Improved Doppler signal processing method and system

Abstract

The invention discloses an improved Doppler signal processing method and system, and belongs to the technical field of Doppler signal processing, and the method comprises the following steps: S1, signal collection and preprocessing: collecting an original Doppler signal, and carrying out the moving average filtering; s2, time-frequency analysis: performing time-domain analysis on the filtered Doppler signal by adopting a Hamming window function and improved wavelet threshold de-noising and combining short-time Fourier transform to obtain a motion spectrogram; s3, feature extraction: calculating a real envelope by adopting Hilbert transform, and traversing data in the motion spectrogram in combination with a peak searching function based on envelope analysis so as to extract a peak point; and S4, judging and outputting: screening peak points meeting peak height threshold conditions, and outputting the peak points and position coordinates thereof. According to the invention, a better signal extraction effect and visual speed measurement are realized, signal separation and extraction are effectively realized, and the precision and reliability of signal processing are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of Doppler signal processing, and particularly to an improved Doppler signal processing method and system. Background Art

[0002] Traditional Doppler signal processing methods mainly include spectral analysis, frequency tracking, counting techniques, photon correlation techniques, and Fourier transform methods. In complex measurement environments, such as multipath propagation, noise interference, etc., a single signal processing method often fails to effectively cope with various signal characteristics and noise types, resulting in the characteristics of Doppler signals may vary depending on the measurement object, environmental conditions, and sensor characteristics. Therefore, a single method is difficult to adapt to all situations.

[0003] Among them, the Fourier transform, as a basic method in the field of signal processing, has been widely used in Doppler signal processing. From the initial Discrete Fourier Transform (DFT) to the later Fast Fourier Transform (FFT), since the Fourier transform is a global transform and cannot provide local information of time and frequency simultaneously, it is difficult to process non-stationary signals. When processing Doppler signals, when the target motion state changes, the Fourier transform cannot accurately reflect the real-time change of the signal frequency. Summary of the Invention

[0004] The purpose of the present invention is to provide an improved Doppler signal processing method and system, aiming to effectively extract key features in the signal and suppress noise interference.

[0005] To achieve the above purpose, the present invention provides an improved Doppler signal processing method, including the following steps: Step S1, signal acquisition and preprocessing: Acquire the original Doppler signal and perform moving average filtering; Step S2, time-frequency analysis: Use the Hamming window function, improved wavelet threshold denoising and combine with the short-time Fourier transform to perform time-domain analysis on the filtered Doppler signal to obtain the motion spectrogram; Step S3, feature extraction: Use the Hilbert transform to calculate the real envelope and combine with the peak-seeking function based on envelope analysis to traverse the data in the motion spectrogram to extract the peak points; Step S4, judgment and output: Screen the peak points that meet the peak height threshold condition and output the peak points and their position coordinates.

[0006] Preferably, step S1 specifically includes: Step S11, set the sampling rate to 200 MHz, acquire the original Doppler signal and input it; Step S12: Set the window length to 9 and perform moving average filtering on the input original Doppler signal. This process suppresses random noise while preserving the trend characteristics of the signal. The output of the moving average filtering is expressed as: (1); where represents the current moment in the time series, and its value range is , is the signal length, is the length of the moving average filter. The moving average filter calculates the average value of the current point at time and the previous points as the filtered signal value. is the average coefficient, is the input signal, is the index variable in the summation operation and . By changing the value of , the values at different times in the input signal values are sequentially selected for accumulation.

