A multi-component phytoplankton concentration measurement method based on fluorescence spectrum layering and zoning

Abstract

The application discloses a kind of multicomponent plankton algae concentration measurement method based on fluorescence spectrum stratification partitioning.The application belongs to the field of marine aquatic ecological environment monitoring technology.The algorithm will unknown to be analyzed mixed three-dimensional fluorescence spectrum according to the following stratification partitioning analysis idea to be analyzed.First layer: full spectrum is involved in analysis, and the purpose is to analyze cyanophyta concentration and cryptophyta concentration, and obtain the difference spectrum in full spectrum region.Second layer: partitioning analysis, take the region containing effective spectral feature to participate in analysis, and the purpose is to analyze chlorophyll a concentration of green algae, diatom, dinoflagellate and yellow algae.The algorithm is significantly superior to common analysis algorithm in reducing misidentification, especially in processing such as diatom, dinoflagellate and yellow algae with high similarity living body fluorescence spectrum.

Description

Technical Field

[0001] This invention belongs to the field of marine aquatic ecological environment monitoring technology, specifically relating to a method for measuring the concentration of multi-component phytoplankton based on fluorescence spectral stratification and partitioning. Background Technology

[0002] Phytoplankton (Algae) refers to a collective term for a group of tiny algae that live in water bodies by floating on the surface. Algae are widely distributed and are the material and energy basis of aquatic ecosystems. Their species composition, community structure, and abundance changes directly affect water quality, energy flow, material flow, and changes in biological resources within the ecosystem.

[0003] Vitro fluorescence spectroscopy offers advantages such as rapid measurement speed, high sensitivity, and no need for sample pretreatment, making it the most promising method for rapid on-site measurement of phytoplankton concentration in large-area water bodies. The basic idea of ​​vitro fluorescence spectroscopy is to invert the chlorophyll a concentration of algae by utilizing the fluorescence intensity generated by stimulated emission of pigment molecules in living algal cells. Different species of living algae possess their own characteristic fingerprint fluorescence spectra. When multiple components are present in a sample, within the range of Lambert-Beer's law, the algal fluorescence spectra satisfy linear additive property. Therefore, the problem of quantifying algal concentration can be transformed into a linear regression problem of a matrix. This is the fundamental principle behind the rapid classification and quantitative measurement of phytoplankton using vitro fluorescence spectroscopy.

[0004] Visible fluorescence, a byproduct of phytoplankton photosynthesis, is closely related to the efficiency of light energy absorption, transfer, and release during photosynthesis in living cells. The instability of phytoplankton vital fluorescence negatively impacts the accurate quantification of chlorophyll a concentration in phytoplankton vital fluorescence. Constructing a standardized spectral library for phytoplankton through interpolation can significantly improve the classification accuracy and measurement precision of multivariate regression analysis results. However, this also introduces new problems: the standardized spectral library constructed by interpolation has a certain distribution range, leading to significant overlap and excessive similarity among standardized spectral libraries of various phytoplankton phyla—especially those with highly similar fluorescence spectral characteristics—resulting in difficulties in analysis. Summary of the Invention

[0005] To address the technical problem mentioned in the background art—the large overlap and high similarity of the standard spectral libraries of planktonic algae with highly similar fluorescence spectral characteristics, leading to difficulties in analysis—this invention proposes a multi-component phytoplankton concentration measurement method based on hierarchical partitioning of fluorescence spectra.

[0006] The technical solution of the present invention is as follows:

[0007] A method for measuring the concentration of multi-component planktonic algae based on fluorescence spectral hierarchical partitioning includes the following steps:

[0008] Step 1: Perform non-negative least squares multiple linear regression analysis on the mixed in vivo three-dimensional fluorescence spectra of the unknown algae samples across the full spectrum for both cyanobacteria and cryptophytes.

