Water body micro-plastic identification method and system based on hyperspectral sensor

Abstract

The invention provides a water body micro-plastic identification method and system based on a hyperspectral sensor, and relates to the technical field of micro-plastic identification. Hyperspectral data of a water body to be analyzed are obtained based on the hyperspectral sensor and preprocessed, and a local water body component end member library is obtained; matching pixels in the effective wave band spectral data with end members in a water body component end member library, and forming a two-dimensional end member matrix by the successfully matched pixels; forming a two-dimensional hyperspectral matrix by using the effective wave band spectral data; constructing a reconstruction cost function based on the abundance matrix and the two-dimensional end member matrix, and performing optimization solution on the reconstruction cost function to determine an optimal abundance matrix; k-means clustering is carried out based on the row vector of the abundance matrix; and completing the recognition of the water body micro-plastics based on the pixels of the clustering center and the abundance matrix.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of microplastic identification, in particular to a water microplastic identification method and system based on a hyperspectral sensor. BACKGROUND

[0002] With the acceleration of industrialization and the large-scale use of plastic products, microplastic pollution has become a global water environmental problem, posing a potential threat to the ecosystem and human health. Hyperspectral imaging technology can obtain continuous and fine ground object spectral information, and is widely used in remote sensing monitoring of water microplastics. Especially in natural reservoirs, lakes and other open waters, hyperspectral sensors can realize rapid and non-destructive detection of large-area water bodies, providing an important technical means for the identification and quantification of microplastic pollution. However, the natural water environment is complex, and there are many interference substances such as algae, suspended solids and biological membranes on the surface of the water body, and their spectral characteristics are highly overlapped with microplastics in the visible-near infrared band, which poses a serious challenge to traditional methods in practical application.

[0003] Existing microplastic identification methods based on hyperspectral data mostly rely on spectral threshold segmentation or supervised classification algorithms, which can achieve plastic identification to some extent, but in complex scenes dominated by algae coverage or mixed pixels, the recognition accuracy is significantly reduced. Specifically, the existing methods have the following defects: first, there is a lack of fine analysis of the spectral composition of mixed pixels, making it difficult to distinguish microplastic regions wrapped in algae membranes or coexisting with organic matter; second, the clustering process is mostly based on original spectral features, without fully considering the spatial correlation and similarity of material composition between pixels, which easily leads to the phenomenon of "same material different spectrum" or "same spectrum different material"; third, the number of clusters is usually set manually, and the cluster center is initialized randomly, resulting in unstable results and poor repeatability. Therefore, it is urgent to develop a microplastic identification method that can adapt to complex water environments, integrate spectral and spatial information, and have adaptive clustering capability.

[0004] The above information disclosed in the background section is only intended to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The present application relates to the technical field of microplastic identification, in particular to a water microplastic identification method and system based on a hyperspectral sensor.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions: A water microplastic identification method based on a hyperspectral sensor, comprising the following specific steps: Step 1: Obtain hyperspectral data of the water body to be analyzed based on a hyperspectral sensor, analyze the hyperspectral data to obtain the signal-to-noise ratio and standard deviation of each band, filter out meaningless bands based on the signal-to-noise ratio and standard deviation, and perform PCA analysis on the remaining bands to eliminate low-contribution bands and obtain effective band spectral data; Step 2: Obtain a local water component endmember library, match each pixel in the effective band spectral data with each endmember in the water component endmember library, and construct a two-dimensional endmember matrix from the pixels that match successfully; and expand the effective band spectral data into a two-dimensional hyperspectral matrix; Step 3: Reconstruct the two-dimensional hyperspectral matrix based on the abundance matrix and the two-dimensional endmember matrix, construct a reconstruction cost function, and optimize and solve the reconstruction cost function to obtain an abundance matrix at which the reconstruction cost is the smallest; Step 4: Extract the eigenvalues of the abundance matrix, determine the number of clusters k using the spectral gap method based on the eigenvalues of the abundance matrix, and perform K-means clustering based on the row vectors of the abundance matrix; Step 5: Complete the identification of water microplastics based on the clustering results and the abundance matrix.

[0007] Further, the mean and standard deviation of the reflectivity of all pixels in each band are calculated, the absolute value of the mean reflectivity of all pixels is divided by the standard deviation of all pixels in the same band, the signal-to-noise ratio of the corresponding band is obtained, a preset signal-to-noise ratio threshold and a standard deviation threshold, and the bands that simultaneously satisfy the signal-to-noise ratio being less than the signal-to-noise ratio threshold and the standard deviation being less than the standard deviation threshold are referred to as meaningless bands.