[0007] Preferably, step S2 specifically includes: Step S21: Process the filtered Doppler signal using the Hamming window function to suppress the edge effect of the Doppler signal. The definition of the Hamming window function is as follows: (2); where is the serial number of the current sampling point, and its value range is , is the length of the Hamming window function; Step S22: Perform dynamic denoising on the Doppler signal obtained after being processed in step S21 using improved wavelet threshold denoising. Through the multi-scale decomposition ability of the wavelet, different frequency components in the Doppler signal obtained after being processed in step S21 are separated. The formula of the improved wavelet threshold denoising function is as follows: (3); where is the output signal, is the input signal value, is the absolute value of the input signal, is the defined sign function, which determines the sign of the output signal according to the positive and negative of the input signal . is the threshold, and α and β are adjustment parameters. α controls the contraction intensity and affects the compression amplitude of the output signal, and β controls the smoothness and affects the adjustment of the signal in the transition region; Step S23: Segment and window the Doppler signal obtained after being processed in Step S22, and apply the short-time Fourier transform to analyze the time-frequency characteristics of the Doppler signal, capture the frequency characteristics of the Doppler signal changing with time, and obtain the original spectrum. The formula for windowing using the window function is: (4); Where, is the signal sequence, represents the input signal at the current moment value, is the starting position identifier of the window function on the signal sequence , indicates segmenting and windowing the input signal , is the windowing expression for the signal sequence, is the length of each segment of the signal sequence; The short-time Fourier transform is used to capture non-stationary signals and mutation signals, and the expression is: (5); Where, represents the result of the short-time Fourier transform, is the current moment, is the index parameter for distinguishing the odd and even segment calculations of the signal sequence , is the starting position of the window function on the signal sequence , is the frequency index used to determine the frequency position, is the signal length, is about odd sequence point transformation, is about even sequence point transformation; Step S24: Process the original spectrum using the spectrum optimization method based on bottom-spectrum correction; Step S241: Collect the static spectrum without moving object information as the bottom spectrum ; Step S242: Perform a difference operation on the bottom spectrum and the dynamic spectrum containing motion information to obtain the real motion spectrum diagram The formula is as follows: (6).

[0008] Preferably, Step S3 specifically includes: Step S31: After short-time Fourier transform, the first frequency point in the true motion spectrum diagram corresponds to a value of 0, denoted as , to remove the DC component of the Doppler signal; Step S32: Use Hilbert transform to calculate the Doppler signal with the DC component removed. The coefficient of the Hilbert transform is , being the frequency variable, used to determine the value of . When , the negative frequencies in the frequency-domain signal shift forward by phase. When , the positive frequencies in the frequency-domain signal shift backward by phase to obtain a complex signal, and then take the modulus of the obtained complex signal to obtain the true envelope of the frequency-domain signal. The formula for the value of is as follows: In the formula, is the imaginary unit; Step S33: Set the peak height threshold. Traverse the entire true motion spectrum data through the peak search function, and judge whether each data meets the peak height threshold condition. The data that meets the peak height threshold condition is the peak point. The method for setting the peak height threshold is: take the absolute value of the calculated true envelope. Among the true envelope data after taking the absolute value, the envelope value greater than 0.5 times of all other data is set as the peak height threshold. The formula for screening peak points using the peak height threshold is: (8); where, represents the peak height threshold; represents the true envelope signal, is the abbreviation of the absolute value function , indicating taking the absolute value of the true envelope signal; The output value of the peak search function is , and the formula is as follows: (9); is the index of the true motion spectrum data, which points to each data point in turn during the process of traversing the entire true motion spectrum data. is the value of the spectrum data at point, and are the adjacent data points of . When the output is 1, point is the peak point.​ When the output is 0, the point is not a peak point.

[0009] Preferably, after step S4, it further includes conducting a disk rotation speed measurement experiment by setting up a single-point laser Doppler velocimetry system to test the performance of the improved Doppler signal processing method.

[0010] A system for implementing an improved Doppler signal processing method includes a preprocessing module, a time-frequency analysis module, and a feature extraction module connected in sequence. Among them, the preprocessing module includes an original signal input module for collecting Doppler signals at a set sampling frequency and a moving average filtering module for filtering the collected Doppler signals with a set window length; the time-frequency analysis module includes a Hamming window function processing module for suppressing signal edge effects, an improved wavelet threshold denoising module for dynamically denoising Doppler signals, and a short-time Fourier transform module for capturing the frequency characteristics of Doppler signals; the feature extraction module includes a Doppler peak extraction module based on envelope analysis and then using Hilbert transform and a spectral feature output module for outputting the peak coordinates corresponding to the effective frequency points. The original signal input module, the moving average filtering module, the Hamming window function processing module, the improved wavelet threshold denoising module, the short-time Fourier transform module, the Doppler peak extraction module, and the spectral feature output module are connected in sequence.