[0009] ,

[0010] in: This indicates the three-dimensional fluorescence spectrum of living phytoplankton with an excitation wavelength range of 370 nm to 650 nm and an emission wavelength range of 600 nm to 720 nm. This represents the average spectrum of the standard spectral library of planktonic algae, excluding cyanobacteria and cryptophytes. These represent the standard spectrum combinations with the smallest errors obtained after traversing all combinations in the standard spectral libraries of phytoplankton other phyla besides cyanobacteria and cryptophytes. The least-squares analytical concentrations of phytoplankton from the phylum Cyanobacteria and Cryptophyta are respectively. Indicates the minimum iteration error;

[0011] Step 2: Obtain the difference fluorescence spectrum across the full spectrum after subtracting cyanobacteria and cryptophytes. :

[0012] ,

[0013] Step 3: Select a new characteristic fluorescence spectral range within the differential fluorescence spectral range. , The fluorescence characteristics of most algae, diatoms, dinoflagellates and xanthophytes are covered, and the excitation wavelength is set in the range of 580 nm to 620 nm and the emission wavelength is set in the range of 660 nm to 700 nm.

[0014] Step 4: Within the same range as in Step 3, extract the standard spectra of Chlorophyta, Diatoms, Dinophyta, and Xanthophyta to construct a new standard spectral library. ;

[0015] Step 5: Within the same scope as in Step 3, perform non-negative least squares multiple linear regression analysis on the phyla Chlorophyta, diatoms, dinoflagellates, and xanthophytes:

[0016] ,

[0017] in: These represent the standard spectral combinations with the smallest errors obtained after traversing all combinations in the standard spectral libraries of Chlorophyta, Diatoms, Dinophyta, and Xanthophyta within the same step three range; The least-squares analytical concentrations of the phyla Chlorophyta, diatoms, dinoflagellates, and xanthophyta are respectively represented. This represents the minimum iteration error.

[0018] Step Six: Output the chlorophyll a concentration of different phyla of planktonic algae, i.e.:

[0019] .

[0020] Beneficial effects:

[0021] This invention discloses a method for measuring the concentration of multi-component planktonic algae based on a hierarchical partitioning approach to fluorescence spectroscopy. During analysis, the unknown, aliased three-dimensional fluorescence spectrum to be analyzed is processed according to the following hierarchical partitioning approach: First layer: Full spectrum. The purpose of the analysis was to determine the concentrations of cyanobacteria and cryptophytes, and to obtain the difference spectrum across the entire spectral region. Second layer: Partition resolution, taking... Regions containing effective spectral features The purpose of this algorithm is to analyze the chlorophyll a concentration of phyla Chlorophyta, Diatoms, Dinophyta, and Xanthophyta. This algorithm significantly outperforms common analytical algorithms in reducing false identifications, especially when dealing with highly similar in vivo fluorescence spectra of phyla such as Diatoms, Dinophyta, and Xanthophyta. Attached Figure Description

[0022] Figure 1 A schematic diagram showing the overlap range of the standard spectral library for phytoplankton (especially diatoms and dinoflagellates).

[0023] Figure 2 This is a schematic diagram of the three-dimensional fluorescence spectrum similarity of living planktonic algae (especially green algae, diatoms, yellow algae, and dinoflagellates).

[0024] Figure 3 This is a schematic diagram of the hierarchical partitioning analysis of aliased spectra.

[0025] Figure 4 This diagram illustrates the comparison of the average relative error of different spectral analysis methods for measuring algal samples in the test collection. Detailed Implementation

[0026] The specific embodiments of the present invention will be described in detail below. This embodiment is implemented based on the technical solution of the present invention, and provides detailed implementation methods and specific operation processes. However, the protection scope of the present invention is not limited to the following embodiments.

[0027] Example 1

[0028] like Figure 1As shown, the standardized spectral libraries constructed using interpolation methods have a certain distribution range, resulting in large overlap and excessive similarity among standardized spectral libraries of various phytoplankton phyla—especially those with highly similar fluorescence spectral characteristics—leading to difficulties in analysis. For example... Figure 2 As shown, green algae, diatoms, and dinoflagellates share similar characteristic spectral regions and similar noise regions, with the characteristic spectral regions accounting for only a small percentage of the full spectrum, approximately 30-50%. Therefore, when substituting into the full spectrum analysis, nearly half or more of the noise region will reduce the accuracy of identification and quantification of these algal phyla. However, directly narrowing the identification spectral region may make it difficult to cover the characteristic fluorescence spectral regions of cyanobacteria and cryptophytes.

