Marine lidar signal noise removal method, device, equipment and medium

Abstract

The present application relates to a kind of marine Lidar signal noise removal method, device, equipment and medium, method includes: the information sampling of the sea area to be measured generates airborne laser radar signal, through pre-processing eliminates background noise and corrects sea surface fluctuation error, based on fast fourier transform extraction underwater depth profile frequency domain feature, to the amplitude replacement and spectrum reconstruction of electric noise concentrated frequency band, generate noise reduction radar signal, the method provided in the present application overcomes the vertical resolution loss caused by traditional vertical sliding average algorithm and weak signal error filtering problem, by frequency domain filtering replacement technology retains the original data of non-noise frequency band, while removing electric signal clutter, maintain original vertical resolution, effectively capture the microscale marine dynamic information such as thermocline, suspended solids concentration mutation interface and thin layer phytoplankton aggregation, improve marine environment monitoring precision.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of marine Lidar, in particular to a marine Lidar signal noise removal method, device, equipment and medium. BACKGROUND

[0002] Marine Lidar (LiDAR) is a high-precision marine detection technology based on active laser remote sensing. By emitting laser pulses and receiving their backscattering signals, it can obtain key information such as water optical parameters, seabed topography, and suspended matter distribution in real time, and has irreplaceable advantages in marine mapping, environmental monitoring, resource exploration, and military fields. Its core principle is to use the interaction mechanism of laser and seawater medium to calculate the target distance by measuring the time difference from emission to reception of laser, and to analyze the echo intensity, polarization characteristics and spectral characteristics to retrieve water components. However, marine LiDAR faces serious signal interference problems in practical application, especially the mixed electrical signal clutter in the echo signal, which becomes the core bottleneck restricting data quality and inversion accuracy. The sources of electrical signal clutter mainly include detector dark current, amplifier thermal noise, laser jitter signal and crosstalk in transmission path, etc.

[0003] The traditional mainstream signal processing is to suppress noise by vertically sliding average (Vertical Move Average, VMA) method by sliding average of adjacent vertical layer signals, but its essence is a kind of low-pass filter, which has obvious defects. First, the vertical resolution is lost. The average operation blurs the fine structure of different depth layers such as thermocline and suspended matter concentration abrupt interface, resulting in distortion of layered information, which reduces the ability of marine plankton detection or coral reef ecological monitoring, etc. Second, signal smoothing causes many weak but key features such as thin layer plankton aggregation and microscale density fluctuation caused by turbulence to be filtered out as noise, reducing the detection ability of small-scale dynamic processes. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a marine Lidar signal noise removal method, device, equipment and medium which can remove electrical signal noise while preserving the original vertical resolution and effectively protect weak signal features.

[0005] The purpose of the present application is achieved by the following scheme:

[0006] In a first aspect, the present application provides a marine Lidar signal noise removal method, comprising the following steps:

[0007] S1: Based on the airborne dual-wavelength marine Lidar system, information sampling and signal extraction are performed on the sea area to be measured to generate airborne Lidar signals containing detection information and radar detection parameters of the sea area to be measured;

[0008] S2: preprocessing the airborne lidar signal to generate a preprocessed radar signal with a unified background noise profile;

[0009] S3: analyzing the underwater depth profile of the preprocessed radar signal based on a fast Fourier algorithm to generate a frequency domain distribution signal;

[0010] S4: processing the frequency domain distribution signal, filtering the spectrum of the electric noise concentrated frequency band and reconstructing the time domain waveform signal to generate a noise-reduced radar signal, which is used to indicate the optical parameters of the ocean water body and the seabed topographic features.

[0011] In one embodiment, the S2 of the ocean Lidar signal noise removal method provided by the application specifically comprises the following steps:

[0012] S21: quality control of the airborne lidar signal, eliminating profile signals below a preset data threshold, pure background noise, and abnormal profiles with a deviation greater than a preset deviation threshold according to the peak value distribution of the airborne lidar signal to generate a radar preliminary screening signal;

[0013] S22: position alignment of the radar preliminary screening signal, taking the maximum value of the rising edge slope of the echo signal as the sea surface reference, using the low-energy channel of each wavelength to align the data with the peak position of the positive channel to generate a radar alignment signal;

[0014] S23: height correction of the radar alignment signal, combining the signal trigger light start time, attitude angle, and GPS height of the radar detection parameters to calculate the actual height of the sea surface and correct the elevation error in the radar alignment signal to generate a radar correction signal;

[0015] S24: background noise removal of the radar correction signal, fitting the radar correction signal to generate and remove the background noise profile to generate a preprocessed radar signal.

[0016] In one embodiment, the S3 of the ocean Lidar signal noise removal method provided by the application specifically comprises the following steps:

[0017] S31: time domain discrete sampling of the preprocessed radar signal, converting the continuous time domain signal into a finite length sequence with discrete time domain sampling and equal interval sampling to generate a discrete sampling signal;

[0018] S32: discrete time Fourier transform of the discrete sampling signal, mapping the time domain discrete signal to a continuous frequency spectrum in the frequency domain to generate a frequency domain optimized sequence;

[0019] S33: fast Fourier algorithm transformation of the frequency domain optimized sequence, calculating the frequency domain amplitude distribution of the underwater depth profile based on the discrete Fourier transform formula to generate a frequency domain distribution signal.

[0020] In one embodiment, the calculation formula of the frequency domain distribution signal of the ocean Lidar signal noise removal method provided by the present invention is:

[0021] ;

[0022] in, is the frequency domain distribution signal Frequency components, The discrete sampling signal sampling points, is the length of the discrete sampling signal, is the imaginary unit, is a natural constant.

[0023] In one embodiment, S4 of a method for removing noise from an ocean Lidar signal provided by the present invention specifically includes the following steps:

[0024] S41: Analyze the spectrum characteristics of the frequency domain distribution signal, detect the amplitude mutation area and mark the noise frequency band range, and generate the electric noise concentration band;

[0025] S42: performing amplitude replacement processing on the concentrated frequency band of the electric noise, assigning the amplitude of the noise frequency point to a random sampling value of the adjacent signal frequency band, and generating a noise suppression spectrum;

[0026] S43: reconstructing the noise suppression spectrum through a filter sequence, combining the retained frequency band and the replaced noise frequency band to generate a complete spectrum, and generating a reconstructed spectrum signal;

[0027] S44: Perform an inverse Fourier transform on the reconstructed spectrum signal to convert the frequency domain signal into a time domain waveform to generate a noise-reduced radar signal.

