A multi-beam sounding speed optimization method based on noise spectrum evolution characteristics

Abstract

The present application relates to the technical field of ship navigation optimization control, and particularly relates to a multi-beam sounding speed optimization method based on noise spectrum evolution characteristics, comprising the following steps: collecting time-domain noise of each receiving channel and generating time-aligned noise spectrum through short-time Fourier transform; performing sliding window analysis on the noise spectrum, extracting main peak drift, flatness gradient and harmonic coherence characteristics; comparing with historical reference, and marking abnormality and encoding if below the threshold value; calculating the variation coefficient and autocorrelation attenuation and weighting to obtain the non-stationarity parameter; determining the speed reduction according to the parameter interval query speed mapping; performing adjustment and monitoring, and keeping the speed if the abnormality disappears, otherwise, cyclic optimization is performed. Through spectrum feature extraction, evolution matching identification and speed adaptive adjustment, the present application realizes accurate identification of noise spectrum non-stationarity and dynamic interference suppression in the multi-beam sounding process, thereby effectively improving the stability and data quality of the sounding operation.

Description

Technical Field

[0001] This invention relates to the field of shipborne navigation optimization and control technology, and in particular to a multibeam bathymetry speed optimization method based on noise spectrum evolution characteristics. Background Technology

[0002] Multibeam bathymetry systems (MBS) are widely used in seabed topography mapping, underwater channel monitoring, and marine resource surveys due to their high resolution and efficiency. However, factors such as water disturbance, speed variations, platform attitude fluctuations, and equipment aging in the operating environment often result in significant noise fluctuations in the received signals acquired by MBS systems, exhibiting time correlation and spectral non-stationarity. These spectral characteristics not only affect the accuracy of subsequent echo signal identification but also cause the accumulation of bottom tracking errors, thereby impacting the quality of nautical chart products and operational reliability. Therefore, research on noise spectrum evolution monitoring and adaptive speed adjustment mechanisms for MBS systems has significant engineering and application value.

[0003] Existing multibeam bathymetry systems mostly operate at fixed speeds, lacking dynamic sensing and response mechanisms when facing complex noise backgrounds, such as propeller interference, sea state disturbances, or enhanced bottom echoes. On the one hand, traditional noise assessment methods are mostly based on static statistics such as mean and variance, which cannot accurately identify abnormal characteristics in the evolution of the spectrum over time, such as peak frequency shifts, spectral structure disturbances, or harmonic mismatches. On the other hand, existing speed adjustment strategies mostly rely on manual experience or global control commands, lacking a local dynamic speed control mechanism driven by channel noise non-stationarity. This makes it difficult for the system to adapt to complex acoustic environments in a timely manner, easily resulting in excessive interference or ineffective deceleration, affecting bathymetry efficiency and data quality. Summary of the Invention

[0004] This invention provides a multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics. By real-time acquisition of multi-channel received signals, construction of a spectrum evolution feature set, and annotation of anomalous spectrum evolution windows, and by driving a speed adjustment mapping table based on weighted non-stationary parameters, it achieves rapid perception and refined response to changes in the noise environment. This method, while ensuring bathymetry accuracy, can dynamically optimize the balance between speed and system power consumption, improving the robustness and autonomy of bathymetry operations in complex environments, and has significant engineering application value and promotion potential.