[0029] To address the above issues, the hierarchical partitioning parsing algorithm in this embodiment follows these steps:

[0030] 1. Establish a standardized spectral library of interpolated spectral data for different phyla of phytoplankton based on non-uniform interpolation, following the steps below:

[0031] Step 1: Obtain the fluorescence spectrum of the phytoplankton in vivo, with an excitation wavelength range of 370 nm to 650 nm and an emission wavelength range of 600 nm to 720 nm. Measure the standard chlorophyll a concentration of the phytoplankton in vivo.

[0032] Step 2: Divide the fluorescence spectrum of the phytoplankton obtained in Step 1 by the standard chlorophyll a concentration of the phytoplankton to obtain the concentration-normalized spectrum of the phytoplankton, i.e., the standard spectrum.

[0033] Step 3: Calculate the mean fluorescence spectrum M of the phytoplankton and the corresponding relative standard deviation S based on the standard spectrum described in Step 2;

[0034] Step 4: Calculate M±2 S, obtains the upper and lower limits of the interpolation range of the standard spectral library of phytoplankton;

[0035] Step 5: Perform uniform interpolation within the interpolation range of the phytoplankton standard spectral library mentioned in Step 5, with the number of interpolations being 10.

[0036] Step 6: Repeat steps 1-5 to obtain standardized spectral libraries of different phyla of phytoplankton;

[0037] 2. Obtain the three-dimensional fluorescence spectrum of the unknown phytoplankton sample in vivo, with the spectral range the same as in step 1 of section 1;

[0038] 3. This invention analyzes the spectrum based on the idea of ​​least squares multiple linear regression analysis.

[0039] The idea behind multiple linear regression analysis is briefly described below: For a certain type of algae, its three-dimensional fluorescence spectrum measurement values ​​are a fluorescence intensity data matrix, theoretically:

[0040] ,

[0041] Among them, M i For the in vivo three-dimensional fluorescence measurement spectrum of a certain type of algae sample, a i This indicates that the concentration of the algae sample is related to the concentration of that algae. i This indicates the standard fluorescence spectrum of algal samples of this phylum.

[0042] When multiple components are present in a sample, within the range of Beer-Lambert law, the fluorescence spectra of algae satisfy linear additive property, that is:

[0043] ,

[0044] ,

[0045] Where: M represents the three-dimensional fluorescence measurement spectrum of the mixed in vivo unknown algal sample; f ij The j-th standard spectrum in the standard spectral library of the i-th phylum of planktonic algae; a i Let χ be the chlorophyll concentration of the i-th phylum, and λ be the error. mn It represents the fluorescence intensity of the i-th algae at the excitation wavelength n and the emission wavelength m.

[0046] To minimize error, a non-negative least squares method is used to analyze the measured spectrum. Furthermore, since the concentration of algae in the mixed algae sample will not be negative, a non-negativity constraint needs to be added to the iterative calculations, and the least squares result is output.

[0047] ,

[0048] 4. During analysis, the unknown, aliased three-dimensional fluorescence spectra to be analyzed are analyzed according to the following hierarchical and partitioned analysis approach, such as... Figure 3 Shown. First layer: Full spectrum The purpose of the analysis was to determine the concentrations of cyanobacteria and cryptophytes, and to obtain the difference spectrum across the entire spectral region. Second layer: Partition resolution, taking... Regions containing effective spectral features The purpose of this analysis is to determine the chlorophyll a concentration of green algae, diatoms, dinoflagellates, and xanthophytes.

[0049] 4.1 First Layer: Optimization and analysis of interpolated spectral libraries of cyanobacteria and cryptophytes across the entire spectral range

[0050] Step 1: Perform non-negative least squares multiple linear regression analysis on the mixed in vivo three-dimensional fluorescence spectra of the unknown algae sample across the full spectrum for both cyanobacteria and cryptophytes. The full spectrum is the spectral range described in Step 1 of section 1.

[0051] ,

[0052] in: This indicates the three-dimensional fluorescence spectrum of living phytoplankton with an excitation wavelength range of 370 nm to 650 nm and an emission wavelength range of 600 nm to 720 nm. This represents the average spectrum of the standard spectral library of planktonic algae, excluding cyanobacteria and cryptophytes. These represent the standard spectrum combinations with the smallest errors obtained after traversing all combinations in the standard spectral libraries of phytoplankton other phyla besides cyanobacteria and cryptophytes. The least-squares analytical concentrations of phytoplankton from the phylum Cyanobacteria and Cryptophyta are respectively. This represents the minimum iteration error.