[0011] Therefore, the improved Doppler signal processing method and system adopted by the present invention have the following beneficial effects: 1) In view of the strict requirements for high precision in the Doppler speed measurement process, an innovative improved wavelet threshold denoising method is proposed. By this method, the local characteristics or noise level of the Doppler signal are dynamically adjusted, realizing the effective separation and extraction of the signal, and significantly improving the accuracy and reliability of signal processing.

[0012] 2) With the short-time Fourier transform as the core, combined with moving average filtering, Hamming window function, improved wavelet threshold denoising, and peak extraction techniques, a complete set of processing methods for laser Doppler velocimetry signals is formed. Compared with a single processing method, this method realizes the accurate extraction and visual speed measurement of Doppler peak signals.

[0013] The technical solutions of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0014] Figure 1 It is a flowchart of the signal processing structure of an embodiment of the present invention; Figure 2 It is a comparison diagram of the improved threshold function curves of an embodiment of the present invention; Figure 3Flow chart of the peak extraction algorithm according to the embodiments of the present invention; Figure 4 User interface diagram according to the embodiments of the present invention, where (a) is the original diagram and (b) is the fast Fourier transform and envelope diagram; Figure 5 Spectrum comparison diagram obtained by the single-point Laser Doppler Velocimeter (LDV) according to the embodiments of the present invention, where (a) is the traditional fast Fourier transform spectrum diagram and (b) is the spectrum diagram of the improved Doppler signal processing method. Detailed implementation manners

[0015] In order to make the purposes, technical solutions and advantages of the embodiments disclosed in the present invention clearer and more understandable, the following further elaborates on the embodiments of the present invention in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not used to limit the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of this application. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout.

[0016] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server that includes a series of steps or units does not necessarily have to be limited to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0017] The following elaborates in detail on the implementation manners of the present invention in conjunction with the accompanying drawings.

[0018] As Figure 1 shown, the present invention provides an improved Doppler signal processing method, including the following steps: Step S1, signal acquisition and preprocessing: acquiring the original Doppler signal and performing moving average filtering; Step S2, time-frequency analysis: performing time-domain analysis on the filtered Doppler signal by using the Hamming window function, improved wavelet threshold denoising and combining with the short-time Fourier transform to obtain the motion spectrum diagram; Step S3, feature extraction: calculating the true envelope by using the Hilbert transform and traversing the data in the motion spectrum diagram in combination with the peak-seeking function based on envelope analysis to extract the peak points; Step S4, judgment and output: screening the peak points that meet the peak height threshold condition and outputting the peak points and their position coordinates.

[0019] Step S1 specifically includes: Step S11: Set the sampling rate to 200 MHz, collect the original Doppler signal and input it; Step S12: Set the window length to 9, perform moving average filtering on the input original Doppler signal. This process suppresses random noise while preserving the trend characteristics of the signal. Set the length of a moving average filter to , and the output is , Expressed as: (1); Among them, represents the current moment in the time series, and the value range is , is the signal length, is the length of the moving average filter. The moving average filter calculates the average value of the current point at time and the previous points as the filtered signal value. is the average coefficient, is the input signal, is the index variable in the summation operation and . By changing the value, the values at different times in the input signal values are sequentially selected for accumulation.

[0020] Step S2 specifically includes: Step S21: Process the filtered Doppler signal using the Hamming window function to suppress the edge effect of the Doppler signal. The definition of the Hamming window function is as follows: (2); Among them, is the serial number of the current sampling point, and the value range is , is the length of the Hamming window function; Step S22: Perform dynamic denoising on the Doppler signal obtained after being processed in Step S21 using improved wavelet threshold denoising. Through the multi-scale decomposition ability of the wavelet, different frequency components in the Doppler signal obtained after being processed in Step S21 are separated. The formula of the improved wavelet threshold denoising function is as follows: (3); Among them, is the output signal, is the input signal value, is the absolute value of the input signal, is the defined sign function, which determines the sign of the output signal according to the positive or negative of the input signal , is the threshold, α and β are adjustment parameters, α controls the contraction intensity and affects the compression amplitude of the output signal, and β controls the smoothness and affects the adjustment of the signal in the transition region.