[0005] A multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics includes the following steps: S1: Real-time acquisition of time-domain noise signals from each receiving channel of the multibeam echo sounding system; short-time Fourier transform of each channel is performed according to a fixed time window to generate a time-aligned multi-channel noise spectrum sequence. S2, performs a sliding window analysis on the time dimension of the multi-channel noise spectrum sequence, extracts the noise spectrum evolution features in each window, including the main peak frequency drift, the gradient of the spectrum flatness change and the harmonic component coherence coefficient, and forms the noise spectrum evolution feature set at the current moment; S3, compare the similarity between the noise spectrum evolution feature set and the historical benchmark feature set. When the similarity is lower than the set threshold, mark the corresponding time window as an abnormal evolution label sequence and record the abnormal type code. S4. For the abnormal evolution labeled sequence, calculate the coefficient of variation and autocorrelation coefficient decay rate of its noise spectrum evolution characteristics, and use the weighted sum of the coefficient of variation and autocorrelation coefficient decay rate as the noise nonstationarity parameter. S5. Based on the numerical range of the noise non-stationarity parameter, query the pre-established speed adjustment mapping table to determine the speed adjustment amount. S6, execute the speed adjustment and monitor the noise spectrum evolution feature set of the next time window after the adjustment. If the abnormal evolution marker disappears, maintain the current speed; otherwise, enter the next round of optimization loop.

[0006] Optionally, S1 includes: S11, during the operation of the multibeam echo sounder system, sequentially from each receiving channel (denoted as channel). ,in, , (Total number of receiving channels) collect time-domain noise signal The time-domain noise signal of each channel is stored as a string of length [length missing]. A continuous time-series data segment; S12, time-domain noise signal acquired for each channel. Based on a fixed sliding time window (window length) Sliding step size Perform a short-time Fourier transform to generate channels. Spectral sequence under each time window ; S13, all channels arrive of Combined to form a multi-channel spectrum matrix and all time windows The sequences form a time-aligned multichannel noise spectrum sequence.

[0007] Optionally, S2 includes: S21, for each receiving channel Spectrum matrix A sliding window (length) is used in the time dimension. Step size is Extract the spectral sequence within a continuous window and perform normalization; S22, within each window, extract the main peak frequency. It then compares the result with the previous window to calculate the main peak frequency shift. ; S23, calculate the spectral flatness within each window using the normalized spectral sequence within the sliding time window, and calculate the gradient of spectral flatness change. ; S24, Select the main peak frequency The integer multiples of the frequency are taken as harmonic frequencies (e.g. ), calculate the coherence coefficient of harmonic components ; S25, for each channel During the time window Final output noise spectrum evolution characteristics The noise spectrum evolution feature set at the current moment is formed by splicing multiple channels. .

[0008] Optionally, S3 includes: S31, compare the noise spectrum evolution feature set extracted under the current sliding window with the pre-constructed historical benchmark feature set to evaluate the degree of deviation between the current feature distribution and the normal reference state. When the similarity is lower than the set threshold, it is determined that the spectrum evolution behavior in the time window is abnormal and it is initially marked as an abnormal evolution subsequence. S32. For time windows that have been identified as abnormal evolution, assign corresponding abnormal type codes according to their noise spectrum evolution characteristics, and record the time window index and coding information together to construct an abnormal evolution annotation sequence.

[0009] Optionally, S31 includes: S311: Extract representative noise spectrum evolution features from multiple time windows during the stable operation phase, construct a historical benchmark feature set, and generate a historical benchmark feature reference vector. ; S312, using cosine similarity to calculate the noise spectrum evolution characteristics of the current time window. Historical baseline feature reference vector similarity ; S313, Setting the similarity threshold ,like If the noise spectrum evolution characteristics of the current time window deviate from the normal state, the noise spectrum evolution characteristics that deviate from the normal state will be recorded in the anomaly index set.

[0010] Optionally, S311 includes: S3111, during the calibration period or a period without anomalies, extracts the noise spectrum evolution feature set corresponding to the current window from multiple continuous or discrete time windows. ; S3112, stack the multi-channel noise spectrum evolution feature sets within all selected stable windows in chronological order to form a historical baseline feature set. ; S3113 uses an arithmetic mean method to aggregate the historical benchmark feature set, generating a historical benchmark feature reference vector. .

