Water body apparent spectrum real-time acquisition method, device and equipment based on buoy, medium and product

Abstract

The invention discloses a buoy-based water body apparent spectrum real-time acquisition method, device and equipment, a medium and a product, and relates to the field of marine environment intelligent monitoring, and the method comprises the following steps: preprocessing original marine spectrum data to generate high-quality marine spectrum data with time-space alignment; reconstructing a three-channel spectral data matrix through multi-scale discrete wavelet decomposition and soft threshold quantization compression, inputting the three-channel spectral data matrix into a dynamic weight model, and identifying a spectral abnormal mode; obtaining an abnormal feature mark set, calculating spectral shape parameters, and generating a shape parameter sequence; and fusing the abnormal feature mark set and the shape parameter sequence to generate a spectral morphological feature curve, integrating the three-channel spectral data matrix, the abnormal mode and the shape parameter sequence, and outputting a real-time water body apparent spectrum acquisition report. Through deep fusion of the dynamic weight model and the multi-dimensional morphological features, the accuracy and reliability of spectrum anomaly detection in a complex marine environment are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent monitoring of marine environment, and in particular to a water surface apparent spectrum real-time acquisition method, device, equipment, medium and product based on a buoy. BACKGROUND

[0002] Water surface apparent spectrum monitoring is an important technical means of marine environment remote sensing and field observation, and is widely used in water quality classification, ecological assessment and pollution early warning. The traditional method usually acquires the uplink radiance of water body based on the spectral sensor carried on the buoy platform, and realizes noise suppression and baseline correction through a pretreatment process, completes the spatio-temporal alignment of data through timestamp synchronization and geographic coordinate matching, and then generates a high-quality spectral data set that can be used for analysis.

[0003] However, the existing method relies on static threshold or single statistical indicator for identification of spectral anomalies in complex marine environment, and it is difficult to fully integrate multi-dimensional morphological features and dynamic environmental context information, resulting in limited representation ability of abnormal events; at the same time, when generating the monitoring report, the spectral data and abnormal results are usually presented independently, lacking deep integration of abnormal feature marking and quantitative shape parameters, which limits the comprehensive description ability of the evolution process of the optical state of the water body. SUMMARY

[0004] The purpose of the present application is to provide a water surface apparent spectrum real-time acquisition method, device, equipment, medium and product based on a buoy, which can solve the problems of poor adaptability of abnormal identification caused by dependence on static threshold and single statistical indicator, and insufficient description of state evolution process caused by separate processing of spectral data and abnormal information.

[0005] To achieve the above purpose, the present application provides the following solutions: In a first aspect, the present application provides a water surface apparent spectrum real-time acquisition method based on a buoy, comprising the following steps: Acquire original marine spectral data and perform pretreatment to generate spatio-temporally aligned high-quality marine spectral data; the original marine spectral data includes radiance value, dark current background noise value and marine environment parameter; Perform multi-scale discrete wavelet decomposition and soft threshold quantization processing on the spatio-temporally aligned high-quality marine spectral data to generate marine spectral information; Perform coefficient reconstruction and principal component analysis on the marine spectral information to generate a three-channel spectral data matrix; Input the three-channel spectral data matrix into a dynamic weight model to calculate the Pearson correlation coefficient between each principal component channel; Compare the Pearson correlation coefficient with the reference Pearson correlation coefficient under the corresponding marine environment condition in the historical database to identify the abnormal mode of the spectrum. performing local morphological analysis on the three-channel spectral data matrix based on the abnormal time points identified based on the abnormal pattern of the spectrum, to obtain an abnormal feature marker set; extracting spectral values from the three-channel spectral data matrix according to the time points corresponding to the abnormal feature marker set, calculating a spectral shape parameter, and generating a shape parameter sequence; aligning the abnormal feature marker set with the shape parameter sequence in time and fusing features, to generate a spectral morphological feature curve; integrating the spectral morphological feature curve, the shape parameter sequence, the abnormal pattern of the spectrum, and the three-channel spectral data matrix, to generate a water apparent spectral real-time acquisition report of the buoy.

[0006] Optionally, the original marine spectral data is acquired and preprocessed to generate high-quality marine spectral data aligned in time and space, specifically including the following steps: The radiance value is subjected to a wavelet threshold denoising method to suppress high-frequency random noise; The dark current background noise value is subjected to dark current calibration, and baseline correction is completed by a polynomial fitting correction method; The original marine spectral data is subjected to spectral normalization based on integrated light flux; The original marine spectral data subjected to spectral normalization is subjected to channel redundancy verification by detecting abnormal channels; The original marine spectral data subjected to channel redundancy verification is subjected to timestamp synchronization, and the acquisition time of each piece of original marine spectral data is uniformly calibrated by using the high-precision GPS time signal carried by the buoy; The original marine spectral data subjected to channel redundancy verification is bound with the latitude and longitude coordinates obtained by GPS at the same timestamp, to realize accurate association of spatial positions; According to the solar zenith angle, atmospheric pressure and humidity parameters at the acquisition time, the spectral radiance value is subjected to path radiation and atmospheric scattering correction by using the MODTRAN method, to compensate for spectral distortion caused by changes in atmospheric conditions, to generate high-quality marine spectral data aligned in time and space.

[0007] Optionally, the high-quality marine spectral data aligned in time and space is subjected to multi-scale discrete wavelet decomposition and soft threshold quantization processing, to generate marine spectral information, specifically including the following steps: The high-quality marine spectral data aligned in time and space is subjected to multi-scale discrete wavelet decomposition by using a compression algorithm of wavelet transform, and a biorthogonal wavelet basis bior4.4 is selected to decompose each spectral curve point by point in a fixed waveband range, and the decomposition layer number is set to 5 layers; In the first layer decomposition, the spatio-temporal aligned high-quality ocean spectrum data is respectively convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet, and then the convolution results are down-sampled by every other point to obtain the first layer low-frequency approximation coefficient and the first layer high-frequency detail coefficient; In the second layer decomposition, the first layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the second layer low-frequency approximation coefficient and the second layer high-frequency detail coefficient; In the third layer decomposition, the second layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the third layer low-frequency approximation coefficient and the third layer high-frequency detail coefficient; In the fourth layer decomposition, the third layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the fourth layer low-frequency approximation coefficient and the fourth layer high-frequency detail coefficient; In the fifth layer decomposition, the fourth layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the fifth layer low-frequency approximation coefficient and the fifth layer high-frequency detail coefficient; The soft threshold quantization processing is respectively performed on the first layer high-frequency detail coefficient, the second layer high-frequency detail coefficient, the third layer high-frequency detail coefficient, the fourth layer high-frequency detail coefficient and the fifth layer high-frequency detail coefficient to generate a detail coefficient matrix; The current available communication bandwidth and the remaining battery capacity of the ocean water body buoy are obtained; The detail coefficient matrix is fused with the current available communication bandwidth and the remaining battery capacity of the ocean water body buoy into a sparse coefficient set; The sparse coefficient set is integrated with the wavelet basis bior4.4, the number of decomposition layers and the quantization parameter to generate ocean spectrum information.