[0008] Further, the hyperspectral data of the remaining bands is constructed into a hyperspectral matrix, the hyperspectral matrix is standardized, the covariance matrix of the standardized hyperspectral matrix is calculated, the eigenvalues and corresponding eigenvectors of the covariance matrix are extracted, the largest and smallest standardized eigenvalues are sorted in descending order, and the cumulative contribution rate is greater than 95% as a constraint condition, the first I eigenvalues that meet the constraint condition are taken as principal components, the component retention coefficient of each band on the principal components is determined, the component retention coefficients of each band are sorted in descending order, the first 10% of the bands are referred to as effective bands, and the remaining bands are referred to as low-contribution bands, and the low-contribution bands are eliminated.

[0009] The formula is: ; ; Wherein, is the cumulative contribution rate of the first principal components, is the component retention coefficient of the jth remaining band, is the eigenvalue corresponding to the i-th principal component, i is the index of the principal component, and j is the index of the remaining waveband, the sum of all eigenvalues, is the value corresponding to the j-th remaining waveband in the eigenvector corresponding to the i-th principal component; Further, the water component endmember library is the complete spectral data of various pure substances in the local water body, and the logic for matching each pixel in the effective waveband spectral data with each endmember in the water component endmember library is as follows: the spectral angle distance of the calculated pixel and the effective waveband spectral data of each endmember in the water component endmember library is calculated, and if the smallest spectral angle distance is less than the spectral angle distance threshold value of the corresponding endmember, the matching is successful.

[0010] Further, the two-dimensional pixel coordinates are flattened into one dimension, and the flattened pixel coordinate information and the pixel spectral information are used to form a two-dimensional hyperspectral matrix, wherein the two-dimensional pixel coordinate space is a pixel space; each pixel in the two-dimensional hyperspectral matrix is reconstructed as a superposition of various endmembers using the abundance matrix and the two-dimensional endmember matrix; and a reconstruction cost function is constructed as: ; wherein, is the reconstruction cost function, is the two-dimensional hyperspectral matrix, specifically a B-row N-column matrix, B is the total number of pixels in the effective waveband spectral data, and N is the total number of effective wavebands in the effective waveband spectral data, is the two-dimensional endmember matrix, specifically an M-row N-column matrix, is the abundance matrix, specifically a B-row M-column matrix, M is the number of matched endmembers, is the abundance vector of the i-th pixel in the abundance matrix, is the abundance vector of the j-th pixel in the 4-adjacent neighborhood in the abundance matrix, wherein, is the 4-adjacent neighborhood of the i-th pixel, is a sparse parameter, is a spatial smoothing parameter, , is the index of the pixel in the pixel space; is the square of the Frobenius norm of is the L1 norm of is the square of the Euclidean distance between and ; is the square of the Euclidean distance between and ; solves the corresponding abundance matrix when minimizing .

[0011] ​​Further, the logic for determining the number of clusters k based on the spectral gap method is as follows: arrange the eigenvalues of the abundance matrix in ascending order, and the arrangement result is: ; wherein, is the tth eigenvalue of the abundance matrix, is the index of the eigenvalue of the abundance matrix, T is the total number of eigenvalues of the abundance matrix, and is selected when the maximum is taken as the number of clusters k.

[0012] Further, randomly select k row vectors in the row vector matrix of the abundance matrix as cluster centers, calculate the Euclidean distance between each row vector of the abundance matrix and each cluster center, assign each row vector of the abundance matrix to the cluster corresponding to the nearest cluster center, update the cluster center position with the mean of the row vectors in each cluster, iterate until a preset number of iterations is reached, and output the clustering result, which includes the pixel coordinate set of each class label.