[0011] Optionally, S32 includes: S321, for time windows that have been identified as abnormal evolution Extract its noise spectrum evolution characteristics And based on a predefined set of exception type discrimination rules Perform matching and judgment, and determine the exception type for each rule. Define the characteristic pattern corresponding to the exception type, if It is determined to be a frequency mutation anomaly (coded A1). It was determined to be a sudden change in spectral structure anomaly (coded A2). It was determined to be a harmonic instability anomaly (code A3), among which, The main peak frequency drift threshold. The gradient threshold for spectral flatness variation. Set the lower limit of the coherence coefficient of the harmonic components and output the code of the successfully matched anomaly type. ; S322, Encode according to the generated exception type Forming code pairs If a window matches multiple exception types, use parallel encoding (e.g., "A1|A3"). S323 combines the encoding pairs of all anomalous time windows in chronological order to construct an anomalous evolution annotation sequence. .

[0012] Optionally, S4 includes: S41, for time window segments already marked as anomalous evolution Extract the noise spectrum evolution features corresponding to each window in the segment to form time series for each feature dimension. ; S42, calculate the coefficient of variation of the noise spectrum evolution characteristics by the ratio of the standard deviation to the mean of the time series of each feature dimension within the time window segment of abnormal evolution. ; S43, calculate the autocorrelation coefficient decay rate by measuring the average decay of the autocorrelation coefficients of the time series of each feature dimension within the time window segment of abnormal evolution at multiple lag orders. ; S44, weighted and fused the coefficient of variation of the noise spectrum evolution characteristics with the autocorrelation coefficient decay rate to form the noise nonstationarity parameter. .

[0013] Optionally, S5 includes: S51, Obtain the noise nonstationarity parameters corresponding to the current anomaly evolution annotation sequence. It then compares the results with a predefined nonstationarity classification threshold to classify the nonstationarity parameter levels. These include stable, slightly unstable, and highly unstable states; S52, based on nonstationary parameter levels The corresponding adjustment strategy, including the speed adjustment amount, is retrieved from a pre-built speed adjustment mapping table, as follows: ; in, For speed adjustment amount, For level The corresponding standard adjustment step size, Adjust the level in the speed map table Configuration, if To maintain stability, keep the original speed. For slight instability, reduce speed by 0.5 knots; if... For highly unstable conditions, reduce speed by 1.0 knots; S53, adjust current speed amount Applicable to existing speeds Generate a new recommended speed .

[0014] Optionally, the nonstationarity parameter level The division is represented as: ; in, As a level 1 threshold, It is a level 2 threshold.

[0015] The beneficial effects of this invention are: This invention constructs a time-aligned multi-channel noise spectrum sequence by performing short-time Fourier transform on the time-domain noise signals of each receiving channel of a multibeam echo sounding system. Furthermore, it employs a sliding window in the time dimension to extract noise spectrum evolution characteristics such as the main peak frequency drift, the gradient of spectral flatness change, and the coherence coefficient of harmonic components. This achieves a fine characterization of the dynamic changes in the noise spectrum structure during echo sounding operations, effectively avoiding misjudgments caused by relying solely on a single noise amplitude or instantaneous indicator, and improving the accuracy and robustness of noise state perception.

[0016] This invention compares the similarity of the noise spectrum evolution feature set of the current time window with the historical benchmark feature set constructed based on the stable operation stage, and combines multi-dimensional spectral feature rules to encode and time-series label the abnormal evolution behavior. It can accurately identify the evolution process of the noise spectrum from a stable state to an abnormal state and distinguish different abnormal types such as frequency mutation, sudden change in spectral structure and harmonic instability.