[0021] Step S23: Segment and window the Doppler signal obtained after being processed in Step S22, and apply the short-time Fourier transform to analyze the time-frequency characteristics of the Doppler signal, capture the frequency characteristics of the Doppler signal changing with time, and obtain the original spectrum. The formula for windowing using the window function is: (4); where is the signal sequence, represents the input signal at the current moment value, is the starting position identifier of the window function on the signal sequence , indicates segmenting and windowing the input signal , is the windowing expression of the signal sequence, is the length of each segment of the signal sequence; The short-time Fourier transform is used to capture non-stationary signals and mutation signals, and the expression is: (5); where represents the result of the short-time Fourier transform, is the current moment, is the index parameter for distinguishing the odd and even segment calculations of the signal sequence , is the starting position of the window function on the signal sequence , is the frequency index used to determine the frequency position, is the signal length, is the odd sequence point transform with respect to , is the even sequence point transform with respect to ; Step S24: Process the original spectrum using the spectrum optimization method based on bottom spectrum correction; Step S241: Collect the static spectrum without moving object information as the bottom spectrum ; Step S242: Perform a difference operation on the bottom spectrum and the dynamic spectrum containing motion information to obtain the true motion spectrum diagram The formula is as follows: (6); Step S3 specifically includes: Step S31: After short-time Fourier transform, the value corresponding to the first frequency point in the true motion spectrum diagram is 0, denoted as , to remove the DC component of the Doppler signal; Step S32: Use the Hilbert transform to calculate the Doppler signal with the DC component removed. The coefficient of the Hilbert transform is , where is the frequency variable used to determine the value of . When , the negative frequency in the frequency-domain signal moves forward by phase. When , the positive frequency in the frequency-domain signal moves backward by phase to obtain a complex signal, and then take the modulus of the obtained complex signal to obtain the true envelope of the frequency-domain signal. The formula for the value of is as follows: (7); In the formula, is the imaginary unit; As shown in Figure 3 , in Step S33, set 0.5 times the maximum envelope amplitude as the peak minimum height threshold to avoid noise interference. Then, traverse the entire true motion spectrum data through the peak search function to determine whether each data meets the peak height threshold condition. The data that meets the peak height threshold condition is the peak point. The method for setting the peak height threshold is: take the absolute value of the calculated true envelope. In the true envelope data after taking the absolute value, take 0.5 times the maximum envelope value. The formula for screening peak points using the peak height threshold is: (8); Among them, represents the peak height threshold; represents the true envelope signal, is the abbreviation of the absolute value function , indicating taking the absolute value of the true envelope signal.

[0022] The output value of the peak search function is , and the formula is as follows: (9); is the index of the true motion spectrum data, which points to each data point in turn during the process of traversing the entire true motion spectrum data. is the spectrum data at The value at a certain point, and are adjacent data points of When the output is 1, the point is a peak point, When the output is 0, the point is not a peak point.

[0023] The Short-Time Fourier Transform (STFT), as the core module of the time-frequency analysis module, performs local analysis on the Doppler signal in the time and frequency domains, providing the instantaneous frequency characteristics of the signal. This processing step lays the foundation for subsequent denoising and feature extraction. As Figure 2 shown, in the comparison graph of the improved threshold function curve, the abscissa is the input signal , and the ordinate is the output signal . In the legend in the upper left corner, the three curves from top to bottom are the hard threshold function curve, the soft threshold function curve, and the improved threshold function curve. The hard threshold algorithm is simple and direct, but it is easy to introduce a large reconstruction error; the soft threshold algorithm can better suppress noise, but it will cause the signal to be over-smoothed and details are easily lost; the improved wavelet denoising threshold algorithm proposed in the present invention dynamically adjusts the processing method of the input signal in different signal regions (sharp changes or flat), and performs better in terms of smoothness and adaptability. It can dynamically adjust the processing method of the input signal according to the local characteristics of the signal or the noise level.