[0053] Step 2: Obtain the difference fluorescence spectrum across the full spectrum after subtracting cyanobacteria and cryptophytes. :

[0054] .

[0055] Step 3: Select a new characteristic fluorescence spectral range within the differential fluorescence spectral range. , The fluorescence characteristics of most algae, diatoms, dinoflagellates and xanthophytes are covered, and the excitation wavelength is set to the range of 580 nm to 620 nm and the emission wavelength to the range of 660 nm to 700 nm.

[0056] 4.2 Second layer: Optimization analysis of interpolated spectral libraries for local features of green algae, diatoms, dinoflagellates and xanthophytes.

[0057] Step 4: Within the same range as in Step 3, extract the standard spectra of Chlorophyta, Diatoms, Dinophyta, and Xanthophyta to construct a new standard spectral library. .

[0058] Step 5: Within the same scope as in Step 3, perform non-negative least squares multiple linear regression analysis on the phyla Chlorophyta, diatoms, dinoflagellates, and xanthophytes:

[0059] ,

[0060] in: These represent the standard spectral combinations with the smallest errors obtained after traversing all combinations in the standard spectral libraries of Chlorophyta, Diatoms, Dinophyta, and Xanthophyta within the same step three range; The least-squares analytical concentrations of the phyla Chlorophyta, diatoms, dinoflagellates, and xanthophyta are respectively represented. This represents the minimum iteration error.

[0061] Step Six: Output the chlorophyll a concentration of different phyla of planktonic algae, i.e.:

[0062] ,

[0063] Example 2: Three-dimensional fluorescence spectra of phytoplankton in vivo under different habitat conditions

[0064] Common phytoplankton species from my country's coastal and freshwater waters were selected, including cyanobacteria, green algae, diatoms, dinoflagellates, xanthophytes, and cryptophytes (see Table 1). They were cultured according to GB / T21805-2008 "Chemicals - Algal Growth Inhibition Test".

[0065] Table 1

[0066]

[0067] Appropriate amounts of algal sample stock solution were taken at different growth stages, diluted according to a certain volume ratio, and the three-dimensional fluorescence spectrum of the live samples was measured using a fluorescence spectrophotometer. The excitation wavelength range was 370 nm to 650 nm, and the emission wavelength range was 600 nm to 720 nm. At the same time, the chlorophyll a standard concentration of the algal samples was measured according to the national standard method.

[0068] Example 3: Obtaining a spectral library of interpolated spectra of different phytoplankton

[0069] The algal samples with known chlorophyll a standard concentration obtained in Example 2 were divided into experimental and test sets in a 7:3 ratio.

[0070] By dividing the three-dimensional fluorescence spectra of the live phytoplankton in the experimental set by their respective chlorophyll a standard concentrations, a series of standard three-dimensional fluorescence spectra of the live phytoplankton were obtained. Based on the standard spectra of different phyla, the mean fluorescence spectrum M and the corresponding relative standard deviation S of the phytoplankton were calculated; M ± 2 was then calculated. S, obtain the upper and lower limits of the interpolation range of the standard spectral library of the phytoplankton; perform uniform interpolation within the interpolation range of the standard spectral library of the phytoplankton in step 4, and the number of interpolations is 10, thereby establishing the standard spectral library of the phytoplankton.

[0071] The methods described above yield standardized spectral libraries of planktonic algae from specific phyla such as cyanobacteria, green algae, diatoms, dinoflagellates, xanthophytes, cryptophytes, and others.

[0072] Example 4: Comparison of Layered Partitioning Resolution Effects

[0073] The analytical method described in this invention was used to analyze the live fluorescence spectra of the phytoplankton in the test set obtained in Example 2. The analytical results were then compared with the direct analytical results obtained by the least squares multiple linear regression algorithm (i.e., a common analytical algorithm) using the standard concentration of chlorophyll a of the samples and without stratification. The test set samples included 209 pure algal samples, of which 44 were cyanobacteria, 60 were green algae, 85 were diatoms, 60 were dinoflagellates, 46 were xanthophytes, and 14 were cryptophytes, along with 20 mixed samples. The false identification rate was used as the basis for the analysis. and measurement relative error This indicates the analytical precision of the algorithm.