[0017] This invention calculates the coefficient of variation and autocorrelation coefficient decay rate of the noise spectrum evolution characteristics in the anomaly evolution annotation sequence, and then weights and fuses them to form a noise nonstationarity parameter. Based on the classification result of this parameter, a speed adjustment mapping table is consulted to implement graded deceleration control. This allows the speed adjustment strategy to be directly related to the degree of noise nonstationarity, realizing adaptive and progressive optimization of speed. As a result, noise interference can be effectively suppressed in complex operating environments, improving the quality of multibeam bathymetry data and ensuring the continuity and stability of bathymetry operations. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the optimization method flow according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the construction of anomaly evolution annotation sequences according to an embodiment of the present invention. Detailed Implementation

[0020] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. Those skilled in the art may employ other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0021] like Figures 1-2As shown, a multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics includes the following steps: S1: Real-time acquisition of time-domain noise signals from each receiving channel of the multibeam echo sounding system; short-time Fourier transform of each channel is performed according to a fixed time window to generate a time-aligned multi-channel noise spectrum sequence. S2, performs a sliding window analysis on the time dimension of the multi-channel noise spectrum sequence, extracts the noise spectrum evolution features in each window, including the main peak frequency drift, the gradient of the spectrum flatness change and the harmonic component coherence coefficient, and forms the noise spectrum evolution feature set at the current moment; S3, compare the similarity between the noise spectrum evolution feature set and the historical benchmark feature set. When the similarity is lower than the set threshold, mark the corresponding time window as an abnormal evolution label sequence and record the abnormal type code. S4. For the abnormal evolution labeled sequence, calculate the coefficient of variation and autocorrelation coefficient decay rate of its noise spectrum evolution characteristics, and use the weighted sum of the coefficient of variation and autocorrelation coefficient decay rate as the noise nonstationarity parameter. S5. Based on the numerical range of the noise non-stationarity parameter, query the pre-established speed adjustment mapping table to determine the speed adjustment amount. S6, execute the speed adjustment and monitor the noise spectrum evolution feature set of the next time window after the adjustment. If the abnormal evolution marker disappears, maintain the current speed; otherwise, enter the next round of optimization loop.

[0022] S1 includes: S11, during the operation of the multibeam echo sounder system, sequentially from each receiving channel (denoted as channel). ,in, , (Total number of receiving channels) collect time-domain noise signal The time-domain noise signal of each channel is stored as a string of length [length missing]. A continuous time-series data segment is represented as: ; S12, time-domain noise signal acquired for each channel. Based on a fixed sliding time window (window length) Sliding step size Perform a short-time Fourier transform to generate channels. Spectral sequence under each time window , is represented as: ; in, For frequency index, For window functions, The start time of the current window, in steps. Slide to update; S13, all channels arrive of Combined to form a multi-channel spectrum matrix and all time windows The sequences form a time-aligned multichannel noise spectrum sequence, represented as: ; in, For channel In frequency The power spectral density on , This represents the total number of frequency components.

[0023] S2 includes: S21, for each receiving channel Spectrum matrix A sliding window (length) is used in the time dimension. Step size is Extract the spectral sequence within the continuous window and normalize it, as follows: ; in, This is the normalized spectral sequence. For the first The starting point of each time window; S22, within each window, extract the main peak frequency. It then compares the result with the previous window to calculate the main peak frequency shift. , is represented as: ; ; S23, calculate the spectral flatness within each window using the normalized spectral sequence within the sliding time window, and calculate the gradient of spectral flatness change. , is represented as: ; ; S24, Select the main peak frequency The integer multiples of the frequency are taken as harmonic frequencies (e.g. ), calculate the coherence coefficient of harmonic components , is represented as: ; in, , These are the normalized spectral sequences at the main frequency and its harmonics, respectively. For covariance, Standard deviation; S25, for each channel During the time window Final output noise spectrum evolution characteristics The noise spectrum evolution feature set at the current moment is formed by splicing multiple channels. , is represented as: .

[0024] S3 includes: S31, compare the noise spectrum evolution feature set extracted under the current sliding window with the pre-constructed historical benchmark feature set to evaluate the degree of deviation between the current feature distribution and the normal reference state. When the similarity is lower than the set threshold, it is determined that the spectrum evolution behavior in the time window is abnormal and it is initially marked as an abnormal evolution subsequence. S32. For time windows that have been identified as abnormal evolution, assign corresponding abnormal type codes according to their noise spectrum evolution characteristics, and record the time window index and coding information together to construct an abnormal evolution annotation sequence.