[0013] Further, for the same class label under the pixel, calculate the abundance matrix component of each pixel matched with the endmember in each abundance matrix, called the endmember abundance component, calculate the sum of all pixels in the same endmember abundance component under the same class label, and calculate the sum of all endmember abundance components of all pixels, and obtain the proportion of the same kind of microplastic in all pixels under the corresponding class label by dividing the sum of all pixels in the same kind of microplastic endmember abundance component by the sum of all pixels in all endmember abundance components. The present application further provides a water microplastic identification system based on a hyperspectral sensor, which is used to implement the water microplastic identification method based on the hyperspectral sensor, and specifically comprises: A component extraction module is configured to obtain hyperspectral data of a water body to be analyzed based on a hyperspectral sensor, analyze the hyperspectral data to obtain the signal-to-noise ratio and standard deviation of each waveband, filter out meaningless wavebands based on the signal-to-noise ratio and standard deviation, perform PCA analysis on the remaining wavebands to eliminate low-contribution wavebands, and obtain effective waveband spectral data. An endmember matching module is configured to obtain a local water component endmember library, match each pixel in the effective waveband spectral data with each endmember in the water component endmember library, and construct a two-dimensional endmember matrix with the pixels that match successfully; and expand the effective waveband spectral data into a two-dimensional hyperspectral matrix. A reconstruction optimization module is configured to reconstruct the two-dimensional hyperspectral matrix based on the abundance matrix and the two-dimensional endmember matrix, construct a reconstruction cost function, and optimize and solve the reconstruction cost function to obtain the abundance matrix when the reconstruction cost is the smallest. The similar clustering module is used for determining a spectral clustering similarity matrix based on the abundance matrix and the spatial adjacency relationship between the pixels; the similarity matrix is subjected to Laplacian matrix normalization processing to obtain a normalized similarity matrix, the number of clusters k is determined based on the spectral gap method, and the normalized similarity matrix is subjected to unitization, and K-means clustering is performed based on the unitized row vectors; The final identification module is used for completing identification of the water body microplastics based on the pixel of the cluster center and the abundance matrix.

[0014] Compared with the prior art, the beneficial effects of the present application are: The present application effectively distinguishes the spectral characteristics of microplastics and interference substances such as algae by constructing a local water body composition endmember library and combining spectral angle distance matching, and significantly improves the accuracy of endmember identification in a complex water body environment. On this basis, each pixel in the original hyperspectral image is expressed as a linear combination of several endmembers in the water body composition endmember library, and a reconstruction cost function containing spatial smoothing and sparsity constraints is constructed to optimize and solve, accurately analyzing the spectral composition of the mixed pixels. This reconstruction mechanism effectively solves the problem of spectral mixing of interference substances such as algae, biofilms and microplastics, and significantly improves the accuracy of microplastic identification in complex natural water bodies. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 The present application is a whole method flowchart.

[0016] Figure 2 The present application is a whole system structure schematic diagram. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with specific embodiments.

[0018] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present application should be understood as the usual meaning understood by those skilled in the art to which the present application belongs. The words "first", "second" and similar words used in the present application do not represent any order, number or importance, but are only used to distinguish different components. The words "include" or "contain" and similar words mean that the elements or objects before the words cover the elements or objects listed after the words and their equivalents, and do not exclude other elements or objects. The words "connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. The words "up", "down", "left", "right" and the like only represent relative positional relationships, which may change accordingly when the absolute position of the described object changes.

[0019] Embodiment: Referring to Figure 1 The present application provides a technical solution: A water microplastic identification method based on a hyperspectral sensor, the specific steps comprising: Step 1: Based on the hyperspectral sensor, the hyperspectral data of the water body to be analyzed is obtained, the signal-to-noise ratio and the standard deviation of each band are analyzed, the meaningless bands are filtered out based on the signal-to-noise ratio and the standard deviation, and the effective band spectrum data is obtained by performing PCA analysis on the remaining bands to eliminate low-contribution bands. Further, the reflectivity mean and standard deviation of all pixels in each band are calculated, the absolute value of the reflectivity mean of all pixels is divided by the standard deviation of all pixels in the same band, the signal-to-noise ratio of the corresponding band is obtained, the signal-to-noise ratio threshold and the standard deviation threshold are preset, and the band that meets the signal-to-noise ratio less than the signal-to-noise ratio threshold and the standard deviation less than the standard deviation threshold is called meaningless band.

[0020] The signal-to-noise ratio of each band is calculated by the uniform area method, and the logic of the uniform area method for calculating the signal-to-noise ratio is that the average value of the signal is estimated as the signal intensity, the standard deviation of the signal is the intensity of the noise, and the average value is divided by the noise to obtain the signal-to-noise ratio. The absolute value of the reflectivity mean of all pixels is divided by the standard deviation of all pixels in the same band to obtain the signal-to-noise ratio of the corresponding band. The setting method of the signal-to-noise ratio threshold is to count the signal-to-noise ratios of all bands, calculate the average value of the signal-to-noise ratios of all bands, and set the average value of 20% as the signal-to-noise ratio threshold.