[0028] In one embodiment, S42 of a method for removing noise from an ocean Lidar signal provided by the present invention specifically includes the following steps:

[0029] S421: Perform noise frequency location processing on the concentrated frequency band of electrical noise, identify the spectrum amplitude gradient change in the spectrum distribution, and mark the frequency points where the spectrum amplitude gradient exceeds a preset spectrum threshold as amplitude mutation frequency points, thereby generating a noise frequency point set;

[0030] S422: Randomly sample the signal frequency bands adjacent to the amplitude mutation frequency points based on the noise frequency point set, extract amplitude values ​​from the preset effective frequency band interval without replacement, and generate a random sampling amplitude sequence;

[0031] S423: Perform amplitude assignment on the noise frequency point set and the random sampling amplitude sequence, replace the original amplitude of the noise frequency point with the corresponding value of the sampling sequence, and generate a noise suppression spectrum.

[0032] In one embodiment, the application provides a marine Lidar signal noise removal method S1, which specifically comprises the following steps:

[0033] S11: Based on the airborne dual-wavelength marine Lidar, information sampling is performed on the sea area to be measured, and the laser radar original signal of the sea area to be measured is obtained at a preset sampling rate;

[0034] S12: The laser radar original signal is segmented, and based on a preset underwater depth threshold, profile sampling point data of a specific length underwater is intercepted to generate an airborne laser radar signal.

[0035] In a second aspect, the application provides a marine Lidar signal noise removal device, which is configured with the following modules:

[0036] A laser radar signal acquisition module is configured to sample and extract signals based on an airborne dual-wavelength marine Lidar system on a sea area to be measured, and generate an airborne laser radar signal containing detection information and radar detection parameters of the sea area to be measured;

[0037] A radar signal preprocessing module is configured to preprocess the airborne laser radar signal to generate a uniform background noise profile of the preprocessed radar signal;

[0038] A frequency domain signal analysis module is configured to analyze the underwater depth profile of the preprocessed radar signal based on a fast Fourier algorithm to generate a frequency domain distribution signal;

[0039] A signal denoising reconstruction module is configured to process the frequency domain distribution signal, filter the spectrum of the electric noise concentrated frequency band, and reconstruct the time domain waveform signal to generate a denoised radar signal, which is used to indicate the optical parameters of the marine water body and the seabed topographic features.

[0040] In a third aspect, the application provides a computer device, which comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize any one of the above marine Lidar signal noise removal methods.

[0041] In a fourth aspect, the application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize any one of the above marine Lidar signal noise removal methods.

[0042] In summary, the ocean Lidar signal noise removal method provided by the application utilizes the FFT spectrum analysis technology to perform spectrum analysis on the ocean laser radar signal, filters in the frequency domain, and reconstructs the laser radar signal, which can effectively remove the electrical signal stray waves in the crosstalk input signal, and can be significantly better than the traditional vertical sliding method and can maintain the original vertical resolution without reducing.

[0043] The application can realize ocean laser radar signal noise removal through the frequency domain filtering reconstruction mechanism, while maintaining the core advantage of maintaining the original detection resolution. On the one hand, based on the fast Fourier transform analysis of the preprocessed radar signal, the system performs amplitude replacement processing on the electrical noise concentrated frequency band, which can achieve the effect of preserving the original data integrity of the non-noise frequency band, to solve the problem of vertical resolution loss caused by the traditional vertical sliding average algorithm; on the other hand, the system replaces the amplitude of the noise frequency point by random sampling value and reconstructs the continuous spectrum, realizes complete preservation of weak ocean signals, effectively avoids the misfiltering of weak signal characteristics such as thin layer of plankton aggregation and turbulent microscale density fluctuation in the denoising process; by means of the technical path of inverse Fourier transform to reconstruct the time domain waveform, the high-fidelity restoration of the microstructure characteristics of the ocean can be achieved, so as to realize the accurate capture of the microscale ocean dynamic information such as the thermocline and the concentration mutation interface of suspended solids.

[0044] In order to better understand and implement, the application will be described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 A flowchart of the ocean Lidar signal noise removal method provided by the embodiment of the application is shown in the figure.

[0046] Figure 2 A flowchart of generating a frequency domain distribution signal provided by the embodiment of the application is shown in the figure.

[0047] Figure 3 A flowchart of generating a denoising radar signal provided by the embodiment of the application is shown in the figure.

[0048] Figure 4 An experimental area location and test flight trajectory diagram provided by another embodiment of the application is shown in the figure.

[0049] Figure 5 A diagram of extracting the laser radar original signal by MATLAB to generate an airborne laser radar signal provided by another embodiment of the application is shown in the figure.

[0050] Figure 6 A laser radar signal profile diagram obtained after preprocessing provided by another embodiment of the application is shown in the figure.

[0051] Figure 7A frequency domain distribution signal FFT spectrum analysis chart provided for another embodiment of the application;

[0052] Figure 8 A filtered frequency spectrum schematic diagram provided for another embodiment of the application;

[0053] Figure 9 A comparison chart of an original signal and a reconstructed signal provided for another embodiment of the application;

[0054] Figure 10 A structural schematic diagram of a marine Lidar signal noise removal device provided for another embodiment of the application. DETAILED DESCRIPTION

[0055] For the purpose of facilitating the understanding of the present application, a more comprehensive description will be made below with reference to the relevant drawings. The drawings show the preferred embodiments of the present application. However, the present application can be realized in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the application more thorough and comprehensive.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0057] In one embodiment, as shown in Figure 1 A marine Lidar signal noise removal method is provided, and the embodiment is exemplified by the method applied to a terminal. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is realized through the interaction of the terminal and the server. In the embodiment, the method includes the following steps:

[0058] S1: Based on an airborne dual-wavelength marine laser radar system, information sampling and signal extraction are performed on a to-be-measured sea area to generate an airborne laser radar signal containing to-be-measured sea area detection information and radar detection parameters.

[0059] Specifically, the airborne dual-wavelength marine laser radar system used in the present application includes four receiving channels, namely a 486nm high-gain channel, a 486nm low-gain channel, a 532nm high-gain channel, and a 532nm low-gain channel. The system sets the laser divergence angle to 5mrad, the receiving field of view angle to 28mrad, and the sampling rate to 1GHz, ensuring that each detection profile can obtain 2000 sampling points, wherein the air layer distance resolution is 0.15m and the water layer distance resolution is 0.11m.

[0060] Specifically, the airborne dual-wavelength marine laser radar system emits laser pulses of a specific wavelength to the sea area to be measured according to the preset flight trajectory and parameters during flight. The laser pulses propagate in the water body and interact with seawater, generating backscattering signals. The system receives these backscattering signals and converts them into electrical signals for recording. At the same time, the system records parameters such as the time of laser emission, pulse width, energy, as well as attitude information and position information of the flight platform. These information are integrated into the airborne laser radar signal, which contains the water optical properties of the sea area to be measured, the seabed topographic features, and various parameters in the radar detection process.

[0061] S2: Preprocessing the airborne laser radar signal to generate a uniform background noise profile of the pretreated radar signal.