[0025] S31 includes: S311: Extract representative noise spectrum evolution features from multiple time windows during the stable operation phase, construct a historical benchmark feature set, and generate a historical benchmark feature reference vector. ; S312, using cosine similarity to calculate the noise spectrum evolution characteristics of the current time window. Historical baseline feature reference vector similarity , is represented as: ; S313, Setting the similarity threshold ,like If the noise spectrum evolution characteristics of the current time window deviate from the normal state, the noise spectrum evolution characteristics that deviate from the normal state will be recorded in the abnormal index set. Similarity threshold Represented as: ; in, This represents the mean of the similarity of spectral evolution features within a historical normal window. The standard deviation of the corresponding similarity It is a regulating factor.

[0026] S311 includes: S3111, during the calibration period or a period without anomalies, extracts the noise spectrum evolution feature set corresponding to the current window from multiple continuous or discrete time windows. ; S3112, stack the multi-channel noise spectrum evolution feature sets within all selected stable windows in chronological order to form a historical baseline feature set. , is represented as: ; in, The number of windows used to construct the reference feature; S3113 uses an arithmetic mean method to aggregate the historical benchmark feature set, generating a historical benchmark feature reference vector. , is represented as: .

[0027] S32 includes: S321, for time windows that have been identified as abnormal evolution Extract its noise spectrum evolution characteristics And based on a predefined set of exception type discrimination rules Perform matching and judgment, and determine the exception type for each rule. Define the characteristic pattern corresponding to the exception type, if It is determined to be a frequency mutation anomaly (coded A1). It was determined to be a sudden change in spectral structure anomaly (coded A2). It was determined to be a harmonic instability anomaly (code A3), among which, The main peak frequency drift threshold. The gradient threshold for spectral flatness variation. Set the lower limit of the coherence coefficient of the harmonic components and output the code of the successfully matched anomaly type. ; ; in, This represents the average peak frequency drift during the stable operation phase. For the corresponding standard deviation, This is the offset tolerance coefficient; ; in, The mean gradient of spectral flatness variation under a stable window. For the corresponding standard deviation, For change response factors; ; in, The mean value of harmonic coherence coefficient under steady-state conditions. For the corresponding standard deviation, This is the structural instability tolerance coefficient; S322, Encode according to the generated exception type Forming code pairs If a window matches multiple exception types, use parallel encoding (e.g., "A1|A3"). S323 combines the encoding pairs of all anomalous time windows in chronological order to construct an anomalous evolution annotation sequence. .

[0028] S4 includes: S41, for time window segments already marked as anomalous evolution Extract the noise spectrum evolution features corresponding to each window in the segment to form time series for each feature dimension. ,in, In the time window The first One noise spectrum evolution characteristic; S42, calculate the coefficient of variation of the noise spectrum evolution characteristics by the ratio of the standard deviation to the mean of the time series of each feature dimension within the time window segment of abnormal evolution. , is represented as: ; in, , Time series Standard deviation and mean; S43, calculate the autocorrelation coefficient decay rate by measuring the average decay of the autocorrelation coefficients of the time series of each feature dimension within the time window segment of abnormal evolution at multiple lag orders. , is represented as: ; in, for In lag order The autocorrelation coefficient under the following conditions for variance The lag order; ; in, , The first , Noise spectrum evolution characteristics of each sample point The length of the time series; ; S44, weighted and fused the coefficient of variation of the noise spectrum evolution characteristics with the autocorrelation coefficient decay rate to form the noise nonstationarity parameter. , is represented as: ; in, Dimensions of spectral evolution features , These are the feature weights for the coefficient of variation and the autocorrelation decay rate, respectively.