[0008] Optionally, the ocean spectrum information is subjected to coefficient reconstruction and principal component analysis to generate a three-channel spectrum data matrix, which specifically includes the following steps: The sparse coefficient set in the ocean spectrum information is subjected to coefficient reconstruction through an inverse wavelet transform algorithm, the fifth layer low-frequency approximation coefficient is taken as the input of the inverse wavelet transform algorithm according to the wavelet basis bior4.4 and the number of decomposition layers recorded in the ocean spectrum information, and inverse transform is performed on the fifth layer high-frequency detail coefficient to generate the reconstructed fourth layer low-frequency approximation coefficient; Inverse transform is performed on the reconstructed fourth layer low-frequency approximation coefficient and the fourth layer high-frequency detail coefficient to generate the reconstructed third layer low-frequency approximation coefficient; performing inverse transform on the reconstructed second layer low frequency approximation coefficients and second layer high frequency detail coefficients to generate reconstructed first layer low frequency approximation coefficients; performing inverse transform on the reconstructed second layer low frequency approximation coefficients and second layer high frequency detail coefficients to generate reconstructed first layer low frequency approximation coefficients; performing up-sampling, filter convolution and summation operations on the reconstructed first layer low frequency approximation coefficients and first layer high frequency detail coefficients to complete the last level of inverse transform to generate a complete high frequency detail coefficient complete wavelet domain; based on the complete high frequency detail coefficient complete wavelet domain, merging the fifth layer low frequency approximation coefficient with the first layer high frequency detail coefficient, the second layer high frequency detail coefficient, the third layer high frequency detail coefficient, the fourth layer high frequency detail coefficient and the fifth layer high frequency detail coefficient layer by layer according to an energy-weighted hierarchical fusion algorithm, and finally fusing into a coefficient spectral curve; According to the coefficient spectral curve, the covariance of the coefficient spectrum is calculated, and the covariance matrix is constructed by a hierarchical filling algorithm; According to the coefficient spectral curve, a sliding window energy analysis method is used for covariance calculation, the sliding window width and step length are set, and the spectral values of the current waveband and its adjacent 7 wavebands before and after each window position are extracted to form a local spectral vector; The covariance estimation value between each waveband is calculated for the set of local spectral vectors at all sampling times; the deviations between each waveband at all window positions are aligned according to the waveband index, and a covariance matrix is constructed by a hierarchical filling algorithm; According to the covariance matrix, the variance contribution rate and the eigenvector between each waveband are calculated, and orthogonal projection and feature compression operations are performed to generate a three-channel spectral data matrix.

[0009] Optionally, the three-channel spectral data matrix is input into a dynamic weight model to calculate the Pearson correlation coefficient between each principal component channel, which specifically includes the following steps: The three-channel spectral data matrix is input into a dynamic weight model, and for each group of time-ordered observation samples in the three-channel spectral data matrix, the first principal component is taken as the dependent variable, and the second principal component is taken as the independent variable. The regression slope and intercept are calculated by a linear regression function to obtain the linear correlation measure between the first principal component and the second principal component; The first principal component is taken as the dependent variable, and the third principal component is taken as the independent variable to calculate the linear correlation measure between the first principal component and the third principal component; The second principal component is taken as the dependent variable, and the third principal component is taken as the independent variable to calculate the linear correlation measure between the second principal component and the third principal component; The three-channel spectral data matrix is subjected to inter-channel statistical analysis by a spectral fingerprint analysis algorithm, and the Pearson correlation coefficients between the first principal component and the second principal component, the first principal component and the third principal component, and the second principal component and the third principal component are calculated.

[0010] Optionally, the Pearson correlation coefficients are compared with reference Pearson correlation coefficients under corresponding marine environmental conditions in a historical database, and the abnormal pattern of the spectrum is identified, specifically including the following steps: The Pearson correlation coefficients calculated by the spectral fingerprint analysis algorithm are compared with the Pearson correlation coefficients in the historical database item by item; the Pearson correlation coefficients in the historical database are reference values of various normal water body states stored in the spectral fingerprint library of the marine water body state, including the mean values of the Pearson correlation coefficients between the first principal component and the second principal component, the first principal component and the third principal component, and the second principal component and the third principal component corresponding to clean water body, eutrophic water body, and high-suspended matter water body; For each pair of Pearson correlation coefficients currently calculated, the corresponding water body type and reference range under similar marine environmental conditions in the historical database are searched, and the deviation between the current value and the reference range is calculated; If the Pearson correlation coefficient deviation of any channel pair exceeds a dynamic threshold value, it is determined that the spectrum shows structural changes inconsistent with the normal pattern, and is identified as an abnormal pattern of the spectrum, and the principal component pair and deviation direction involved in the abnormality are recorded.

[0011] In a second aspect, the present application provides a water body apparent spectrum real-time acquisition device based on a buoy, comprising: A data acquisition module is configured to acquire original marine spectral data and perform preprocessing to generate high-quality marine spectral data aligned in time and space; the original marine spectral data includes radiation brightness values, dark current background noise values, and marine environmental parameters; A wavelet decomposition and soft threshold quantization processing module is configured to perform multi-scale discrete wavelet decomposition and soft threshold quantization processing on the high-quality marine spectral data aligned in time and space to generate marine spectral information; A data reconstruction module is configured to perform coefficient reconstruction and principal component analysis on the marine spectral information to generate a three-channel spectral data matrix; A Pearson correlation coefficient calculation module is configured to input the three-channel spectral data matrix into a dynamic weight model to calculate the Pearson correlation coefficients between the principal components; An abnormality identification module is configured to compare the Pearson correlation coefficients with reference Pearson correlation coefficients under corresponding marine environmental conditions in a historical database to identify an abnormal pattern of the spectrum; a local morphology analysis module, configured to perform local morphology analysis on the three-channel spectral data matrix based on the abnormal time points identified by the abnormal pattern of the spectrum, to obtain an abnormal feature marker set; a spectral shape parameter calculation module, configured to extract spectral values from the three-channel spectral data matrix according to the time points corresponding to the abnormal feature marker set, to calculate spectral shape parameters, and to generate a shape parameter sequence; a feature fusion module, configured to perform time alignment and feature fusion on the abnormal feature marker set and the shape parameter sequence, to generate a spectral morphology feature curve; a report generation module, configured to integrate the spectral morphology feature curve, the shape parameter sequence, the abnormal pattern of the spectrum, and the three-channel spectral data matrix, to generate a water apparent spectral real-time acquisition report of a buoy.

[0012] In a third aspect, the present application provides a computer device, comprising a memory, a processor, a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the steps of the water apparent spectral real-time acquisition method based on a buoy according to any one of the above.

[0013] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the steps of the water apparent spectral real-time acquisition method based on a buoy according to any one of the above.

[0014] In a fifth aspect, the present application provides a computer program product, comprising a computer program, and the computer program is executed by a processor to implement the steps of the water apparent spectral real-time acquisition method based on a buoy according to any one of the above.