[0021] The low signal-to-noise ratio band with a signal-to-noise ratio less than the signal-to-noise ratio threshold is usually located at the response edge of the sensor or the strong atmospheric absorption band, and its signal is severely contaminated by noise, so the data is unreliable, and retaining them will introduce a large amount of error and interfere with the accuracy of subsequent key steps such as endmember matching, spectral unmixing and clustering analysis.

[0022] The setting method of the standard deviation threshold is to count the standard deviations of all bands, calculate the average value of the smallest 10% of the standard deviations among all standard deviations, and compare it with 5% of the maximum standard deviation, and set the smaller one as the standard deviation threshold.

[0023] The standard deviation of a band is small, which means that the reflectivity values of all pixels in the band are very close and change little. Such a band is like a "constant" and cannot provide effective information to distinguish different substances. In hyperspectral identification, we rely on the difference in reflectivity of different substances at different bands (i.e. spectral characteristics) to distinguish them. A band with no change has zero information entropy and is ineffective for classification and identification, and belongs to redundant information that needs to be filtered out.

[0024] Furthermore, the hyperspectral data of the remaining bands are used to construct a hyperspectral matrix. The hyperspectral matrix is ​​then subjected to min-max normalization, and the covariance matrix of the min-max normalized hyperspectral matrix is ​​calculated. Eigenvalues ​​and corresponding eigenvectors of the covariance matrix are extracted. The min-max normalized eigenvalues ​​are then sorted in descending order, and their cumulative contribution rate is used as the eigenvalues. With a constraint of greater than 95%, the first I eigenvalues ​​that meet the constraint are taken as principal components. The component retention coefficients of the remaining bands are determined based on the components of each band on the principal components. The component retention coefficients of each band are sorted from largest to smallest. The first 10% of the bands are called effective bands, and the remaining bands are called low contribution bands. Low contribution bands are then removed.

[0025] The formula is: ; ; in, For the front The cumulative contribution rate of each principal component Let be the component retention coefficient of the j-th residual band. Let be the eigenvalue corresponding to the i-th principal component, where i is the index of the principal component and j is the index of the residual band. The sum of all eigenvalues, This represents the value of the j-th residual band in the eigenvector corresponding to the i-th principal component. In PCA, each eigenvalue represents the amount of variance explained by its corresponding principal component (eigenvector). The principal component corresponding to the largest eigenvalue explains the largest variance in the data, which usually means that this principal component contains the most important information in the data. The cumulative contribution rate is determined by stacking the sorted eigenvalues ​​to determine how much information the stacked eigenvalues ​​retain, expressed as the cumulative contribution rate. With a constraint of greater than 95%, principal component selection can remove a large number of components with little information while retaining most of the information. For each principal component's corresponding eigenvector, the components reflecting the eigenvectors under each band... The larger the absolute value, the more sensitive the corresponding spectral band is to spectral differences. This represents the contribution of the corresponding principal component. Using the contribution of each principal component as a weight, the sensitivity of a certain band to the spectral differences under each principal component is weighted to obtain the comprehensive sensitivity of the corresponding band to the spectral differences among all principal components. The larger the value, the greater the overall sensitivity of the corresponding band to the spectral differences among all principal components. This is also beneficial for subsequent analysis of the degree of spectral differences, and its importance is also higher. Therefore, bands with larger retention coefficients are retained.

[0026] PCA can reduce the dimension of the data by condensing the information of the original numerous bands into a few principal components. Eliminating those bands with low contribution rate in the principal components is equivalent to removing redundant information. This can greatly reduce the amount of data needed to be processed in subsequent steps (such as endmember matching, spectral unmixing, clustering), thereby greatly improving the running speed and computational efficiency of the entire algorithm. There is usually a high degree of correlation (i.e. multicollinearity) between adjacent bands of spectral data. This collinearity can interfere with the stability of the optimization algorithm and clustering. The principal components generated by PCA are orthogonal (uncorrelated) to each other. By selecting components with high contribution rate and bands sensitive to spectral differences, we indirectly reduce the redundant correlation within the data set by retaining the original bands corresponding to the principal components that constitute the main information. This makes the numerical calculation more stable and the solution more reliable when optimizing the reconstruction cost function in subsequent steps, avoiding the problem of non-convergence or excessive error caused by ill-conditioned data. The final identification relies on spectral clustering and K-means clustering. If the input clustering data contains a large number of low-contribution bands that are not discriminative, these "noise" dimensions will dilute the truly effective spectral features, leading to "dimension disaster" and making it difficult for the clustering algorithm to accurately capture meaningful class structures. By filtering effective bands through PCA, we provide the clustering algorithm with purer, more compact data, which helps to form clearer inter-class separation and tighter intra-class aggregation, resulting in more accurate and stable microplastic pixel clustering results.