[0062] Specifically, the airborne laser radar signal inevitably interfered by various noises during the acquisition process, including detector dark current, amplifier thermal noise, laser jitter signal, and crosstalk in the transmission path, etc. These noise signals are mixed with the effective signal, reducing the signal-to-noise ratio of the signal and affecting the subsequent data processing and inversion accuracy.

[0063] Based on this, the system performs filtering processing on the acquired original signal to remove high-frequency interference components; and by analyzing the statistical characteristics of the original airborne laser radar signal, the strength and distribution characteristics of the background noise are calculated. By calculating the statistical parameters such as the mean and variance of the signal, the level of the background noise is estimated; further, the system corrects the original signal according to the estimation result, and adjusts the background noise to a standard noise profile.

[0064] S3: Analyzing the underwater depth profile of the pretreated radar signal based on the fast Fourier algorithm to generate a frequency domain distribution signal.

[0065] Specifically, the Fast Fourier Transform (FFT) is a discrete Fourier transform algorithm that can quickly convert time-domain signals into frequency-domain signals. Specifically, the system performs discrete sampling on the preprocessed radar signal, converting the continuous signal into a discrete sample sequence, which is then input into the FFT algorithm for processing. The FFT algorithm converts the signal from the time domain to the frequency domain through complex number operations on the discrete samples, generating the signal spectrum, which reflects the energy distribution of the signal at different frequencies, including both useful components and noise components in the signal. By analyzing the spectrum, the system can identify the frequency range corresponding to the noise components in the signal, providing a basis for further filtering. This process simplifies complex time-domain signals into frequency-domain signals that are easy to analyze, allowing the system to more effectively remove noise and extract useful information about ocean water optical parameters and seabed topography.

[0066] S4: processing the frequency-domain distributed signal, filtering the spectrum of the electric noise concentrated frequency band and reconstructing the time-domain waveform signal to generate a noise-reduced radar signal, which is used to indicate the ocean water optical parameters and seabed topography features.

[0067] Specifically, the system determines the frequency range of the electric noise concentrated frequency band based on the spectrum analysis results, and filters the spectrum of these frequency ranges, replacing the amplitude of the electric noise with random sampling values of other frequency intervals. The filtered spectrum retains the frequency characteristics of the useful components in the signal while removing the interference of the noise components. Then, the system reconstructs the signal using the filtered spectrum, converting the frequency-domain signal back to the time-domain signal to generate a noise-reduced radar signal. The noise-reduced radar signal has a higher signal-to-noise ratio compared to the original signal, and can more accurately reflect the ocean water optical parameters and seabed topography features. The computer system further analyzes the noise-reduced radar signal to extract water optical parameters and seabed topography information, providing high-quality data support for applications such as marine mapping, environmental monitoring, and resource exploration.

[0068] In summary, the ocean Lidar signal noise removal method provided by the present application uses FFT spectrum analysis technology to analyze the spectrum of the ocean laser radar signal, filters and reconstructs the laser radar signal in the frequency domain, which can effectively remove the electric signal stray waves in the crosstalk signal, and can significantly outperform the traditional vertical sliding method and maintain the original vertical resolution without reducing it.

[0069] The application can realize ocean Lidar signal noise removal through a frequency domain filtering reconstruction mechanism, while maintaining the core advantage of original detection resolution. On the one hand, based on the fast Fourier transform analysis of the preprocessed radar signal, the system performs amplitude replacement processing on the concentrated frequency band of the electrical noise, which can achieve the effect of preserving the original data integrity of the non-noise frequency band, to solve the problem of vertical resolution loss caused by the traditional vertical sliding average algorithm. On the other hand, the system replaces the amplitude of the noise frequency point by random sampling value and reconstructs the continuous frequency spectrum, realizes the complete preservation of weak ocean signals, effectively avoids the misfiltration of weak signal characteristics such as thin layer of plankton aggregation and turbulent microscale density fluctuation in the denoising process; by means of the technical path of inverse Fourier transform to reconstruct the time domain waveform, the high-fidelity restoration of the microstructure characteristics of the ocean can be achieved, so as to realize the accurate capture of the microscale ocean dynamic information such as the thermocline and the concentration mutation interface of suspended solids.

[0070] In one embodiment, the application provides a method for removing ocean Lidar signal noise, which specifically comprises the following steps:

[0071] S11: Based on the information sampling of the airborne dual-wavelength ocean Lidar to the measured sea area, the original laser radar signal of the measured sea area is obtained at a preset sampling rate.

[0072] Specifically, the airborne dual-wavelength ocean Lidar system used in the present application includes four receiving channels, namely 486nm high-gain channel, 486nm low-gain channel, 532nm high-gain channel and 532nm low-gain channel. During flight, the airborne dual-wavelength ocean Lidar system transmits laser pulses of specific wavelengths to the measured sea area according to the preset flight trajectory and parameters. These laser pulses propagate in the water body and interact with seawater, generating backscattering signals. The system receives these backscattering signals and converts them into electrical signals for recording. At the same time, the system records the time, pulse width, energy and other parameters of laser emission, as well as the attitude information and position information of the flight platform. These information are integrated into the airborne laser radar signal, which contains the water optical properties of the measured sea area, the seabed topographic features and various parameters in the radar detection process.

[0073] S12: The laser radar original signal is segmented, and the profile sampling point data of a specific length underwater is intercepted based on a preset underwater depth interception threshold to generate an airborne laser radar signal.

[0074] Specifically, MATLAB is a numerical calculation and data processing tool that can efficiently analyze and process raw signals. The system implements signal segmentation based on MATLAB, reads the raw signals of the laser radar through a pre-set interface, and divides the continuous raw signals into multiple independent detection profile signals according to the time sequence characteristics of the signals. The system performs signal segmentation on the raw signals, divides the continuous signals into multiple independent profiles according to the time sequence, and each profile corresponds to the emission and reception process of a laser pulse, containing water body information within a specific depth range. The system intercepts the profile sampling point data of a specific length under water according to a pre-set underwater depth interception threshold.

[0075] The pre-set underwater depth interception threshold is pre-set according to the requirements of the detection task and the depth characteristics of the target sea area, for example, it can be set to intercept data within a specific depth range under water. The system determines the starting point and ending point of each profile by analyzing the time delay and intensity distribution of the signals, and intercepts the corresponding sampling point data. The intercepted data generates an airborne laser radar signal containing detection information of the sea area to be measured and radar detection parameters, wherein the detection information includes water backscattering characteristics, seabed reflection signals, etc., and the radar detection parameters include laser wavelength, sampling rate, flight height, etc.