[0024] As Figure 4 shown, Figure 4 the (a) of Figure 4 is the original image, and the (b) of Figure 4 is the fast Fourier transform and envelope graph. Comparing these two images, Figure 4 the (b) is more stable and the peak is more obvious. The output panel displays detailed data, which is more intuitive and convenient for users to view. The single-point Laser Doppler Velocimeter (LDV) is a device used to measure the motion speed information of an object at a specific point. In the present invention, using the laser Doppler effect, when a laser irradiates a moving object, the frequency of the scattered light will change, and the Doppler signal related to the speed is obtained by detecting this frequency change; the disk rotation speed measurement experiment uses the disk as the moving object, and by rotating it, the single-point Laser Doppler Velocimeter measures the speed information during the disk rotation process. In the present invention, the disk is in a rotating state, and the single-point Laser Doppler Velocimeter continuously monitors and collects data on the relevant information during the disk rotation process to obtain the Doppler signal related to the disk motion at different rotation frequencies. The improved Doppler speed measurement method is verified by building a single-point speed measurement system and conducting a disk rotation speed measurement experiment. AsFigure 5 As shown Figure 5 Figure (a) is a spectrum diagram of the traditional Fast Fourier Transform (FFT). The abscissa is frequency with the unit of Hertz (Hz), and the ordinate is amplitude with the unit of decibel milliwatt (dBm). Figure 5 Figure (b) is a spectrum diagram of an improved Doppler signal processing method and system proposed by the present invention. The abscissa is frequency with the unit of Hertz (Hz), and the ordinate is the normalized amplitude. By comparing the two figures, it can be seen that the improved Doppler signal processing method does not deviate from the characteristics of the original spectrum diagram, and the spectrum diagram is clear, has a low signal-to-noise ratio, and prominent peaks. Therefore, the improved Doppler signal processing method and system proposed by the present invention have obvious effects.

[0025] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An improved Doppler signal processing method, characterized in that: It includes the following steps: Step S1, signal acquisition and preprocessing: Acquire the original Doppler signal and perform moving average filtering; Step S2, time-frequency analysis: Use the Hamming window function, improved wavelet threshold denoising and combine with short-time Fourier transform to perform time-domain analysis on the filtered Doppler signal to obtain the motion spectrogram; Step S3, feature extraction: Use Hilbert transform to calculate the true envelope and combine with the peak-seeking function based on envelope analysis to traverse the data in the motion spectrogram to extract peak points; Step S4, judgment and output: Screen the peak points that meet the peak height threshold condition and output the peak points and their position coordinates.

2. An improved Doppler signal processing method according to claim 1, characterized in that: Step S1 specifically includes: Step S11, set the sampling rate to 200 MHz, acquire the original Doppler signal and input it; Step S12: Set the window length to 9 and perform moving average filtering on the input original Doppler signal. This process suppresses random noise while retaining the trend characteristics of the signal. The output of the moving average filtering is expressed as: （1）； Among them, represents the current moment in the time series, and the value range is , is the signal length, is the length of the moving average filter. The moving average filter calculates the average value of the current point at time and the previous points as the filtered signal value. is the average coefficient, is the input signal, is an index variable in the summation operation and , by changing the value, the values at different times in the input signal value are successively selected for accumulation.