[0027] Step 2: Obtain the local water component endmember library, match each pixel in the effective band spectral data with each endmember in the water component endmember library, and construct a two-dimensional endmember matrix with the pixels that match successfully; expand the effective band spectral data into a two-dimensional hyperspectral matrix; Number each pixel in the pixel space in order from top to bottom and left to right, use the number of each pixel as the index of the row of the two-dimensional hyperspectral matrix, and number the effective bands from small to large as the index of the column of the two-dimensional hyperspectral matrix to form a two-dimensional hyperspectral matrix.

[0028] Construct a row vector for each pixel that matches the endmember according to the order of the effective bands from small to large, sort the pixels that match the endmember according to the order of the numbers from small to large, and group the row vectors of the pixels that match the endmember according to the sorting order to form a matrix, which is the two-dimensional endmember matrix; Furthermore, the water component end-member library contains complete spectral data of various pure substances in the local water body. The logic for matching each pixel in the effective band spectral data with each end-member in the water component end-member library is as follows: calculate the spectral angular distance between the pixel and each effective band of each end-member in the water component end-member library. If the minimum spectral angular distance is less than the spectral angular distance threshold of the corresponding end-member, the match is successful. The spectral angular distance threshold for each substance can be obtained by consulting relevant literature for the corresponding substance, or it can be provided by inviting experts in the field to demonstrate the properties of the corresponding substance. This is a conventional technical method in the field and will not be elaborated here. The various pure substances include, but are not limited to, water, various microplastics, and various algae.

[0029] Step 3: Reconstruct the two-dimensional hyperspectral matrix based on the abundance matrix and the two-dimensional endmember matrix, construct the reconstruction cost function, and optimize the reconstruction cost function to obtain the abundance matrix with the minimum reconstruction cost. Furthermore, for ease of computation, the two-dimensional pixel coordinates are flattened into one dimension. The flattened pixel coordinate information and pixel spectral information are then used to construct a two-dimensional hyperspectral matrix, where the two-dimensional pixel coordinate space is the pixel space. Each pixel in the two-dimensional hyperspectral matrix is ​​reconstructed as a superposition of its endmembers using an abundance matrix and a two-dimensional endmember matrix. The reconstruction cost function is constructed as follows: ; in, To reconstruct the cost function, It is a two-dimensional hyperspectral matrix, specifically a B-row, N-column matrix, where B represents the total number of pixels in the effective band spectral data, and N represents the total number of effective bands in the effective band spectral data. It is a two-dimensional endmember matrix, specifically an M-row N-column matrix. The abundance matrix is ​​specifically a B-row, M-column matrix, where M is the number of successfully matched endmembers. For the first pixel space The abundance vector of each cell in the abundance matrix The fourth neighbor in the 4th neighboring neighborhood The abundance vector of each cell in the abundance matrix, where... For the first A 4-neighbor neighborhood of 1 pixel The sparsity parameter is typically set between 0.001 and 0.1. This is a spatial smoothing parameter, typically taken as 0.01–1. , For the index of a cell in the cell space; for The square of the Frobenius norm, for L1 norm, For And The square of the Euclidean distance; To reconstruct the original spectral matrix by representing each pixel as a linear combination of endmembers superimposed by each endmember, is the "reconstruction error sum of squares" of the entire data set; for reflecting the accuracy of the linear combination of endmembers to reconstruct the original spectral matrix, the smaller the error, the more correct the unmixed model can explain the observed spectrum. is the sparsity constraint on the abundance matrix, which is used to encourage each pixel to be composed of only a few endmembers, and to reduce the complexity of mixing as much as possible. is the abundance difference between the pixel and its adjacent pixels in the abundance space (abundance matrix), which is used to represent the overall smoothness of the adjacent pixel abundance in the entire image, to ensure that the mixing ratio of spatially adjacent pixels does not appear "abruptly", and to avoid the unmixed result from jumping randomly between adjacent pixels due to noise. The reconstruction cost function is composed of the above three parts (reconstruction error part, sparsity constraint and spatial smoothing constraint), and minimizing the reconstruction cost function is to ensure that the reconstruction total error of the unmixed is smaller, the mixing complexity is lower and the reconstruction space is more smooth.

[0030] The abundance matrix reflects the mixing ratio and spatial distribution of different materials (such as plastics, algae, water, etc.) in each pixel. The abundance matrix describes the deconstruction process of "spectral mixing" in hyperspectral observation - that is, the reflection signal of each pixel is formed by superimposing several pure components in proportion. The abundance matrix not only reveals the spectral composition rule of materials, but also presents the distribution form and mutual interaction of materials in space.