[0076] It should be noted that when collecting signals, the airborne laser radar system is inevitably disturbed by various factors, resulting in a large amount of noise components mixed in the signals. These noises may come from the dark current of the detector, the thermal noise of the amplifier, the jitter signal of the laser, and the crosstalk in the transmission path, etc. In addition, due to the influence of atmospheric conditions, changes in the attitude of the flight platform, and factors such as sea wave fluctuations, the sea surface position and height of the signal may also deviate. These interference factors not only reduce the signal-to-noise ratio of the signal, but also may lead to misjudgment of the optical parameters of the ocean water body and the seabed topographic characteristics.

[0077] Based on this, in one of the embodiments, the marine Lidar signal noise removal method provided by the application specifically includes the following steps:

[0078] S21: Perform quality control on the airborne laser radar signal, eliminate profile signals below a pre-set data threshold, pure background noise, and abnormal profiles with a deviation greater than a pre-set deviation threshold according to the peak value distribution of the airborne laser radar signal, and generate a radar preliminary screening signal.

[0079] Specifically, the operator sets a preset data threshold in the noise removal system provided in the embodiment, the threshold is determined based on the statistical analysis result of historical effective signals, and is used to screen out profiles with signal intensity not meeting the detection requirements. The system marks all profiles with signal intensity lower than the preset data threshold as invalid and removes them from the data set. At the same time, the system identifies and removes pure background noise, which refers to the natural noise signal received by the detector without target reflection. It does not contain any ocean environment information, but will interfere with subsequent analysis. Further, the system sets a preset deviation threshold according to the peak value distribution of the airborne laser radar signal. For abnormal profiles whose peak value deviates from the normal range by more than the threshold, the system removes them, thereby generating radar pre-screening signals that are not affected by environmental interference, equipment abnormalities or transmission errors. The signals contain effective detection profiles that have undergone preliminary screening.

[0080] S22: Position alignment is performed on the radar pre-screening signal, the maximum slope of the rising edge of the echo signal is taken as the sea surface reference, the peak position of the low-energy channel and the orthogonal channel of each wavelength is used for data alignment, and a radar aligned signal is generated.

[0081] Specifically, the system analyzes the signal characteristics of different wavelength channels. Since the high-energy channel and the parallel polarization channel are prone to signal saturation in the sea surface area, leading to inaccurate identification of the sea surface position, the low-energy channel and the orthogonal channel of each wavelength are selected as the data source for sea surface reference identification. Specifically, the system calculates the slope of the echo signal of the selected channel, extracts the maximum slope of the rising edge of the echo signal, determines the position corresponding to the maximum slope as the sea surface reference point, and adjusts the spatial coordinates of the signals of all detection profiles with the sea surface reference point as the reference, so that the sea surface reference points of different profiles are in the same depth coordinate. Thus, a radar aligned signal that eliminates the position deviation of the profile caused by the change of flight attitude is generated.

[0082] S23: Height correction is performed on the radar aligned signal, the actual height of the sea surface is calculated in combination with the signal trigger light start time, attitude angle and GPS height of the radar detection parameters, and the height error in the radar aligned signal is corrected, thereby generating a radar corrected signal.

[0083] Specifically, the system extracts the start time of the signal trigger light from the radar detection parameters, combines the time information corresponding to the sea surface reference point in the radar aligned signal, calculates the propagation time of the laser pulse from emission to the sea surface, and converts the propagation time into the propagation distance of the laser in the atmosphere according to the propagation speed of the laser in the atmosphere and the atmospheric refractive index. Secondly, the system calls the pitch angle, roll angle and azimuth angle data in the flight attitude parameters, simultaneously obtains the aircraft height information recorded by the GPS, and converts the relative height data of the aircraft into the actual height data of the sea surface through a three-dimensional coordinate conversion algorithm.

[0084] The system calculates the difference between the actual height of the sea surface and the initial height data in the radar alignment signal, corrects the height information of each sampling point in the radar alignment signal based on the difference, and eliminates the height measurement error caused by sea wave fluctuation and flight attitude change. After the correction is completed, the computer system generates a radar correction signal, and the height data of the signal is consistent with the actual geographic space position.

[0085] S24: background noise removal is performed on the radar correction signal, and a background noise profile is generated and removed by fitting the radar correction signal to generate a preprocessed radar signal.

[0086] Specifically, the background noise profile reflects the natural noise level received by the detector when there is no target reflection signal. Such noise can be caused by electronic noise of the detector itself, environmental light interference or other non-target signal sources. The system extracts sampling point data from the radar correction signal, and performs fitting processing on the depth data of each detection profile. When the data shows a linear change trend, a least squares method is used for linear fitting, and when the data shows an exponential decay characteristic, a nonlinear least squares method is used for exponential fitting to generate a background noise trend line for each profile.

[0087] Further, the system aggregates the background noise trend lines of all profiles, and statistically analyzes each depth layer to take the median value of the noise value of each depth layer, thereby constructing a unified background noise profile covering the entire detection depth range. The depth resolution of the profile is consistent with the radar correction signal, and the system subtracts the background noise profile from the radar correction signal to remove the influence of background noise and generate preprocessed radar signals. These signals have been significantly reduced in noise level while retaining effective information of water optical parameters and seabed topographic features, thereby providing high-quality data support for subsequent signal analysis and feature extraction.

[0088] In one embodiment, as shown in FIG. 3, the method provided by the present application specifically includes the following steps: Figure 2

[0089] S31: time-domain discrete sampling is performed on the preprocessed radar signal to convert the continuous time-domain signal into a discrete, equally spaced, finite-length sequence to generate a discrete sampling signal.

[0090] Specifically, the system performs time-domain discrete sampling on the preprocessed radar signal to convert the continuous time-domain signal into a discrete, equally spaced, finite-length sequence. Preferably, the sampling process can use the Nyquist theorem to ensure that the sampling frequency is higher than twice the highest frequency of the signal to avoid aliasing. The specific steps are as follows:

[0091] Let the continuous time-domain signal be , and the sampling period be , then the discrete sampling signal​ may be expressed as:

[0092] ;

[0093] wherein, , N is the number of sampling points, and represents the length of the finite-length sequence. Through the above operation, the system converts the continuous time-domain signal into a computer-processable, time-domain-discrete and equidistant finite-length sequence, and generates a discrete sampling signal.

[0094] S32: performing a discrete-time Fourier transform on the discrete sampling signal, mapping the time-domain discrete signal to a frequency-domain continuous spectrum, and generating a frequency-domain optimization sequence.

[0095] Specifically, the system performs a discrete-time Fourier transform (DTFT) on the discrete sampling signal , maps the time-domain discrete signal to a frequency-domain continuous spectrum. The definition of DTFT is:

[0096] ;

[0097] wherein, is a continuous angular frequency, is an imaginary unit; since is only nonzero when 0≤ n < N , the actual calculation interval is n=0 to , after transformation, the frequency-domain is a continuous function of angular frequency (unit: rad / s), reflecting the distribution of the signal at different frequency components. The system traverses the value of the angular frequency by a numerical method, calculates the value of corresponding to , maps the time-domain discrete signal to the frequency-domain continuous spectrum, and generates the frequency-domain optimization sequence. Preferably, the calculation formula of the frequency-domain optimization sequence is:

[0098] ;

[0099] wherein, N is the length of the discrete sampling signal.