[0015] According to the specific embodiments provided by the present application, the present application has the following technical effects: The present application provides a water apparent spectral real-time acquisition method, device, equipment, medium and product based on a buoy, which realizes efficient compression and feature extraction of high-quality marine spectral data with time and space alignment through the steps of multi-scale discrete wavelet decomposition and soft threshold quantization processing, generates marine spectral information containing key information, and further reconstructs into a three-channel spectral data matrix; the present application realizes accurate identification of spectral abnormal patterns by inputting the three-channel spectral data matrix into a dynamic weight model and calculating the Pearson correlation coefficient; the present application dynamically adapts to water quality changes under different environmental backgrounds by combining marine environmental conditions with reference values in a historical database, and improves the accuracy and reliability of abnormal detection. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed in the embodiments will be briefly introduced as follows. Obviously, the accompanying drawings in the following description only only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of these drawings.

[0017] Figure 1 A flowchart of a water surface apparent spectrum real-time acquisition method based on a buoy provided by an embodiment of the present application is shown in the figure. Figure 2 A functional module diagram of a water surface apparent spectrum real-time acquisition device based on a buoy provided by an embodiment of the present application is shown in the figure. Figure 3 A structural diagram of a computer device provided by an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments only only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0019] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] In an exemplary embodiment, as shown in Figure 1 A water surface apparent spectrum real-time acquisition method based on a buoy is provided, which is executed by a computer device, specifically, can be executed by a terminal or a server, or can be executed by a terminal and a server together. In the embodiments of the present application, the method includes the following steps. Step 101: acquiring original marine spectrum data and preprocessing to generate high-quality marine spectrum data aligned in time and space; the original marine spectrum data includes radiation brightness value, dark current background noise value and marine environment parameters.

[0021] It should be noted that the original marine spectrum data is automatically acquired by a hyperspectral radiometer integrated on the buoy platform, which continuously measures the uplink radiance of the water body and the downlink irradiance of the sky in the visible light to near-infrared band with high spectral resolution, obtains the radiation brightness value, and performs dark current calibration before and after each spectrum scan, and obtains the dark current background noise value by closing the sensor shutter; the marine environment parameters are synchronously acquired by the multi-parameter water quality instrument carried by the buoy, including water temperature, salinity, pH value, dissolved oxygen and turbidity.

[0022] The preprocessing includes denoising, baseline correction, spectral normalization and quality evaluation; Specifically, the radiation brightness value in the original marine spectrum data is denoised, the wavelet threshold denoising method is used to suppress high-frequency random noise and improve the signal-to-noise ratio; the dark current background noise value in the original marine spectrum data is calibrated, and the baseline correction is completed by the polynomial fitting correction method; the original marine spectrum data is standardized based on the integral light flux to eliminate the interference of non-water factors caused by the daily variation of light intensity and the fluctuation of weather conditions, and the spectral normalization is completed; and the original marine spectrum data after spectral normalization is verified by the channel redundancy of the abnormal channel detection, and the quality evaluation is completed.

[0023] It should be noted that the integral light flux refers to the total energy value obtained by integrating the spectral radiance or irradiance in a fixed wavelength range, which is used to represent the overall intensity of the light signal.

[0024] The original marine spectrum data after preprocessing is subjected to time stamp synchronization, geographic coordinate matching and atmospheric delay correction operation to generate high-quality marine spectrum data aligned in time and space.

[0025] Further, the original marine spectrum data after preprocessing is subjected to time stamp synchronization: the collection time of each piece of original marine spectrum data is uniformly calibrated by using the high-precision GPS timing signal carried by the buoy, ensuring that the time mark of all original marine spectrum data is accurate to the millisecond level, and eliminating the time deviation caused by asynchronous collection of the sensor; Geographic coordinate matching is performed: the original marine spectrum data after preprocessing is bound with the longitude and latitude coordinates obtained by GPS at the same time stamp, realizing accurate association of spatial position; Atmospheric delay correction operation is performed: according to the solar zenith angle, atmospheric pressure and humidity parameters at the collection time, the MODTRAN method is used to correct the path radiation and atmospheric scattering of the spectral radiance value, compensate for the spectral distortion caused by changes in atmospheric conditions, and generate high-quality marine spectrum data aligned in time and space.

[0026] Step 102: performing multi-scale discrete wavelet decomposition and soft threshold quantization processing on the high-quality marine spectrum data aligned in time and space to generate marine spectrum information.

[0027] The high-quality marine spectrum data aligned in time and space is subjected to multi-scale discrete wavelet decomposition to obtain low-frequency approximation coefficients and high-frequency detail coefficients; Further, the high-quality marine spectrum data aligned in time and space is subjected to multi-scale discrete wavelet decomposition by using the compression algorithm of wavelet transform, and the biorthogonal wavelet basis bior4.4 is selected to decompose each spectrum curve point by point in a fixed wavelength range, and the decomposition layer is set to 5 layers; In the first layer decomposition, the spatio-temporally aligned high-quality marine spectral data is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet respectively, and then the convolution results are down-sampled by every other point to obtain the first layer low-frequency approximation coefficient and the first layer high-frequency detail coefficient, wherein the first layer low-frequency approximation coefficient reflects the overall contour feature of the spectrum, and the first layer high-frequency detail coefficient reflects the local fluctuation of the spectrum at the finest scale; In the second layer decomposition, the first layer low-frequency approximation coefficient is taken as the input, and the low-pass and high-pass filters of the bior4.4 wavelet are convolved and down-sampled again to generate the second layer low-frequency approximation coefficient and the second layer high-frequency detail coefficient, wherein the second layer low-frequency approximation coefficient represents the trend information at a coarser scale, and the second layer high-frequency detail coefficient captures the mutation characteristics at a medium scale; In the third layer decomposition, the second layer low-frequency approximation coefficient is taken as the input, and the filtering and down-sampling operations are repeated to obtain the third layer low-frequency approximation coefficient and the third layer high-frequency detail coefficient; In the fourth layer decomposition, the third layer low-frequency approximation coefficient is taken as the input, and the same operation is performed to generate the fourth layer low-frequency approximation coefficient and the fourth layer high-frequency detail coefficient; In the fifth layer decomposition, the fourth layer low-frequency approximation coefficient is taken as the input, and the filtering and down-sampling are performed again to finally obtain the fifth layer low-frequency approximation coefficient and the corresponding high-frequency detail coefficients of the five layers, i.e., the first layer high-frequency detail coefficient, the second layer high-frequency detail coefficient, the third layer high-frequency detail coefficient, the fourth layer high-frequency detail coefficient and the fifth layer high-frequency detail coefficient.