[0031] In algae-covered environments such as lakes or reservoirs, pixels are often formed by mixing of microplastic particles, algal biofilms, water background and other components. Traditional classification or clustering methods cannot distinguish such "mixed spectra", resulting in too large spectral difference during subsequent clustering identification, too many clustering labels, and decreased microplastic identification rate. By constructing the abundance matrix, complex mixed pixels can be decomposed into independent contributions of each material in the spectrum, thereby achieving accurate description of the "covered" "partially mixed" or "weak signal" plastic area.

[0032] Solve the minimization The corresponding abundance matrix can use existing technologies, including but not limited to multiplicative update method and alternating least squares method.

[0033] The solution process of the multiplicative update method is: first, under the condition of fixing the end member matrix, the optimization objective of the abundance matrix is decomposed into a form that can be gradually approximated, then according to the idea of non-negative matrix decomposition, the gradient of the objective function with respect to the abundance matrix is separated into positive and negative terms, so that the update of each term is in the form of proportional correction. Specifically, the algorithm calculates the positive and negative effects of the reconstruction term and the constraint term on the abundance matrix in each iteration, respectively, and updates the two terms by ratio, so that the abundance value gradually approaches the optimal distribution while keeping non-negative. With the iteration, the reconstruction error of the pixel spectrum decreases, and the abundance matrix gradually stabilizes at the local optimal solution, so that the proportion of each end member in each pixel is obtained.

[0034] The solution process of the alternating least squares method is to divide the overall optimization problem into two analytically solvable sub-problems, and gradually approach the global optimum through alternating iteration. The algorithm first fixes the end member matrix, and solves a least squares problem with non-negative constraints for each pixel to obtain the optimal abundance vector under the current condition; then the end member spectrum is corrected in the opposite direction under the condition of fixing the abundance matrix, so that the new end member combination can better reconstruct the original data. The above two steps are alternately executed, and each iteration ensures that the objective function monotonically decreases until convergence.

[0035] Step 4: Extract the eigenvalues of the abundance matrix, determine the number of clusters k based on the spectral gap method based on the eigenvalues of the abundance matrix, and perform K-means clustering based on the row vectors of the abundance matrix; Further, the logic for determining the number of clusters k based on the spectral gap method is: arrange the eigenvalues of the abundance matrix in ascending order, and the arrangement result is: ; wherein, is the tth eigenvalue of the abundance matrix, is the index of the eigenvalue of the abundance matrix, T is the total number of eigenvalues of the abundance matrix, and is selected as the number of clusters k when is the maximum.

[0036] Further, randomly select k row vectors in the row vector matrix of the abundance matrix as cluster centers, calculate the Euclidean distance between each row vector of the abundance matrix and each cluster center, and assign each row vector of the abundance matrix to the cluster corresponding to the nearest cluster center. Update the cluster center position with the mean of the row vectors in each cluster, iterate until a preset number of iterations is reached, output the clustering result, and the clustering result includes the pixel coordinate set of each class label.

[0037] Step 5: Complete the identification of water microplastics based on the clustering result and the abundance matrix.

[0038] For the same class label under the pixel, the abundance matrix component of each pixel matched with the endmember is calculated, which is called endmember abundance component, the sum of all pixels in the same endmember abundance component under the same class label is calculated, and the sum of all endmember abundance components of all pixels is calculated, and the ratio of the sum of all pixels in the same endmember abundance component of the same kind of microplastic to the sum of all endmember abundance components of all pixels is obtained.

[0039] Please refer to Figure 2 The application further provides a water body microplastic identification system based on a hyperspectral sensor, which is used to implement the water body microplastic identification method based on the hyperspectral sensor, and specifically comprises: The component extraction module is used to obtain hyperspectral data of the water body to be analyzed based on the hyperspectral sensor, analyze the hyperspectral data to obtain the signal-to-noise ratio and the standard deviation of each wave band, filter out meaningless wave bands based on the signal-to-noise ratio and the standard deviation, perform PCA analysis on the remaining wave bands to eliminate low-contribution wave bands, and obtain effective wave band spectral data. The endmember matching module is used to obtain a local water body component endmember library, match each pixel in the effective wave band spectral data with each endmember in the water body component endmember library, and construct a two-dimensional endmember matrix of the matched pixels; and the effective wave band spectral data is unfolded into a two-dimensional hyperspectral matrix. The reconstruction optimization module is used to reconstruct the two-dimensional hyperspectral matrix based on the abundance matrix and the two-dimensional endmember matrix, construct a reconstruction cost function, and optimize and solve the reconstruction cost function to obtain the abundance matrix when the reconstruction cost is the smallest. The similar clustering module is used to determine a spectral clustering similarity matrix based on the spatial adjacency relationship between the abundance matrix and the pixels, perform Laplacian matrix normalization processing on the similar matrix to obtain a normalized similar matrix, determine the clustering number k based on the spectral gap method, unitize the normalized similar matrix, and perform K-means clustering based on the unitized row vectors. The final identification module is used to complete the identification of the water body microplastic based on the pixel and the abundance matrix of the clustering center.