[0074] ,

[0075] ,

[0076] in: and These represent the total number of samples and the number of misidentified samples for a specific phylum of phytoplankton, respectively. and These represent the standard concentration and the elucidated concentration of chlorophyll a in a sample of a specific phylum of planktonic algae, respectively.

[0077] Table 2

[0078]

[0079] Table 2 and Figure 4 The results show that the common parsing algorithm misidentified 34 algae in the test set, accounting for 11.0% of the total samples. The misidentification numbers for diatoms, dinoflagellates, and xanthophytes were 14, 10, and 9, respectively, with misidentification rates of 16.5%, 16.7%, and 64.3%. The misidentification rate for green algae and cryptophytes was 0%. In contrast, the hierarchical partitioning parsing algorithm significantly reduced the number of misidentified algae samples in the test set, with only 4 misidentified samples, accounting for 1.3% of the total samples. The misidentification numbers for cyanobacteria, diatoms, dinoflagellates, and xanthophytes decreased to 0, 0, 3, and 1, respectively, with misidentification rates decreasing by 2.3%, 16.5%, 11.7%, and 57.1%. (Table 2 and...) Figure 4 This indicates that the hierarchical partitioning analysis algorithm significantly outperforms common analysis algorithms in reducing false identifications, especially when dealing with highly similar live fluorescence spectra of diatoms, dinoflagellates, and xanthophytes.

[0080] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for measuring the concentration of multi-component planktonic algae based on fluorescence spectral stratification and partitioning, characterized in that, Includes the following steps: Step 1: Perform non-negative least squares multiple linear regression analysis on the mixed in vivo three-dimensional fluorescence spectra of unknown algae samples across the full spectrum for both cyanobacteria and cryptophytes. ， in: This indicates the three-dimensional fluorescence spectrum of living phytoplankton with an excitation wavelength range of 370 nm to 650 nm and an emission wavelength range of 600 nm to 720 nm. This represents the average spectrum of the standard spectral library of planktonic algae, excluding cyanobacteria and cryptophytes. This represents the standard spectrum combination with the smallest error obtained after traversing all combinations in the standard spectral libraries of Cyanobacteria and Cryptophytes. The least-squares analytical concentrations of phytoplankton from the phylum Cyanobacteria and Cryptophyta are respectively. Indicates the minimum iteration error; Step 2: Obtain the difference fluorescence spectrum across the full spectrum after subtracting cyanobacteria and cryptophytes. : ， Step 3: Select a new characteristic fluorescence spectral range within the differential fluorescence spectral range. , The fluorescence characteristics of most algae, diatoms, dinoflagellates and xanthophytes are covered, and the excitation wavelength is set in the range of 580 nm to 620 nm and the emission wavelength is set in the range of 660 nm to 700 nm. Step 4: Within the same range as in Step 3, extract the standard spectra of Chlorophyta, Diatoms, Dinophyta, and Xanthophyta to construct a new standard spectral library. ; Step 5: Within the same scope as in Step 3, perform non-negative least squares multiple linear regression analysis on the phyla Chlorophyta, diatoms, dinoflagellates, and xanthophytes: ， in: These represent the standard spectral combinations with the smallest errors obtained after traversing all combinations in the standard spectral libraries of Chlorophyta, Diatoms, Dinophyta, and Xanthophyta within the same step three range; The least-squares analytical concentrations of the phyla Chlorophyta, diatoms, dinoflagellates, and xanthophyta are respectively represented. Indicates the minimum iteration error; Step Six: Output the chlorophyll a concentration of different phyla of planktonic algae, i.e.: ， M represents the resolution of the three-dimensional fluorescence spectra of the mixed in vivo samples of unknown algae, f i For the standard spectral library of phytoplankton of the i-th phylum, a i denoted as chlorophyll concentration of the i-th phylum, χ as error, λ as algal fluorescence intensity, m as emission wavelength, and n as excitation wavelength.

Images