[0028] The soft threshold quantization processing is performed on the high-frequency detail coefficients to generate a detail coefficient matrix, and the current available communication bandwidth and the remaining battery capacity of the marine water body buoy are fused for analysis to generate marine spectral information; Further, the first layer high-frequency detail coefficient, the second layer high-frequency detail coefficient, the third layer high-frequency detail coefficient, the fourth layer high-frequency detail coefficient and the fifth layer high-frequency detail coefficient are respectively subjected to soft threshold quantization processing, the root mean square of each layer high-frequency detail coefficient is calculated, and smoothing compression is realized; the retained non-zero high-frequency detail coefficients are arranged in the order of decomposition level and waveband to generate a detail coefficient matrix; then the current available communication bandwidth and the remaining battery capacity of the marine water body buoy are obtained, the detail coefficient matrix is fused with the current available communication bandwidth and the remaining battery capacity of the marine water body buoy into a sparse coefficient set, and the wavelet basis bior4.4, the number of decomposition layers 5 and the quantization parameter are integrated to generate marine spectral information.

[0029] Step 103: performing coefficient reconstruction and principal component analysis on the marine spectral information to generate a three-channel spectral data matrix.

[0030] The marine spectral information is subjected to coefficient reconstruction to generate a complete high-frequency detail coefficient wavelet domain, which is fused with the low-frequency approximation coefficient to generate a coefficient spectral curve; The sparse coefficient set in the marine spectral information is reconstructed by the inverse wavelet transform algorithm. According to the wavelet base bior4.4 recorded in the marine spectral information and the decomposition layer number 5, the fifth layer low frequency approximation coefficient is taken as the input of the inverse wavelet transform algorithm, and a first level inverse transform is performed combined with the fifth layer high frequency detail coefficient, that is, the fifth layer low frequency approximation coefficient and the fifth layer high frequency detail coefficient are respectively up-sampled and convolved with the reconstruction filter to generate the fourth layer low frequency approximation coefficient. Then the fourth layer low frequency approximation coefficient and the fourth layer high frequency detail coefficient are operated in the same way to restore the third layer low frequency approximation coefficient. The third layer low frequency approximation coefficient and the third layer high frequency detail coefficient are continuously inverse transformed to obtain the second layer low frequency approximation coefficient. The second layer low frequency approximation coefficient and the second layer high frequency detail coefficient are inverse transformed again to restore the first layer low frequency approximation coefficient. Finally, the first layer low frequency approximation coefficient and the first layer high frequency detail coefficient are up-sampled, filtered and convolved, and summed to complete the last level inverse transform, and generate the complete high frequency detail coefficient complete wavelet domain. On this basis, the fifth layer low frequency approximation coefficient and all levels of high frequency detail coefficients (first layer high frequency detail coefficient, second layer high frequency detail coefficient, third layer high frequency detail coefficient, fourth layer high frequency detail coefficient, fifth layer high frequency detail coefficient) are merged layer by layer according to the energy weighted hierarchical fusion algorithm, and finally fused into a coefficient spectral curve.

[0031] According to the coefficient spectral curve, the covariance of the coefficient spectrum is calculated, and the covariance matrix is constructed by the hierarchical filling algorithm. According to the coefficient spectral curve, the sliding window energy analysis method is used for covariance calculation. The sliding window width and step length are set, the spectral values of the current waveband and its adjacent 7 wavebands before and after are extracted to form a local spectral vector at each window position, and the covariance estimation value between each waveband is calculated for the local spectral vector set at all sampling times. The deviation between each waveband at all window positions is aligned according to the waveband index, and the covariance matrix is constructed by the hierarchical filling algorithm (an initialized zero matrix is filled with the covariance estimation value corresponding to the center waveband of each sliding window to the corresponding row and column positions of the matrix, and a weighted average strategy is used to fuse the contributions of multiple windows for the overlapping area. The non-diagonal elements represent the covariance between different wavebands, and the diagonal elements represent the variance of each waveband. Finally, a complete covariance matrix is generated.

[0032] According to the covariance matrix, the variance contribution rate and the characteristic vector between each waveband are calculated, and the orthogonal projection and feature compression operation are performed to generate a three-channel spectral data matrix.

[0033] Further, according to the covariance matrix, eigenvalue decomposition is performed to solve all eigenvalues and corresponding eigenvectors, each eigenvalue representing the variance size of the corresponding principal component, the eigenvalues are arranged in descending order, and the variance contribution rate between each band is calculated; the eigenvectors corresponding to the first three largest eigenvalues are selected as the principal component direction to form the projection matrix; for each spectral sample in the coefficient spectral curve, the spectral values are combined into a column vector, and the matrix fusion is performed with the projection matrix to complete the orthogonal projection operation, and the original high-dimensional spectral data is mapped to a low-dimensional space formed by three principal components; after orthogonal projection, each spectral data is converted into three values, corresponding to the projection coefficients of the first principal component, the second principal component and the third principal component respectively; finally, a three-channel spectral data matrix is generated by the coefficient matrix analysis method.

[0034] Step 104: inputting the three-channel spectral data matrix into a dynamic weight model to calculate the Pearson correlation coefficients between the principal component channels.

[0035] The three-channel spectral data matrix is spatiotemporally aligned and matched with the marine water quality categories, and the typical pattern of spectral change is extracted; Further, each spectral record arranged in time sequence in the three-channel spectral data matrix is point-by-point matched with the marine water quality category at the corresponding time, and the marine water quality category includes clean water body, eutrophic water body, high-suspended solids water body, oil film pollution water body and colored soluble organic matter enrichment water body. Each category of label is obtained based on synchronous sampling analysis or historical verification data; based on a unified timestamp, it is ensured that each row of data in the three-channel spectral data matrix corresponds to the marine water quality category labeled at the same time accurately; for each marine water quality category, the time series mean curve and the inter-channel correlation characteristics of the three channels under each category are calculated according to the corresponding three-channel spectral data subset; by comparing the time series mean curve and the inter-channel correlation characteristics of the three channels under different categories, the typical pattern of spectral change associated with a specific water quality state is identified, for example, the high-suspended solids water body is characterized by the overall lifting of the first principal component and the enhanced fluctuation of the third principal component. Finally, the spatiotemporal alignment and matching between the three-channel spectral data matrix and the marine water quality category are completed, and the typical pattern of spectral change is extracted.

[0036] It should be noted that the water quality category refers to a classification type according to the concentration or state of physical, chemical and biological parameters in the water body, such as clean water body, eutrophic water body, high-suspended solids water body, oil film pollution water body and colored soluble organic matter enrichment water body.

[0037] The typical pattern of spectral change is combined with the marine environmental conditions to generate a spectral fingerprint library of the marine water body state, and a dynamic weight model is constructed by a multi-channel correlation coefficient weighting method; Further, the typical pattern of spectral change is combined with the marine environmental conditions, and for each type of marine water quality category, such as eutrophic water, high suspended solids water, etc., the corresponding typical pattern of spectral change is sorted, including the response characteristics of each principal component in the three-channel spectral data matrix, the correlation intensity between channels, and the change range, and the marine environmental conditions during the occurrence of the marine water quality category, such as water temperature, salinity, pH value, and turbidity, are associated, for example, eutrophic water often occurs under the condition of water temperature 25~30℃ and salinity 30~33 psu; the typical pattern of spectral change of each type of water quality is paired and stored with the corresponding marine environmental conditions to form an entry with environmental context information; all environmental context information entries are summarized to construct a spectral fingerprint library of marine water body state, and the spectral fingerprint library contains the standard spectral behavior of various types of water bodies under different environmental backgrounds; a dynamic weight model is constructed based on the above data using a multi-channel correlation coefficient weighting method.