[0040] The above formulas are all dimensionless values, and the formulas are obtained by software simulation of a large amount of data to obtain a formula closest to the actual situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0041] The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, the above-described embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be realized by electronic hardware, or by a combination of computer software and electronic hardware. Whether the functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solutions.

[0042] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, and can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiments according to actual needs.

[0043] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art cannot easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for identifying microplastics in water based on a hyperspectral sensor, characterized in that, The specific steps include: Step 1: Acquire hyperspectral data of the water body to be analyzed based on a hyperspectral sensor, analyze the hyperspectral data to obtain the signal-to-noise ratio and standard deviation of each band, filter out meaningless bands based on the signal-to-noise ratio and standard deviation, perform PCA analysis on the remaining bands to remove low-contribution bands, and obtain effective band spectral data. Step 2: Obtain the local water composition end-member library, match each pixel in the effective band spectral data with each end-member in the water composition end-member library, and construct a two-dimensional end-member matrix from the successfully matched pixels; expand the effective band spectral data into a two-dimensional hyperspectral matrix; Step 3: Reconstruct the two-dimensional hyperspectral matrix based on the abundance matrix and the two-dimensional endmember matrix, construct the reconstruction cost function, and optimize the reconstruction cost function to obtain the abundance matrix with the minimum reconstruction cost. Step 4: Extract the eigenvalues ​​of the abundance matrix. Based on the eigenvalues ​​of the abundance matrix, use the spectral gap method to determine the number of clusters k. Perform K-means clustering based on the row vectors of the abundance matrix. Step 5: Identify microplastics in water bodies based on clustering results and abundance matrix.

2. The method for identifying microplastics in water based on a hyperspectral sensor according to claim 1, characterized in that: Calculate the mean and standard deviation of reflectance for all pixels in each band. Divide the absolute value of the mean reflectance of all pixels by the standard deviation of all pixels in the same band to obtain the signal-to-noise ratio (SNR) of the corresponding band. Set a preset SNR threshold and a standard deviation threshold. Bands that simultaneously satisfy the conditions of SNR less than the SNR threshold and standard deviation less than the standard deviation threshold are called meaningless bands.

3. The method for identifying microplastics in water based on a hyperspectral sensor according to claim 1, characterized in that: The remaining hyperspectral data are used to construct a hyperspectral matrix. This hyperspectral matrix is ​​then standardized, and the covariance matrix of the standardized hyperspectral matrix is ​​calculated. Eigenvalues ​​and corresponding eigenvectors of the covariance matrix are extracted. The standardized eigenvalues ​​are then sorted in descending order of their cumulative contribution rate. The constraint condition is greater than 95%. The first I eigenvalues ​​that meet the constraint condition for the first time are taken as principal components. The component retention coefficient of the remaining bands is determined based on the component of each band on the principal components. The component retention coefficient of each band is sorted from large to small. The first 10% of the bands are called effective bands, and the remaining bands are called low contribution bands. Low contribution bands are removed. The formula is: ； ； in, For the front The cumulative contribution rate of each principal component Let be the component retention coefficient of the j-th residual band. Let be the eigenvalue corresponding to the i-th principal component, where i is the index of the principal component and j is the index of the residual band. The sum of all eigenvalues, This represents the value of the j-th residual band in the eigenvector corresponding to the i-th principal component. The water composition end-member library contains complete spectral data of various pure substances in local water bodies. The logic for matching each pixel in the effective band spectral data with each end-member in the water composition end-member library is as follows: calculate the spectral angular distance between the pixel and each effective band of each end-member in the water composition end-member library. If the minimum spectral angular distance is less than the spectral angular distance threshold of the corresponding end-member, the match is successful.