[0100] S33: performing a fast Fourier algorithm transform on the frequency-domain optimization sequence, calculating the frequency-domain amplitude distribution of the underwater depth profile based on the discrete Fourier transform formula, and generating a frequency-domain distribution signal.

[0101] Specifically, the system performs a fast Fourier algorithm transform on the frequency-domain optimization sequence Perform Fast Fourier Transform (FFT), which is an efficient implementation of Discrete Fourier Transform (DFT). Preferably, the calculation formula for the frequency domain distribution signal is:

[0102] ;

[0103] in, is the frequency domain distribution signal Frequency components, The discrete sampling signal sampling points, is the length of the discrete sampling signal, is the imaginary unit, is a natural constant. The system substitutes each sampling point of the discrete sampling signal into the formula to calculate the various frequency components of the frequency domain distribution signal. The generated frequency domain distribution signal is used to analyze the noise frequency components in the lidar signal.

[0104] The above-mentioned ocean Lidar signal noise removal method can achieve high-fidelity spectral representation of ocean Lidar signals through the synergistic effect of time-frequency domain conversion technology, thereby solving the problem of vertical resolution loss caused by traditional noise suppression methods. Specifically, performing time-domain discrete sampling operations on pre-processed radar signals can achieve accurate signal digitization conversion, converting continuous time-domain signals into discrete sequences of equally spaced samples; mapping time-domain discrete signals to frequency-domain continuous spectra through discrete-time Fourier transform can achieve continuous representation of frequency-domain energy distribution to capture the complete frequency characteristics of water backscatter signals; performing fast Fourier transform on the optimized sequence based on the discrete Fourier transform formula can achieve high-resolution analysis of micro-scale ocean dynamic characteristics, and generate frequency-domain signals representing the optical properties of water bodies through precise calculation of frequency-domain amplitude distribution.

[0105] This application uses the technical path of time domain digitization → frequency domain continuation → discrete spectrum to achieve the integrity preservation of weak signal components, effectively avoiding the traditional time domain averaging algorithm from falsely filtering out weak features such as thin-layer plankton aggregation and turbulent microstructure; at the same time, with the help of the frequency domain continuous mapping mechanism, the original vertical resolution can be losslessly maintained, which is used to realize the precise detection of micro-scale ocean structures such as thermocline interface and suspended matter concentration mutation zone, and significantly improve the credibility of the inversion of ocean environmental parameters.

[0106] In one embodiment, Figure 3 As shown, S4 of the method for removing noise from ocean Lidar signals provided by the present invention specifically includes the following steps:

[0107] S41: Perform spectral feature analysis on the frequency domain distribution signal, detect amplitude mutation regions and mark noise frequency band range, and generate an electric noise concentrated frequency band.

[0108] Specifically, the system performs spectral feature analysis on the frequency domain distribution signal, reads the amplitude data of the frequency domain distribution signal, which is arranged in ascending order of frequency, and contains the amplitude values of each frequency point. Then, the system calculates the amplitude difference value of adjacent frequency points. When the amplitude difference value of a frequency point and its adjacent frequency points exceeds a preset mutation threshold, and a plurality of consecutive frequency points exhibit similar mutation characteristics, the system marks the interval as a potential noise frequency band. The system calculates the amplitude mean value of each potential noise frequency band and compares it with the amplitude mean value of the full frequency band. The frequency band with a significantly higher amplitude mean value than the mean value is determined as an electric noise concentrated frequency band.

[0109] S42: Perform amplitude replacement processing on the electric noise concentrated frequency band, assign noise frequency point amplitude to adjacent signal frequency band random sampling value, and generate a noise suppression spectrum.

[0110] Specifically, the system determines the adjacent frequency bands on both sides of each electric noise concentrated frequency band according to the start frequency and end frequency of each electric noise concentrated frequency band, which are set as adjacent signal frequency bands. The system extracts the amplitude values of all frequency points in each adjacent signal frequency band, and aggregates these amplitude values to form a corresponding sampling data set. For each frequency point in each electric noise concentrated frequency band, the system randomly selects an amplitude value from the adjacent signal frequency band sampling data set corresponding to the noise frequency band, and replaces the original amplitude value of the frequency point with the selected amplitude value.

[0111] During the replacement process, the system keeps the original phase information of each frequency point unchanged and only modifies the amplitude parameter. After the system completes the amplitude replacement of all frequency points in the electric noise concentrated frequency band, it generates a noise suppression spectrum containing the replaced noise frequency band amplitude data and the un-replaced other frequency band amplitude data. Preferably, the noise suppression spectrum of the present application is generated by the following steps:

[0112] S421: Perform noise frequency point positioning processing on the electric noise concentrated frequency band, identify the spectral amplitude gradient changes in the spectral distribution, and mark the frequency points with spectral amplitude gradient exceeding a preset spectral threshold as amplitude mutation frequency points, and generate a noise frequency point set.

[0113] Specifically, the system reads the spectral data of the concentrated frequency band of the electrical noise, calculates the spectral amplitude difference between each frequency point and its adjacent next frequency point in the band, defines the ratio of the difference to the frequency point interval as the spectral amplitude gradient, and at the same time, the system calls a preset spectral threshold, which is determined based on the statistical characteristics of the spectral amplitude gradient of the full frequency band. The system compares the spectral amplitude gradient of each frequency point with the preset spectral threshold. When the spectral amplitude gradient of a certain frequency point exceeds the preset spectral threshold, the frequency point is marked as an amplitude mutation frequency point. Subsequently, all the marked amplitude mutation frequency points are collected, and the frequency values and corresponding index positions of these frequency points are recorded to generate a noise frequency point set, which contains all the frequency point information of the noise concentrated frequency band that needs to be replaced in amplitude.

[0114] S422: Random sampling of the adjacent signal frequency band of the amplitude mutation frequency point based on the noise frequency point set, non-replacement extraction of amplitude values from the preset effective frequency band interval to generate a random sampling amplitude sequence.

[0115] Specifically, the system determines the adjacent signal frequency bands on both sides of the amplitude mutation frequency point in the noise frequency point set that are not marked as noise, and divides these signal frequency bands into a preset effective frequency band interval. The amplitude values of all frequency points in the preset effective frequency band interval are extracted to form an amplitude data set that can be sampled. At the same time, the system performs non-replacement extraction operation on the amplitude data set according to the number of frequency points contained in the noise frequency point set. The number of extracted values is equal to the number of frequency points in the noise frequency point set. The system arranges the extracted amplitude values in the order of extraction to generate a random sampling amplitude sequence. Each element in the sequence corresponds to a noise frequency point to be replaced.