[0038] It should be noted that the training process of the dynamic weight model is based on the three-channel spectral data matrix and the corresponding marine water quality category. First, the typical pattern of spectral change of each type of water quality under different marine environmental conditions is extracted from the spectral fingerprint library of marine water body state, including the projection range of each principal component, the average Pearson correlation coefficient between channels, and the covariance structure; the three-channel spectral data matrix is used as input and the corresponding marine water quality category is used as label to construct a training sample set; the weighted least squares method is used to optimize the weight parameters of the correlation coefficient between channels, and the goal is to minimize the Mahalanobis distance between the weighted correlation coefficient vector and the marine water quality category template in the spectral fingerprint library; during the training process, the weight update direction is dynamically adjusted according to the simultaneously recorded marine environmental conditions, for example, the discriminant ability of the correlation between the first principal component and the second principal component is strengthened when the water temperature is higher than 28℃; through batch iterative calculation, a set of optimal weight coefficients is obtained, so that the dynamic weight model can accurately distinguish between different types of water quality under different environmental backgrounds; finally, the weight parameters, discriminant threshold and environmental response function obtained by training are solidified as the core parameter set of the dynamic weight model, and the training of the dynamic weight model is completed.

[0039] The three-channel spectral data matrix is input into the dynamic weight model, the linear correlation measure is calculated, and the Pearson correlation coefficient is calculated through the spectral fingerprint analysis algorithm; Further, the three-channel spectral data matrix is input into a dynamic weight model, and for each group of observation samples arranged in time sequence in the three-channel spectral data matrix, the linear correlation between the first principal component and the second principal component is calculated by taking the first principal component as the dependent variable and the second principal component as the independent variable through a linear regression function; the linear correlation between the first principal component and the third principal component is calculated by taking the first principal component as the dependent variable and the third principal component as the independent variable; and the linear correlation between the second principal component and the third principal component is calculated by taking the second principal component as the dependent variable and the third principal component as the independent variable. Subsequently, the three-channel spectral data matrix is subjected to inter-channel statistical analysis by a spectral fingerprint analysis algorithm to calculate the Pearson correlation coefficients between the first principal component and the second principal component, the first principal component and the third principal component, and the second principal component and the third principal component.

[0040] Step 105: Comparing the Pearson correlation coefficients with reference Pearson correlation coefficients corresponding to the marine environmental conditions in the historical database to identify abnormal patterns of the spectrum.

[0041] Comparing the Pearson correlation coefficients with the Pearson correlation coefficients in the historical database to identify abnormal patterns of the spectrum.

[0042] Further, the Pearson correlation coefficients calculated by the spectral fingerprint analysis algorithm are compared with the Pearson correlation coefficients in the historical database item by item. The Pearson correlation coefficients in the historical database are reference values of various normal water body states stored in the spectral fingerprint library of the marine water body state, including the mean values of the Pearson correlation coefficients between the first principal component and the second principal component, the first principal component and the third principal component, and the second principal component and the third principal component corresponding to categories such as clean water body, eutrophic water body, and high-suspended matter water body. For each pair of Pearson correlation coefficients calculated currently, the corresponding water body type and reference range under similar marine environmental conditions (such as water temperature 20°C-25°C and salinity 32-34 psu) in the historical database are found, the deviation between the current value and the reference range is calculated, for example, the current value of the Pearson correlation coefficient between the first principal component and the second principal component is 0.42, while the historical mean value of clean water body under the same conditions is 0.78, and the deviation is 0.36. If the Pearson correlation coefficient of any channel pair deviates more than a dynamic threshold (usually the value range is not more than ±0.3), it is determined that the spectrum shows structural changes inconsistent with the normal pattern, and is identified as an abnormal pattern of the spectrum, and the principal component pair involved in the abnormality and the deviation direction are recorded.

[0043] Step 106: Based on the abnormal time point identified by the abnormal pattern of the spectrum, performing local morphological analysis on the three-channel spectral data matrix to obtain an abnormal feature marker set.

[0044] Based on the abnormal pattern of spectrum, the time window sequence is expanded forward and backward by multi-scale sliding window analysis, and point-by-point local morphological analysis is performed to identify abnormal mutation points. Further, based on the abnormal pattern of spectrum, the three-layer time window sequence is constructed respectively forward and backward from the identified abnormal occurrence time point as the center, forming a time analysis range covering different response periods; in each layer of time window, the time-aligned data in the three-channel spectral data matrix is called to perform point-by-point local morphological analysis on the spectral value sequence of each principal component channel, calculate the first-order difference to detect the slope mutation point, and calculate the second-order difference to identify the inflection point, and confirm the significant change in combination with the amplitude threshold (usually the value range is: 0~1); for the points that meet the mutation condition, record the time stamp, the principal component channel number and the change direction, and finally summarize the detection results in all windows to identify multiple abnormal mutation points.

[0045] According to the abnormal mutation points, local peak and valley values are extracted by extreme value detection algorithm, and local curvature weighting and time-frequency domain double verification operations are performed to obtain an abnormal feature marker set; Further, according to the abnormal mutation points, for the time neighborhood where each abnormal mutation point is located, the local maximum and minimum values within the time step range are searched by the extreme value detection algorithm, and the condition is that the spectrum value of a point is greater than or less than the values of its two adjacent points, so as to extract the local peak and valley values; then the extracted extreme value points are subjected to local curvature weighting operation, the average absolute value of the second-order difference in the window where the extreme value point is located is calculated as the curvature intensity, and the high curvature points are given higher weights, and then time-frequency domain double verification is performed, the time series is converted to frequency domain by using short-time Fourier transform, and it is checked whether the extreme value point corresponding to the time is in a specific frequency band Energy concentration phenomenon occurs only to keep the extreme value points that are significant in time domain and frequency domain; finally, each extreme value point that passes the verification is generated into a structured marker, including time stamp, principal component channel number, extreme value type (peak or valley), weighted curvature value and confidence score, and all markers are collected to form an abnormal feature marker set.

[0046] Step 107: Extracting spectral values from the three-channel spectral data matrix according to the time points corresponding to the abnormal feature marker set, calculating spectral shape parameters, and generating a shape parameter sequence.

[0047] The abnormal feature marker set is input into the three-channel spectral data matrix, the spectral shape parameters are calculated, and the shape parameter sequence is generated in the order of time windows.

[0048] The specific expression of calculating the spectral shape parameter is: ; Wherein, represents the spectral shape parameter. Spectrum value (principal component projection value from the three-channel spectrum data matrix) representing the first Spectrum value (principal component projection value from the three-channel spectrum data matrix) representing the first Number of wavebands within the analysis window; Shape standard deviation; Mean value of spectrum value within the local window.