[0029] S5 includes: S51, Obtain the noise nonstationarity parameters corresponding to the current anomaly evolution annotation sequence. It then compares the results with a predefined nonstationarity classification threshold to classify the nonstationarity parameter levels. These include stable, slightly unstable, and highly unstable states; S52, based on nonstationary parameter levels The corresponding adjustment strategy, including the speed adjustment amount, is retrieved from a pre-built speed adjustment mapping table, as follows: ; in, For speed adjustment amount, For level The corresponding standard adjustment step size, Adjust the level in the speed map table Configuration, if To maintain stability, keep the original speed. For slight instability, reduce speed by 0.5 knots; if... For highly unstable conditions, reduce speed by 1.0 knots; S53, adjust current speed amount Applicable to existing speeds Generate a new recommended speed , is represented as: .

[0030] Nonstationarity parameter levels The division is represented as: ; in, As a level 1 threshold, It is a level 2 threshold.

[0031] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0032] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics, characterized in that, Includes the following steps: S1: Real-time acquisition of time-domain noise signals from each receiving channel of the multibeam echo sounding system; short-time Fourier transform of each channel is performed according to a fixed time window to generate a time-aligned multi-channel noise spectrum sequence. S2, performs a sliding window analysis on the time dimension of the multi-channel noise spectrum sequence, extracts the noise spectrum evolution features in each window, including the main peak frequency drift, the gradient of the spectrum flatness change and the harmonic component coherence coefficient, and forms the noise spectrum evolution feature set at the current moment; S3, compare the similarity between the noise spectrum evolution feature set and the historical benchmark feature set. When the similarity is lower than the set threshold, mark the corresponding time window as an abnormal evolution label sequence and record the abnormal type code. S4. For the abnormal evolution labeled sequence, calculate the coefficient of variation and autocorrelation coefficient decay rate of its noise spectrum evolution characteristics, and use the weighted sum of the coefficient of variation and autocorrelation coefficient decay rate as the noise nonstationarity parameter. S5. Based on the numerical range of the noise non-stationarity parameter, query the pre-established speed adjustment mapping table to determine the speed adjustment amount. S6, execute the speed adjustment and monitor the noise spectrum evolution feature set of the next time window after the adjustment. If the abnormal evolution marker disappears, maintain the current speed; otherwise, enter the next round of optimization loop.

2. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 1, characterized in that, S1 includes: S11, during the operation of the multibeam echo sounder system, sequentially acquires time-domain noise signals from each receiving channel. The time-domain noise signal of each channel is stored as a string of length [length missing]. A continuous time-series data segment; S12, time-domain noise signal acquired for each channel. Short-time Fourier transform is performed using a fixed sliding time window to generate channels. Spectral sequence under each time window ; S13, all channels arrive of Combined to form a multi-channel spectrum matrix and all time windows The sequences form a time-aligned multichannel noise spectrum sequence.

3. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 2, characterized in that, S2 includes: S21, for each receiving channel Spectrum matrix In the time dimension, a sliding window is used to extract the spectral sequence within a continuous window and perform normalization processing; S22, within each window, extract the main peak frequency. It then compares the result with the previous window to calculate the main peak frequency shift. ; S23, calculate the spectral flatness within each window using the normalized spectral sequence within the sliding time window, and calculate the gradient of spectral flatness change. ; S24, Select the main peak frequency The harmonic frequencies are taken as integer multiples of the harmonic frequency, and the coherence coefficients of the harmonic components are calculated. ; S25, for each channel During the time window Final output noise spectrum evolution characteristics The noise spectrum evolution feature set at the current moment is formed by splicing multiple channels. .

4. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 3, characterized in that, S3 includes: S31, compare the noise spectrum evolution feature set extracted under the current sliding window with the pre-constructed historical benchmark feature set to evaluate the degree of deviation between the current feature distribution and the normal reference state. When the similarity is lower than the set threshold, it is determined that the spectrum evolution behavior in the time window is abnormal and it is initially marked as an abnormal evolution subsequence. S32. For time windows that have been identified as abnormal evolution, assign corresponding abnormal type codes according to their noise spectrum evolution characteristics, and record the time window index and coding information together to construct an abnormal evolution annotation sequence.

5. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 4, characterized in that, S31 includes: S311: Extract representative noise spectrum evolution features from multiple time windows during the stable operation phase, construct a historical benchmark feature set, and generate a historical benchmark feature reference vector. ; S312, using cosine similarity to calculate the noise spectrum evolution characteristics of the current time window. Historical baseline feature reference vector similarity ; S313, Setting the similarity threshold ,like If the noise spectrum evolution characteristics of the current time window deviate from the normal state, the noise spectrum evolution characteristics that deviate from the normal state will be recorded in the anomaly index set.

6. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 5, characterized in that, S311 includes: S3111, during the calibration period or a period without anomalies, extracts the noise spectrum evolution feature set corresponding to the current window from multiple continuous or discrete time windows. ; S3112, stack the multi-channel noise spectrum evolution feature sets within all selected stable windows in chronological order to form a historical baseline feature set. ; S3113 uses an arithmetic mean method to aggregate the historical benchmark feature set, generating a historical benchmark feature reference vector. .

7. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 6, characterized in that, S32 includes: S321, for time windows that have been identified as abnormal evolution Extract its noise spectrum evolution characteristics And based on a predefined set of exception type discrimination rules Perform matching and judgment, and determine the exception type for each rule. Define the characteristic pattern corresponding to the exception type, if If it is determined to be a frequency mutation type of abnormality, If it is determined to be an anomaly of sudden change in spectral structure, It was determined to be a harmonic instability anomaly, among which, The main peak frequency drift threshold. The gradient threshold for spectral flatness variation. Set the lower limit of the coherence coefficient of the harmonic components and output the code of the successfully matched anomaly type. ; S322, Encode according to the generated exception type Forming code pairs If a window matches multiple exception types, parallel encoding is used; S323 combines the encoding pairs of all anomalous time windows in chronological order to construct an anomalous evolution annotation sequence. .

8. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 7, characterized in that, S4 includes: S41, for time window segments already marked as anomalous evolution Extract the noise spectrum evolution features corresponding to each window in the segment to form time series for each feature dimension. ; S42, calculate the coefficient of variation of the noise spectrum evolution characteristics by the ratio of the standard deviation to the mean of the time series of each feature dimension within the time window segment of abnormal evolution. ; S43, calculate the autocorrelation coefficient decay rate by measuring the average decay of the autocorrelation coefficients of the time series of each feature dimension within the time window segment of abnormal evolution at multiple lag orders. ; S44, weighted and fused the coefficient of variation of the noise spectrum evolution characteristics with the autocorrelation coefficient decay rate to form the noise nonstationarity parameter. .

9. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 8, characterized in that, S5 includes: S51, Obtain the noise nonstationarity parameters corresponding to the current anomaly evolution annotation sequence. It then compares the results with a predefined nonstationarity classification threshold to classify the nonstationarity parameter levels. These include stable, slightly unstable, and highly unstable states; S52, based on nonstationary parameter levels The corresponding adjustment strategy, including the speed adjustment amount, is retrieved from a pre-built speed adjustment mapping table, as follows: ； in, For speed adjustment amount, For level The corresponding standard adjustment step size, Adjust the level in the speed map table Configuration, if To maintain stability, keep the original speed. For slight instability, reduce speed by 0.5 knots; if... For highly unstable conditions, reduce speed by 1.0 knots; S53, adjust current speed amount Applicable to existing speeds Generate a new recommended speed .

10. The multi-beam bathymetry speed optimization method based on noise spectrum evolution characteristics according to claim 9, characterized in that, The nonstationary parameter level The division is represented as: ； in, As a level 1 threshold, It is a level 2 threshold.

Images