[0116] S423: Amplitude assignment to the noise frequency point set and the random sampling amplitude sequence, replacing the original amplitude of the noise frequency point with the corresponding value of the sampling sequence to generate a noise suppression spectrum.

[0117] Specifically, the system establishes a one-to-one correspondence between each frequency point in the noise frequency point set and each element in the random sampling amplitude sequence. The corresponding order is consistent with the arrangement order of the frequency points in the noise frequency point set and the arrangement order of the amplitude values in the random sampling amplitude sequence. At the same time, the system replaces the original amplitude value of each frequency point in the noise frequency point set with the amplitude value at the corresponding position in the random sampling amplitude sequence. The phase information of each frequency point is kept unchanged during the replacement process. After the system completes the amplitude replacement of all noise frequency points, the processed spectral data and other frequency band data not involved in the replacement are integrated to generate a noise suppression spectrum. The spectrum retains the spectral characteristics of the effective signal and eliminates the abnormal amplitude of the noise frequency point.

[0118] S43: Filter sequence reconstruction of the noise suppression spectrum, combination of the reserved frequency band and the replaced noise frequency band to generate a complete spectrum, generation of a reconstructed spectrum signal.

[0119] Specifically, the system combines the amplitude data of the reserved frequency bands in the noise suppression spectrum that are not marked as the electric noise concentrated frequency bands, with the amplitude data of the electric noise concentrated frequency bands that have undergone the amplitude replacement processing, in the order of frequency from low to high; the system checks the amplitude continuity at the junctions of adjacent frequency bands in the combined spectrum, calculates the amplitude difference between the boundary frequency points of the electric noise concentrated frequency bands and the boundary frequency points of the adjacent reserved frequency bands, and when the difference exceeds a set continuity threshold, the system performs linear interpolation processing on the amplitudes of a number of continuous frequency points at the boundary, so that the amplitude transition of the adjacent frequency bands meets the continuity requirement; at the same time, the system verifies whether the combined spectrum covers the entire frequency range of the original frequency domain distribution signal and whether the number of frequency points is consistent with the original signal. After the verification is completed, the system generates a reconstructed spectrum signal, which contains complete frequency components and corresponding amplitude and phase information.

[0120] S44: inverse Fourier transform is performed on the reconstructed spectrum signal to convert the frequency domain signal into a time domain waveform, and a noise reduction radar signal is generated.

[0121] Specifically, the system performs inverse Fourier transform on the reconstructed spectrum signal. The transform formula is:

[0122] ;

[0123] wherein, represents the reconstructed time domain sampling point, represents the th frequency point component of the reconstructed spectrum signal, represents the total number of frequency points, represents the imaginary unit, represents the natural constant. The system performs complex number operation on each frequency point component in the reconstructed spectrum signal according to the formula, and obtains the value of each time domain sampling point by accumulating the operation results of all frequency point components and dividing by .

[0124] The system arranges the obtained time domain sampling points in the order of time sequence to form a complete time domain waveform, and associates each sampling point in the time domain waveform with corresponding underwater depth information to generate a noise reduction radar signal, which contains time domain waveform data and corresponding depth data after noise processing, and can be used for subsequent analysis of marine water optical parameters and seabed topographic features.

[0125] In summary, the ocean Lidar signal noise removal method provided by the application can realize high-fidelity noise reduction processing of ocean laser radar signals through the frequency domain noise suppression and reconstruction mechanism, so as to solve the vertical resolution loss and weak signal misfiltering problems caused by the traditional vertical sliding average algorithm. Specifically, the spectral feature analysis operation on the frequency domain distributed signal can achieve accurate positioning of the electrical noise frequency band, and the system inherent noise distribution characteristics are identified by detecting the amplitude mutation area; based on the operation of replacing the amplitude of the noise frequency point by the random sampling value of the adjacent signal frequency band, the integrity of the non-noise frequency band data can be maintained, and the misfiltering of the traditional algorithm to the weak backscattering signal can be effectively avoided; through the processing of combining the reserved frequency band and the reconstructed complete frequency spectrum after processing, the optimization control of the frequency domain phase continuity can be achieved, so as to eliminate the waveform distortion at the frequency band junction; finally, the operation of inverse Fourier transform on the reconstructed frequency spectrum can realize the high-precision restoration of the time domain waveform, while removing the electrical signal clutter and maintaining the time resolution of the original signal.

[0126] Through the closed-loop operation of "analysis-replacement-reconstruction-inverse transform", the application can achieve reliable analysis of the micro-scale dynamic characteristics of the ocean, so as to realize complete retention of weak signals such as thin-layer phytoplankton aggregation and turbulent microstructure density fluctuation, and lossless detection of key ocean structures such as thermocline interface and suspended matter concentration mutation zone.

[0127] For example, as shown in Figure 4 , the system controls the laser radar to perform information sampling in the measured sea area, and each profile obtained at a sampling rate of 1GHz has 2000 sampling points, the distance resolution in the air and water is 0.15m and 0.11m respectively, the airborne dual-wavelength ocean laser radar system synchronously records the laser echo signal, flight attitude parameters and GPS positioning information, obtains the laser radar original signal, and extracts the laser radar original signal by using MATLAB, as shown in Figure 5 , the system intercepts the data of a specific length underwater to reduce the data amount, and generates the airborne laser radar signal.

[0128] After obtaining the airborne laser radar signal, based on the processing method provided above, the system performs data quality control, sea surface position alignment, sea surface height correction and background noise removal operations on the airborne laser radar signal, and obtains the profile of the preprocessed radar signal of the uniform background noise profile, as shown in Figure 6 , it can be seen that the signal is obviously mixed with noise with a specific spectrum.

[0129] Subsequently, as shown in Figure 7As shown, the system performs FFT processing on the profile of the preprocessed radar signal to generate a frequency domain distribution signal. Through FFT spectrum analysis of the signal, it can be concluded that the laser radar signal is doped with multiple frequency electrical signal background noise. After completing the FFT processing, the system filters the spectrum of the electrical noise concentrated frequency band, replaces the amplitude P2(f) of the frequency f in the interval (0.9 1.1), (1.9 2.1), (2.9 3.1) and (3.9 4.1) with a random sampling of the amplitude of f in the interval (1.2 1.8), (2.2 2.8), (3.2 3.8) and (4.2 4.8), and the obtained spectrum is as shown in Figure 8 Then, the filtered spectrum is used to reconstruct the signal, and the reconstructed signal and the original signal are compared as shown in Figure 9

[0130] Finally, as shown in Figure 9 The reconstructed signal and the original signal are compared, and it can be seen that the reconstructed signal has obviously filtered out the crosstalk stray wave in the original signal, which can effectively improve the signal-to-noise ratio of the signal.