[0049] Further, the timestamp corresponding to each abnormal feature marker in the abnormal feature marker set is time-aligned with the three-channel spectrum data matrix, and the spectrum vector at the same time point in the three-channel spectrum data matrix is located; for each matched time point, a set of spectrum shape parameters is calculated based on the spectrum value of the principal component channel to which it belongs, including the integral area under the spectrum curve within the window, the local peak intensity, the half-width, the skewness and the kurtosis, for example, the half-width value within the waveband range at the time point where the local peak is detected; all the calculated spectrum shape parameters are arranged in the order of the timestamps in the abnormal feature marker set in turn, forming a time-ordered multi-dimensional parameter sequence, and finally generating a shape parameter sequence.

[0050] It should be noted that the order of the time windows refers to the time sequence from early to late according to the starting time of the time window.

[0051] Step 108: Time-aligning the abnormal feature marker set with the shape parameter sequence and fusing the features to generate a spectrum morphology feature curve.

[0052] Step 109: Integrating the spectrum morphology feature curve, the shape parameter sequence, the abnormal mode of the spectrum, and the three-channel spectrum data matrix to generate a water apparent spectrum real-time acquisition report of the buoy.

[0053] Resampling and linear interpolation are performed on the abnormal feature marker set and the shape parameter sequence according to the unified time reference of the buoy to generate a time-aligned abnormal feature marker set and a shape parameter sequence; Further, the timestamp sequence of the abnormal feature marker set and the shape parameter sequence is extracted respectively, and the two time sequences are mapped to the same equidistant time axis with the unified time reference of the buoy as the reference. For the abnormal feature marker set, if there is no corresponding marker at the target time point, an empty value is retained, and if there are multiple markers, the marker at the nearest time point is taken. For the shape parameter sequence, the spectrum shape parameter value at the target time point is calculated by using the linear interpolation method. After resampling and interpolation, a time-aligned abnormal feature marker set and a time-aligned shape parameter sequence are generated, which are completely synchronized on the time axis.

[0054] Depth fusion is performed on the time-aligned abnormal feature marker set and the shape parameter sequence to generate a multi-dimensional feature vector sequence; Further, the time-aligned abnormal feature label set and the time-aligned shape parameter sequence are fused horizontally at the same timestamp by a feature splicing method. For each common time point, the abnormal type label, principal component channel number, extreme value type, weighted curvature value and confidence score contained in the time-aligned abnormal feature label set are extracted, and the area under the spectrum curve, local peak intensity, half-height width, skewness and kurtosis of the corresponding time point in the time-aligned shape parameter sequence are extracted. The above two types of feature parameters are combined into a fixed-dimensional vector, and the splicing operation is repeated for all time points to finally generate a multi-dimensional feature vector sequence arranged in chronological order.

[0055] Based on the multi-dimensional feature vector sequence, the curve reconstruction is performed by a piecewise function fitting and event labeling superposition method to generate a spectral morphological feature curve. Further, based on the multi-dimensional feature vector sequence, first, the spectral shape parameter part of each feature vector in the sequence is normalized to map the area under the spectrum curve, local peak intensity, half-height width, skewness and kurtosis to the 0-1 interval, and then a weighted comprehensive index is calculated as a quantitative representation of the spectral morphology, for example, the weights of the parameters are 0.3, 0.25, 0.15, 0.15 and 0.15 respectively, and the weighted sum is used to obtain the normalized spectral morphology index of each time point. Taking time as the horizontal axis and the normalized spectral morphology index as the vertical axis, a function fitting is performed on the discrete index points by using a piecewise cubic spline interpolation method to generate a continuous and smooth baseline curve. Then, the event labeling superposition method is performed to extract the abnormal type label, principal component channel number, extreme value type and confidence score contained in the multi-dimensional feature vector sequence as event information, and to superimpose visual labeling symbols at the corresponding time points of the baseline curve, for example, a red triangle is used to mark an “algal bloom” event, a blue square is used to mark a “suspended matter surge” event, and the confidence score is displayed beside the label. Finally, a spectral morphological feature curve is generated to fully present the dynamic evolution process of the water body spectral morphology and the spatiotemporal distribution of abnormal events.

[0056] The spectral morphological feature curve, shape parameter sequence, abnormal pattern of spectrum and three-channel spectrum data matrix are integrated to generate a real-time acquisition report of the apparent spectrum of the water body of the buoy.

[0057] Further, the spectral morphological feature curve, the shape parameter sequence, the abnormal pattern of the spectrum and the three-channel spectral data matrix are aligned according to a unified timestamp, and all data in a complete monitoring period are extracted with the report generation time as a reference; the spectral morphological feature curve is embedded as a main graph in the report to show the dynamic evolution process of the water body spectrum morphology and abnormal event labeling; the area under the curve, the local peak intensity, the half-height width, the skewness and the kurtosis in the shape parameter sequence are drawn into time sequence curves in the form of subgraphs to assist in explaining the quantitative characteristics of the morphological changes; the abnormal type label, the occurrence time, the confidence score and the principal component channel number involved in the abnormal pattern of the spectrum are arranged into an abnormal event summary table; and the three-channel spectral data matrix is converted into a pseudo-color time sequence graph or a line graph to reflect the continuous change trend of the three principal components, and finally a water body apparent spectrum real-time acquisition report of the buoy is generated.

[0058] Based on the same inventive concept, the embodiments of the present application also provide a device for implementing the above-mentioned water body apparent spectrum real-time acquisition based on a buoy. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the above method, and therefore the specific limitations in one or more water body apparent spectrum real-time acquisition devices based on a buoy embodiments provided below can refer to the limitations of the water body apparent spectrum real-time acquisition method based on a buoy described above, and will not be repeated here.

[0059] In one exemplary embodiment, as shown in Figure 2 a water body apparent spectrum real-time acquisition device based on a buoy is provided, which includes: a data acquisition module 201 configured to acquire raw marine spectrum data and perform preprocessing to generate high-quality marine spectrum data that is spatio-temporally aligned; the raw marine spectrum data includes radiation brightness values, dark current background noise values and marine environment parameters; a wavelet decomposition and soft threshold quantization processing module 202 configured to perform multi-scale discrete wavelet decomposition and soft threshold quantization processing on the high-quality marine spectrum data that is spatio-temporally aligned to generate marine spectrum information; a data reconstruction module 203 configured to perform coefficient reconstruction and principal component analysis on the marine spectrum information to generate a three-channel spectral data matrix; a Pearson correlation coefficient calculation module 204 configured to input the three-channel spectral data matrix into a dynamic weight model to calculate the Pearson correlation coefficients between the principal component channels; an abnormality identification module 205 configured to compare the Pearson correlation coefficients with reference Pearson correlation coefficients corresponding to marine environment conditions in a historical database to identify an abnormal pattern of the spectrum; The local morphology analysis module 206 is configured to perform local morphology analysis on the three-channel spectral data matrix based on the abnormal time points identified by the abnormal pattern of the spectrum, to obtain an abnormal feature marker set; The spectral shape parameter calculation module 207 is configured to extract spectral values from the three-channel spectral data matrix according to the time points corresponding to the abnormal feature marker set, to calculate spectral shape parameters, and to generate a shape parameter sequence. The feature fusion module 208 is configured to perform time alignment and feature fusion on the abnormal feature marker set and the shape parameter sequence, to generate a spectral morphology feature curve. The report generation module 209 is configured to integrate the spectral morphology feature curve, the shape parameter sequence, the abnormal pattern of the spectrum, and the three-channel spectral data matrix, to generate a buoy water apparent spectral real-time acquisition report.