4. The method for identifying microplastics in water based on a hyperspectral sensor according to claim 3, characterized in that: The two-dimensional pixel coordinates are flattened into one dimension. The flattened pixel coordinate information and pixel spectral information are used to construct a two-dimensional hyperspectral matrix, where the two-dimensional pixel coordinate space is the pixel space. Each pixel in the two-dimensional hyperspectral matrix is ​​reconstructed as a superposition of its endmembers using an abundance matrix and a two-dimensional endmember matrix. The reconstruction cost function is then constructed as follows: ； in, To reconstruct the cost function, It is a two-dimensional hyperspectral matrix, specifically a B-row, N-column matrix, where B represents the total number of pixels in the effective band spectral data, and N represents the total number of effective bands in the effective band spectral data. It is a two-dimensional endmember matrix, specifically an M-row N-column matrix. The abundance matrix is ​​specifically a B-row, M-column matrix, where M is the number of successfully matched endmembers. For the first pixel space The abundance vector of each cell in the abundance matrix The fourth neighbor in the 4th neighboring neighborhood The abundance vector of each cell in the abundance matrix, where... For the first A 4-neighbor neighborhood of 1 pixel Sparse parameters, For spatial smoothing parameters, , For the index of a cell in the cell space; for The square of the Frobenius norm, for L1 norm, for and The square of the Euclidean distance; Solve for minimization The abundance matrix corresponding to the time.

5. The method for identifying microplastics in water based on a hyperspectral sensor according to claim 4, characterized in that: The logic for determining the cluster number k based on the spectral gap method is as follows: Arrange the eigenvalues ​​of the abundance matrix in ascending order, and the arrangement result is: ;in, Let be the t-th eigenvalue of the abundance matrix. Here, T represents the index of the eigenvalues ​​of the abundance matrix, and T is the total number of eigenvalues ​​in the abundance matrix. Maximum k is the cluster number.

6. The method for identifying microplastics in water based on a hyperspectral sensor according to claim 5, characterized in that: In the row vector matrix of the abundance matrix, k row vectors are randomly selected as cluster centers. The Euclidean distance between each row vector of the abundance matrix and each cluster center is calculated. Each row vector of the abundance matrix is ​​assigned to the cluster corresponding to its nearest cluster center. The cluster center position is updated with the mean of the row vectors in each cluster. The process is iterated until the preset number of iterations is reached. The clustering result is output, which includes the pixel coordinate set of each category label.

7. The method for identifying microplastics in water based on a hyperspectral sensor according to claim 6, characterized in that: For pixels under the same category label, calculate the abundance matrix component of each pixel that successfully matches an endmember in each abundance matrix. This component is called the endmember abundance component. Calculate the sum of the endmember abundance components of all pixels under the same category label, and then calculate the sum of all endmember abundance components of all pixels. Divide the sum of the endmember abundance components of all pixels for the same microplastic by the sum of all endmember abundance components of all pixels to obtain the proportion of that microplastic in all pixels under the corresponding category label.

8. A water microplastic identification system based on a hyperspectral sensor, characterized in that: The system is used to implement the water microplastic identification method based on a hyperspectral sensor as described in any one of claims 1-7, specifically including: The component extraction module is used to acquire hyperspectral data of the water body to be analyzed based on a hyperspectral sensor, analyze the hyperspectral data to obtain the signal-to-noise ratio and standard deviation of each band, filter out meaningless bands based on the signal-to-noise ratio and standard deviation, perform PCA analysis on the remaining bands to remove low-contribution bands, and obtain effective band spectral data. The endmember matching module is used to obtain the local water composition endmember library, match each pixel in the effective band spectral data with each endmember in the water composition endmember library, and construct a two-dimensional endmember matrix from the successfully matched pixels; and expand the effective band spectral data into a two-dimensional hyperspectral matrix. The reconstruction optimization module is used to reconstruct the two-dimensional hyperspectral matrix based on the abundance matrix and the two-dimensional endmember matrix, construct the reconstruction cost function, and optimize the reconstruction cost function to obtain the abundance matrix with the minimum reconstruction cost. The similarity clustering module is used to determine the spectral clustering similarity matrix based on the abundance matrix and the spatial adjacency relationship between pixels; the similarity matrix is ​​normalized by Laplacian matrix to obtain the normalized similarity matrix; the number of clusters k is determined based on the spectral gap method; the normalized similarity matrix is ​​normalized; and K-means clustering is performed based on the normalized row vectors. The final identification module is used to identify microplastics in water bodies based on the pixels of the cluster centers and the abundance matrix.

Images