[0131] It should be understood that although each step in the flowchart involved in each of the above embodiments is shown in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each of the above embodiments can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0132] Based on the same inventive concept, the embodiments of the present application also provide an ocean Lidar signal noise removal device for implementing the above-mentioned ocean Lidar signal noise removal method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more ocean Lidar signal noise removal device embodiments provided below can refer to the limitations of the ocean Lidar signal noise removal method described above, which will not be repeated here.

[0133] Preferably, as shown in Figure 10 The present application provides an ocean Lidar signal noise removal device 500, which is configured with the following modules:

[0134] ​The laser radar signal collection module 510 is configured to collect information and extract signals from the sea area to be measured based on the airborne dual-wavelength marine laser radar system, and generate an airborne laser radar signal containing detection information of the sea area to be measured and radar detection parameters;

[0135] The radar signal preprocessing module 520 is configured to preprocess the airborne laser radar signal to generate a preprocessed radar signal of a uniform background noise profile;

[0136] The frequency domain signal analysis module 530 is configured to analyze the underwater depth profile of the preprocessed radar signal based on a fast Fourier algorithm to generate a frequency domain distribution signal;

[0137] The signal denoising reconstruction module 540 is configured to process the frequency domain distribution signal, filter the spectrum of the electric noise concentrated frequency band, and reconstruct the time domain waveform signal to generate a denoised radar signal, which is used to indicate the optical parameters of the marine water body and the seabed topographic features.

[0138] Preferably, the laser radar signal collection module 510 provided in the application is configured with the following units:

[0139] The sea area laser sampling unit is configured to collect information from the sea area to be measured based on the airborne dual-wavelength marine laser radar, and obtain the laser radar original signal of the sea area to be measured at a preset sampling rate.

[0140] The signal profile intercepting unit is configured to divide the laser radar original signal, intercept profile sampling point data of a specific length underwater based on a preset underwater depth intercepting threshold, and generate the airborne laser radar signal.

[0141] Preferably, the radar signal preprocessing module 520 provided in the application is configured with the following units:

[0142] The signal quality control unit is configured to control the quality of the airborne laser radar signal, eliminate profile signals below a preset data threshold, pure background noise, and abnormal profiles with a deviation greater than a preset deviation threshold according to the peak value distribution of the airborne laser radar signal, and generate a radar preliminary screening signal.

[0143] The position alignment processing unit is configured to align the positions of the radar preliminary screening signal, take the maximum value of the rising slope of the echo signal as the sea surface reference, align the data of the peak positions of the low-energy channels and the positive channels of each wavelength, and generate a radar aligned signal.

[0144] The height correction unit is configured to correct the height of the radar aligned signal, calculate the actual height of the sea surface in combination with the signal trigger light start time, attitude angle, and GPS height of the radar detection parameters, correct the height error in the radar aligned signal, and generate a radar corrected signal.

[0145] The background noise removal unit is used to remove the background noise of the radar correction signal, fit the radar correction signal to generate and remove the background noise profile, and generate a preprocessed radar signal.

[0146] Preferably, the frequency domain signal analysis module 530 provided in this application is configured with the following units:

[0147] The time domain discrete sampling unit is used to perform time domain discrete sampling on the pre-processed radar signal, convert the continuous time domain signal into a finite length sequence of time domain discrete and equally spaced sampling, and generate a discrete sampling signal;

[0148] The time-domain-frequency-domain mapping unit is used to perform discrete-time Fourier transform on the discrete sampling signal, map the time-domain discrete signal to the frequency-domain continuous spectrum, and generate a frequency-domain optimized sequence;

[0149] The frequency domain distribution generation unit is used to perform fast Fourier algorithm transformation on the frequency domain optimization sequence, calculate the frequency domain amplitude distribution of the underwater depth profile based on the discrete Fourier transform formula, and generate a frequency domain distribution signal.

[0150] Preferably, the signal noise reduction and reconstruction module 540 provided in this application is configured with the following units:

[0151] The noise frequency band identification unit is used to analyze the spectrum characteristics of the frequency domain distribution signal, detect the amplitude mutation area and mark the noise frequency band range, and generate the concentrated frequency band of electrical noise;

[0152] The noise amplitude replacement unit is used to perform amplitude replacement processing on the concentrated frequency band of electrical noise, assign the amplitude of the noise frequency point to the random sampling value of the adjacent signal frequency band, and generate a noise suppression spectrum;

[0153] a complete spectrum reconstruction unit, configured to reconstruct the noise suppression spectrum through a filter sequence, combine the retained frequency band with the replaced noise frequency band to generate a complete spectrum, and generate a reconstructed spectrum signal;

[0154] The time domain signal conversion unit is used to perform inverse Fourier transform on the reconstructed spectrum signal, convert the frequency domain signal into a time domain waveform, and generate a noise-reduced radar signal.

[0155] Preferably, the noise amplitude replacement unit comprises a noise frequency point positioning subunit, an adjacent frequency band sampling subunit and an amplitude replacement processing subunit, wherein the noise frequency point positioning subunit is configured to perform noise frequency point positioning processing on the electric noise concentrated frequency band, identify the spectral amplitude gradient change in the spectral distribution, mark the frequency point with a spectral amplitude gradient exceeding a preset spectral threshold as an amplitude mutation frequency point, and generate a noise frequency point set; the adjacent frequency band sampling subunit is configured to perform random sampling on the signal frequency band adjacent to the amplitude mutation frequency point based on the noise frequency point set, and extract amplitude values from a preset effective frequency band interval without replacement to generate a random sampling amplitude sequence; and the amplitude replacement processing subunit is configured to perform amplitude assignment on the noise frequency point set and the random sampling amplitude sequence, replace the original amplitude of the noise frequency point with the corresponding value of the sampling sequence, and generate a noise suppression spectrum.

[0156] In one embodiment, the present application also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned ocean Lidar signal noise removal method when executing the computer program.

[0157] In one embodiment, the present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the above-mentioned ocean Lidar signal noise removal method.

[0158] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples without contradiction.