[0060] In an exemplary embodiment, a computer device is provided, which can be a server or a terminal, and an internal structure diagram thereof can be as shown in Figure 3 The computer device includes a processor, a memory, an input / output interface (I / O), and a communication interface. The processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The database of the computer device is configured to store buoy water apparent spectral real-time acquisition data. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to communicate with external terminals through network connection. The computer program is executed by the processor to implement a buoy water apparent spectral real-time acquisition method.

[0061] Those skilled in the art can understand that Figure 3 The structure shown in the above embodiment is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. A specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0062] In an exemplary embodiment, a computer readable storage medium storing a computer program is provided, the computer program, when executed by a processor, implements the steps of any of the above method embodiments.

[0063] In an exemplary embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of any of the above method embodiments.

[0064] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0065] It can be understood by those skilled in the art that all or part of the processes in the above embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above embodiments. Any reference to memory, database or other medium used in the embodiments provided by the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0066] The database involved in each of the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a blockchain, and the like, without being limited thereto. The processor involved in each of the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, and the like, without being limited thereto.

[0067] The technical features of the above embodiments can be combined in any manner. To make the description concise, all possible combinations of the technical features in the above embodiments are not described, but it should be considered that any combination of the technical features is within the scope of the present disclosure, as long as there is no contradiction.

[0068] The principles and implementation manners of the present application are described by using specific examples herein, and the above embodiments are only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, the specific implementation manners and application ranges can be changed according to the idea of the present application. In summary, the content of the present description should not be understood as a limitation of the present application.

Claims

1. A method for real-time acquisition of apparent spectrum of water body based on buoy, characterized in that, The buoy-based water surface apparent spectrum real-time acquisition method comprises the following steps: Collecting original marine spectrum data and preprocessing to generate high-quality marine spectrum data aligned in time and space; the original marine spectrum data includes radiation brightness value, dark current background noise value and marine environment parameters; Performing multi-scale discrete wavelet decomposition and soft threshold quantization processing on the high-quality marine spectrum data aligned in time and space to generate marine spectrum information; Performing coefficient reconstruction and principal component analysis on the marine spectrum information to generate a three-channel spectrum data matrix; Inputting the three-channel spectrum data matrix into a dynamic weight model to calculate the Pearson correlation coefficients between the principal component channels; Comparing the Pearson correlation coefficients with the reference Pearson correlation coefficients under the corresponding marine environment conditions in the historical database to identify the abnormal mode of the spectrum; Based on the abnormal time point identified by the abnormal mode of the spectrum, performing local morphological analysis on the three-channel spectrum data matrix to obtain an abnormal feature marker set; According to the time point corresponding to the abnormal feature marker set, extracting the spectrum value from the three-channel spectrum data matrix, calculating the spectrum shape parameter, and generating a shape parameter sequence; Time aligning and feature fusing the abnormal feature marker set and the shape parameter sequence to generate a spectrum morphological feature curve; Integrating the spectrum morphological feature curve, the shape parameter sequence, the abnormal mode of the spectrum and the three-channel spectrum data matrix to generate a buoy water surface apparent spectrum real-time acquisition report.

2. The buoy-based water surface apparent spectral real-time acquisition method according to claim 1, characterized in that, The collecting original marine spectrum data and preprocessing to generate high-quality marine spectrum data aligned in time and space comprises the following steps: Using a wavelet threshold denoising method to suppress high-frequency random noise on the radiation brightness value; Performing dark current calibration on the dark current background noise value to complete baseline correction by polynomial fitting correction method; Performing spectral normalization on the original marine spectrum data based on integral light flux; Verifying channel redundancy by detecting abnormal channels after spectral normalization of the original marine spectrum data; Synchronizing the time stamp of the original marine spectrum data after channel redundancy verification, and uniformly calibrating the collection time of each original marine spectrum data by using the high-precision GPS time signal carried by the buoy; Binding the original marine spectrum data after channel redundancy verification with the latitude and longitude coordinates obtained by GPS at the same time stamp to realize accurate association of spatial position; According to the solar zenith angle, atmospheric pressure and humidity parameters at the collection time, using MODTRAN method to correct the path radiation and atmospheric scattering of the spectrum radiance value, compensating the spectrum distortion caused by the change of atmospheric conditions, and generating high-quality marine spectrum data aligned in time and space.

3. The buoy-based water surface apparent spectral real-time acquisition method according to claim 1, characterized in that, The multi-scale discrete wavelet decomposition and soft threshold quantization processing on the high-quality marine spectrum data aligned in time and space to generate marine spectrum information comprises the following steps: Performing multi-scale discrete wavelet decomposition on the high-quality marine spectrum data aligned in time and space by wavelet compression algorithm, selecting biorthogonal wavelet basis bior4.4 to decompose each spectrum curve point by point in a fixed waveband range, and setting the decomposition layer number to 5 layers; In the first layer decomposition, the spatio-temporal aligned high-quality ocean spectrum data is respectively convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to obtain the first layer low-frequency approximation coefficient and the first layer high-frequency detail coefficient; In the second layer decomposition, the first layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the second layer low-frequency approximation coefficient and the second layer high-frequency detail coefficient; In the third layer decomposition, the second layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the third layer low-frequency approximation coefficient and the third layer high-frequency detail coefficient; In the fourth layer decomposition, the third layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the fourth layer low-frequency approximation coefficient and the fourth layer high-frequency detail coefficient; In the fifth layer decomposition, the fourth layer low-frequency approximation coefficient is taken as the input, and is convolved with the low-pass filter and the high-pass filter of the bior4.4 wavelet and down-sampled to generate the fifth layer low-frequency approximation coefficient and the fifth layer high-frequency detail coefficient; The first layer high-frequency detail coefficient, the second layer high-frequency detail coefficient, the third layer high-frequency detail coefficient, the fourth layer high-frequency detail coefficient and the fifth layer high-frequency detail coefficient are respectively subjected to soft threshold quantization processing to generate a detail coefficient matrix; The current available communication bandwidth and the remaining battery capacity of the ocean water body buoy are obtained; The detail coefficient matrix is fused with the current available communication bandwidth and the remaining battery capacity of the ocean water body buoy into a sparse coefficient set; The sparse coefficient set is integrated with the wavelet basis bior4.4, the number of decomposition layers and the quantization parameter to generate ocean spectrum information.

