module calculates the Doppler frequency shift: ,in, For Doppler frequency shift, The current main frequency at the measurement point. This refers to the transmission frequency of the acoustic-optical radar. Convert frequency shift to radial wind speed The calculation formula is: ,in: Radial wind speed, The speed at which sound waves or light waves propagate in the detection medium; Radial wind speed of each beam With the corresponding beam direction vector Perform joint inversion to construct a three-dimensional wind speed vector. .

3. The acoustic-optical radar wind measurement data correction system based on adaptive noise suppression according to claim 2, characterized in that: The interpolation module determines the spatial location of the current invalid measuring point in the detection grid and selects several measuring points with valid echo signals from the spatial neighborhood of the current measuring point. For all nearby valid measurement points participating in the interpolation, the corresponding interpolation weights are calculated based on their spatial relative distances, and the dominant frequency values ​​of the nearby measurement points are used as the basis for the calculation. and corresponding weights Calculate the estimated main frequency of the current invalid measurement points. : , where k represents the number of all neighboring valid measurement points involved in the interpolation.

4. The acoustic-optical radar wind measurement data correction system based on adaptive noise suppression according to claim 3, characterized in that: Signal-to-noise ratio The calculation formula is: ,in, It is the peak power density corresponding to the dominant frequency in the spectrum. It is a representative value of the average power within a frequency range on both sides of the main frequency.

5. The acoustic-optical radar wind measurement data correction system based on adaptive noise suppression according to claim 1, characterized in that: The dynamically adjusted signal-to-noise ratio threshold output by the dynamic adjustment module is as follows: In the formula, This is the dynamically adjusted signal-to-noise ratio threshold. The average ambient noise power. For the degree of environmental noise fluctuation, This is the system state offset factor. These are the weighting coefficients. This is the initial signal-to-noise ratio threshold.

6. The acoustic-optical radar wind measurement data correction system based on adaptive noise suppression according to claim 5, characterized in that: Ambient average noise power: ,in, Indicates the first Background spectral power within a second or sampling period, The length of the time window used for statistics; Standard deviation of ambient noise power: , The average ambient noise power. This refers to the degree of environmental noise fluctuation.

7. The acoustic-optical radar wind measurement data correction system based on adaptive noise suppression according to claim 6, characterized in that: The acquisition module receives echo signals from the acoustic-optical radar system and preprocesses the echo signals. Fast Fourier Transform is performed on the digitized echo signal to extract the spectral features related to wind speed. The spectral feature data corresponding to the current time, current measurement direction, and altitude are packaged to form a standardized spectral data structure.

8. The acoustic-optical radar wind measurement data correction system based on adaptive noise suppression according to claim 7, characterized in that: The acquisition module preprocesses the echo signal, including filtering, amplification, and analog-to-digital conversion of the original echo signal, converting the analog echo signal into a digital signal. The spectral features related to wind speed are extracted, including the position of the main peak, the peak amplitude, the spectral bandwidth, and the background noise energy. After forming a standardized spectrum data structure, timestamps, measurement point location information, and operating status information are added. The data structure includes multiple frequency components and their corresponding power values, background noise baseline, and a set of level change trends.

Images