[0159] For the device embodiment, since it basically corresponds to the method embodiment, the relevant part is described in the part of the method embodiment. The device embodiments described above are only schematic, wherein the components described as separate components can or can not be physically separate, and the components displayed as a unit can or can not be a physical unit, i.e. can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present disclosure according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0160] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of various changes or replacements within the technical range disclosed by the present application, and these should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for ocean Lidar signal noise removal, the method comprising: The method comprises the following steps: S1: based on the airborne dual-wavelength marine laser radar system, information sampling and signal extraction are performed on the sea area to be measured to generate an airborne laser radar signal containing detection information and radar detection parameters of the sea area to be measured; S2: preprocessing the airborne laser radar signal to generate a uniform background noise profile preprocessed radar signal; S3: analyzing the underwater depth profile of the preprocessed radar signal based on a fast Fourier algorithm to generate a frequency domain distribution signal; S4: processing the frequency domain distribution signal, filtering the spectrum of the electric noise concentrated frequency band, and reconstructing the time domain waveform signal to generate a noise reduction radar signal, which is used to indicate the optical parameters of the marine water body and the seabed topographic features; Wherein, the noise reduction radar is generated by the following steps: S41: performing spectrum feature analysis on the frequency domain distribution signal, detecting amplitude mutation regions and marking noise frequency band ranges to generate an electric noise concentrated frequency band; S42: performing amplitude replacement processing on the electric noise concentrated frequency band, assigning noise frequency point amplitude to adjacent signal frequency band random sampling values to generate a noise suppression spectrum; S43: performing filter sequence reconstruction on the noise suppression spectrum to combine the reserved frequency band and the replaced noise frequency band to generate a complete spectrum and generate a reconstructed spectrum signal; S44: performing inverse Fourier transform on the reconstructed spectrum signal to convert the frequency domain signal into a time domain waveform to generate a noise reduction radar signal; Wherein, the noise suppression spectrum is generated by the following steps: S421: performing noise frequency point positioning processing on the electric noise concentrated frequency band, identifying the gradient change of the spectrum amplitude in the spectrum distribution, and marking the frequency points with a spectrum amplitude gradient exceeding a preset spectrum threshold as amplitude mutation frequency points to generate a noise frequency point set; S422: based on the noise frequency point set, randomly sampling the signal frequency band adjacent to the amplitude mutation frequency point to extract amplitude values from a preset effective frequency band interval without replacement to generate a random sampling amplitude sequence; S423: assigning amplitudes to the noise frequency point set and the random sampling amplitude sequence, replacing the original amplitude of the noise frequency point with the corresponding value of the sampling sequence to generate a noise suppression spectrum.

2. The method of claim 1, wherein, The S2 comprises: S21: quality control of the airborne laser radar signal, eliminating profile signals below a preset data threshold, pure background noise, and abnormal profiles with a deviation greater than a preset deviation threshold according to the peak value distribution of the airborne laser radar signal to generate a radar preliminary screening signal; S22: position alignment of the radar preliminary screening signal, taking the maximum value of the rising slope of the echo signal as the sea surface reference, and using the low energy channel and the positive communication channel peak position of each wavelength to perform data alignment to generate a radar alignment signal; S23: height correction of the radar alignment signal, combining the signal trigger light start time, attitude angle, and GPS height of the radar detection parameters to calculate the actual height of the sea surface and correct the height error in the radar alignment signal to generate a radar correction signal; S24: background noise removal of the radar correction signal, fitting the radar correction signal to generate and remove the background noise profile to generate a preprocessed radar signal.

3. The method of claim 1, wherein, The S3 comprises: S31: performing time-domain discrete sampling on the preprocessed radar signal, converting the continuous time-domain signal into a finite-length sequence of time-domain discrete and equally spaced sampling, to generate a discrete sampling signal; S32: Performing discrete-time Fourier transform on the discrete sampling signal, mapping the time-domain discrete signal to a frequency-domain continuous spectrum, and generating a frequency-domain optimized sequence; S33: Performing a fast Fourier transform on the frequency domain optimization sequence, calculating the frequency domain amplitude distribution of the underwater depth profile based on a discrete Fourier transform formula, and generating a frequency domain distribution signal.

4. The method of claim 3, wherein, The calculation formula of the frequency domain distribution signal is: ; in, is the frequency domain distribution signal k Frequency components, The discrete sampling signal n sampling points, is the length of the discrete sampling signal, is the imaginary unit, e is a natural constant.

5. The method according to any one of claims 1 to 4, characterized in that, Said S1 comprises: S11: Sampling information of the sea area to be measured based on the airborne dual-wavelength ocean lidar, and obtaining the original lidar signal of the sea area to be measured at a preset sampling rate; S12: performing signal segmentation on the original laser radar signal, intercepting the profile sampling point data of a specific underwater length based on a preset underwater depth interception threshold, and generating an airborne laser radar signal.

6. An apparatus for ocean Lidar signal noise removal, characterized in that, The device comprises: The laser radar signal acquisition module is used to perform information sampling and signal extraction on the sea area to be measured based on the airborne dual-wavelength ocean laser radar system, and generate an airborne laser radar signal containing the sea area detection information and radar detection parameters; A radar signal preprocessing module, configured to preprocess the airborne laser radar signal to generate a preprocessed radar signal with a uniform background noise profile; A frequency domain signal analysis module, configured to analyze the underwater depth profile of the pre-processed radar signal based on a fast Fourier algorithm to generate a frequency domain distribution signal; a signal noise reduction and reconstruction module, configured to process the frequency domain distributed signal, filter the spectrum of the concentrated frequency band of the electrical noise, and reconstruct the signal of the time domain waveform to generate a noise-reduced radar signal, wherein the noise-reduced radar signal is used to indicate the optical parameters of the ocean water body and the seabed topography characteristics; The noise reduction radar is generated by the following steps: Performing spectrum feature analysis on the frequency domain distribution signal, detecting amplitude mutation areas and marking noise frequency band ranges, and generating electric noise concentrated frequency bands; Performing amplitude replacement processing on the concentrated frequency band of the electric noise, assigning the amplitude of the noise frequency point to a random sampling value of the adjacent signal frequency band, and generating a noise suppression spectrum; Performing filter sequence reconstruction on the noise suppression spectrum, combining the retained frequency band and the replaced noise frequency band to generate a complete spectrum, and generating a reconstructed spectrum signal; Performing an inverse Fourier transform on the reconstructed spectrum signal to convert the frequency domain signal into a time domain waveform to generate a noise-reduced radar signal; The noise suppression spectrum is generated by the following steps: Performing noise frequency location processing on the concentrated frequency band of electrical noise, identifying the spectrum amplitude gradient change in the spectrum distribution and marking the frequency points where the spectrum amplitude gradient exceeds a preset spectrum threshold as amplitude mutation frequency points, thereby generating a noise frequency point set; Randomly sampling the signal frequency bands adjacent to the amplitude mutation frequency points based on the noise frequency point set, extracting amplitude values ​​from a preset effective frequency band interval without replacement, and generating a random sampling amplitude sequence; Amplitude assignment is performed on the noise frequency point set and the random sampling amplitude sequence, and the original amplitude of the noise frequency point is replaced with the corresponding value of the sampling sequence to generate a noise suppression spectrum. 7.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-6 when the computer program is executed by the processor. The computer program is executed by the processor to implement the method of any one of claims 1 to 5.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1 to 5.

Images