4. The buoy-based water surface apparent spectral real-time acquisition method according to claim 3, characterized in that, The coefficient reconstruction and principal component analysis of the ocean spectrum information generate a three-channel spectrum data matrix, which specifically includes the following steps: The sparse coefficient set in the ocean spectrum information is subjected to coefficient reconstruction through an inverse wavelet transform algorithm, the fifth layer low-frequency approximation coefficient is taken as the input of the inverse wavelet transform algorithm according to the wavelet basis bior4.4 and the number of decomposition layers recorded in the ocean spectrum information, and inverse transformation is performed in combination with the fifth layer high-frequency detail coefficient to generate the reconstructed fourth layer low-frequency approximation coefficient; Inverse transformation is performed on the reconstructed fourth layer low-frequency approximation coefficient and the fourth layer high-frequency detail coefficient to generate the reconstructed third layer low-frequency approximation coefficient; Inverse transformation is performed on the reconstructed third layer low-frequency approximation coefficient and the third layer high-frequency detail coefficient to generate the reconstructed second layer low-frequency approximation coefficient; Inverse transformation is performed on the reconstructed second layer low-frequency approximation coefficient and the second layer high-frequency detail coefficient to generate the reconstructed first layer low-frequency approximation coefficient; The reconstructed first layer low-frequency approximation coefficient and the first layer high-frequency detail coefficient are subjected to up-sampling, filter convolution and summation operation to complete the last level of inverse transformation and generate complete high-frequency detail coefficient complete wavelet domain; Based on the complete high-frequency detail coefficient complete wavelet domain, the fifth layer low-frequency approximation coefficient is merged with the first layer high-frequency detail coefficient, the second layer high-frequency detail coefficient, the third layer high-frequency detail coefficient, the fourth layer high-frequency detail coefficient and the fifth layer high-frequency detail coefficient according to an energy-weighted hierarchical fusion algorithm layer by layer, and finally fused into a coefficient spectrum curve; According to the coefficient spectrum curve, the covariance of the coefficient spectrum is calculated, and a covariance matrix is constructed through a hierarchical filling algorithm; According to the coefficient spectrum curve, a sliding window energy analysis method is used for covariance calculation, the sliding window width and step length are set, and the spectral values of the current wave band and its adjacent 7 wave bands are extracted to form a local spectral vector at each window position; The covariance estimation value between each wave band is calculated for the set of local spectral vectors at all sampling times; the deviations between each wave band at all window positions are aligned according to the wave band index, and a covariance matrix is constructed through a hierarchical filling algorithm; According to the covariance matrix, the variance contribution rate and the eigenvector between each wave band are calculated, and orthogonal projection and feature compression operations are performed to generate a three-channel spectral data matrix.

5. The buoy-based water surface apparent spectral real-time acquisition method according to claim 1, characterized in that, The three-channel spectral data matrix is input into a dynamic weight model, and the Pearson correlation coefficients between each principal component channel are calculated, which specifically includes the following steps: The three-channel spectral data matrix is input into a dynamic weight model, and for each group of observation samples arranged in time sequence in the three-channel spectral data matrix, the linear correlation between the first principal component and the second principal component is calculated by taking the first principal component as the dependent variable and the second principal component as the independent variable through a linear regression function. The linear correlation between the first principal component and the third principal component is calculated by taking the first principal component as the dependent variable and the third principal component as the independent variable. The linear correlation between the second principal component and the third principal component is calculated by taking the second principal component as the dependent variable and the third principal component as the independent variable. The Pearson correlation coefficients between each pair of principal components are calculated through spectral fingerprint analysis algorithm for inter-channel statistical analysis of the three-channel spectral data matrix.

6. The buoy-based water surface apparent spectral real-time acquisition method according to claim 5, characterized in that, The Pearson correlation coefficients are compared with the reference Pearson correlation coefficients under corresponding marine environmental conditions in the historical database to identify abnormal patterns of the spectrum, which specifically includes the following steps: The Pearson correlation coefficients calculated by the spectral fingerprint analysis algorithm are compared with the Pearson correlation coefficients in the historical database item by item; the Pearson correlation coefficients in the historical database are reference values of various normal water body states stored in the spectral fingerprint library of marine water body state, including the mean values of the Pearson correlation coefficients between the first principal component and the second principal component, the first principal component and the third principal component, and the second principal component and the third principal component corresponding to clean water body, eutrophic water body and high-suspended solids water body; For each pair of Pearson correlation coefficients currently calculated, the corresponding water body type and reference range under similar marine environmental conditions in the historical database are found, and the deviation between the current value and the reference range is calculated. If the deviation of the Pearson correlation coefficient of any channel pair exceeds the dynamic threshold, it is determined that the spectrum exhibits structural changes inconsistent with the normal mode, identified as an abnormal mode of the spectrum, and the principal component pair and deviation direction involved in the abnormality are recorded.

7. A buoy-based water surface apparent spectrum real-time acquisition device, characterized in that, The buoy-based water body apparent spectrum real-time acquisition device comprises: A data acquisition module is configured to acquire original marine spectrum data and perform preprocessing to generate high-quality marine spectrum data aligned in time and space; the original marine spectrum data includes radiation brightness values, dark current background noise values, and marine environment parameters; A wavelet decomposition and soft threshold quantization processing module is configured to perform multi-scale discrete wavelet decomposition and soft threshold quantization processing on the high-quality marine spectrum data aligned in time and space to generate marine spectrum information; A data reconstruction module is configured to perform coefficient reconstruction and principal component analysis on the marine spectrum information to generate a three-channel spectrum data matrix; A Pearson correlation coefficient calculation module is configured to input the three-channel spectrum data matrix into a dynamic weight model to calculate the Pearson correlation coefficients between the principal components channels; An anomaly identification module is configured to compare the Pearson correlation coefficients with reference Pearson correlation coefficients corresponding to marine environment conditions in a historical database to identify an abnormal mode of the spectrum; A local morphology analysis module is configured to perform local morphology analysis on the three-channel spectrum data matrix based on an abnormal time point identified by the abnormal mode of the spectrum to obtain an abnormal feature marker set; A spectrum shape parameter calculation module is configured to extract spectrum values from the three-channel spectrum data matrix according to time points corresponding to the abnormal feature marker set to calculate spectrum shape parameters and generate a shape parameter sequence; A feature fusion module is configured to perform time alignment and feature fusion on the abnormal feature marker set and the shape parameter sequence to generate a spectrum morphology feature curve; A report generation module is configured to integrate the spectrum morphology feature curve, the shape parameter sequence, the abnormal mode of the spectrum, and the three-channel spectrum data matrix to generate a buoy-based water body apparent spectrum real-time acquisition report.

8. A computer device comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to implement the buoy-based water body apparent spectrum real-time acquisition method of any one of claims 1-6.

9. 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 buoy-based water body apparent spectrum real-time acquisition method of any one of claims 1-6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the buoy-based water body apparent spectrum real-time acquisition method of any one of claims 1-6.

Images