3. An improved Doppler signal processing method according to claim 1, characterized in that: Step S2 specifically includes: Step S21, use the Hamming window function to process the filtered Doppler signal to suppress the edge effect of the Doppler signal. The definition of the Hamming window function is as follows: （2）； Among them, is the serial number of the current sampling point, and the value range is , is the length of the Hamming window function; Step S22, use improved wavelet threshold denoising to perform dynamic denoising on the Doppler signal obtained after being processed in Step S21. Through the multi-scale decomposition ability of wavelets, separate the different frequency components in the Doppler signal obtained after being processed in Step S21. The formula of the improved wavelet threshold denoising function is as follows: （3）； Among them, is the output signal, is the input signal, is the absolute value of the input signal, is the defined sign function, which determines the sign of the output signal according to the positive or negative of the input signal ; is the threshold, α and β are adjustment parameters, α controls the contraction intensity and affects the compression amplitude of the output signal, and β controls the smoothness and affects the adjustment of the signal in the transition region; Step S23, perform segmented windowing and apply short-time Fourier transform to the Doppler signal obtained after being processed in Step S22, analyze the time-frequency characteristics of the Doppler signal, capture the frequency characteristics of the Doppler signal changing with time, obtain the original spectrum. The formula for windowing with the window function is: （4）； Among them, is the signal sequence, representing the input signal at the current moment value, is the starting position identifier of the window function on the signal sequence , indicating segmental windowing of the input signal is the windowing expression for the signal sequence, is the length of each segment of the signal sequence;​​ The short-time Fourier transform is used to capture non-stationary signals and abrupt signals. The expression is: （5）； Among them, represents the result of the short-time Fourier transform, is the current moment, is the index parameter for calculating the odd and even segmentation of the signal sequence ; is the starting position of the window function on the signal sequence ; is the frequency index used to determine the frequency position, is the signal length, is the odd sequence point transformation with respect to ; is the even sequence point transformation with respect to ; Step S24, use the spectrum optimization method based on bottom spectrum correction to process the original spectrum; Step S241: Collect a static spectrum without moving object information as the base spectrum ; Step S242: Take the bottom spectrum and the dynamic spectrum containing motion information to perform a difference operation to obtain the true motion spectrum diagram The formula is as follows: （6）。 4. An improved Doppler signal processing method according to claim 1, characterized in that: Step S3 specifically includes: Step S31: After short-time Fourier transform, the first frequency point in the true motion frequency spectrum diagram corresponds to a value of 0, denoted as to remove the DC component of the Doppler signal; ​ Step S32: Calculate the Doppler signal with the DC component removed using Hilbert transform. The coefficient of the Hilbert transform is , being the frequency variable, which is used to determine the value of . When , the negative frequencies in the frequency-domain signal shift forward by phase. When , the positive frequencies in the frequency-domain signal shift backward by phase to obtain a complex signal, and then take the modulus of the obtained complex signal to obtain the true envelope of the frequency-domain signal. The formula for the value of is as follows: （7）； In the formula, is the imaginary unit; Step S33, set the peak height threshold, traverse the entire true motion spectrum data through the peak-seeking function, judge whether each data meets the peak height threshold condition. The data that meets the peak height threshold condition is the peak point. The method for setting the peak height threshold is: Take the absolute value of the calculated true envelope. Among the true envelope data after taking the absolute value, the envelope value greater than 0.5 times of all other data envelope values is set as the peak height threshold. The formula for screening peak points using the peak height threshold is: （8）； Among them, represents the peak height threshold; represents the true envelope signal, is the abbreviation of the absolute value function and represents taking the absolute value of the true envelope signal; The output value of the peak search function is , and the formula is as follows: （9）； is the index of the real motion spectrum data, and during the process of traversing the entire real motion spectrum data, it sequentially points to each data point. is the value of the spectrum data at point. and are adjacent data points of When the output is 1, point is the peak point. When the output is 0, point is not the peak point.

5. An improved Doppler signal processing method according to claim 1, characterized in that: After Step S4, it also includes conducting a disk rotation speed measurement experiment through building a single-point laser Doppler velocimetry system to test the performance of the improved Doppler signal processing method.

6. A system for implementing the improved Doppler signal processing method described in claim 1 above, characterized in that: It includes a preprocessing module, a time-frequency analysis module, and a feature extraction module connected in sequence. Among them, the preprocessing module includes an original signal input module for collecting Doppler signals at a set sampling frequency and a moving average filtering module for filtering the collected Doppler signals with a set window length; the time-frequency analysis module includes a Hamming window function processing module capable of suppressing signal edge effects, an improved wavelet threshold denoising module for dynamically denoising Doppler signals, and a short-time Fourier transform module for capturing the frequency characteristics of Doppler signals; the feature extraction module includes a Doppler peak extraction module based on envelope analysis and then using Hilbert transform and a spectral feature output module for outputting the peak coordinates corresponding to the effective frequency points; the original signal input module, the moving average filtering module, the Hamming window function processing module, the improved wavelet threshold denoising module, the short-time Fourier transform module, the Doppler peak extraction module, and the spectral feature output module are connected in sequence.

Images