A method for on-site detection of seawater carbon parameters using particulate matter polarization data

Abstract

A method for on-site detection of seawater carbon parameters using polarized light data of particulate matter includes: using polarized light technology to classify and extract features of particulate matter in seawater samples, and constructing a polarized scattering database (S1); establishing a correlation between the polarized light data of particulate matter and seawater carbon parameters through statistical methods or regression analysis, and forming an analytical formula or digital model (S2); constructing a detection algorithm based on the correlation relationship, taking into account data correspondence, scope of application, and error control (S3); executing the detection algorithm, and calculating through the analytical formula or digital model to obtain the seawater carbon parameters from the polarized light data (S4). The present invention measures the polarized light data of particulate matter in seawater, and uses the detection algorithm combined with an analytical formula of a statistical method or a digital model based on a machine learning method to realize on-site detection of seawater carbon parameters, thereby achieving rapid and accurate detection of seawater carbon parameters and providing an efficient means for marine carbon cycle research and environmental monitoring.

Description

Technical Field

[0001] The present invention relates to a seawater carbon parameter detection technology, and in particular to a seawater carbon parameter on-site detection method using particulate matter polarized light data. Background Art

[0002] The ocean carbon sink refers to the mechanisms and processes by which the ocean absorbs carbon dioxide (CO2) from the atmosphere and fixes it. It involves biological, chemical, and physical processes in the ocean that together cause carbon dioxide in seawater to dissolve and be stored in the ocean. The ocean is one of the largest carbon sinks on Earth. Through physical and biological processes on the ocean surface, carbon dioxide dissolved in seawater can be absorbed by the ocean. This process helps to mitigate the increase in carbon dioxide in the atmosphere and reduce the effects of global warming. The detection and monitoring of ocean carbon sinks is of great significance for understanding the global carbon cycle, climate change, and the stability of ecosystems.

[0003] According to Marine Ecology, particulate matter participates in the carbon cycle in seawater. From the perspective of the transfer and transformation of carbon in the ocean water layer, in the upper seawater (the euphotic zone), phytoplankton, through photosynthesis, converts dissolved carbon dioxide into organic matter, which is then passed on and consumed in turn through the grazing food chain. Non-living particulate carbon is utilized by the detritus food chain, with some sinking to the lower ocean layer, reaching the seafloor. Meanwhile, some dissolved organic carbon is utilized by microbial food chains and reenters organisms (particles). Therefore, it can be seen that carbon parameters in seawater, including total carbon, particulate carbon, dissolved carbon, organic carbon, and inorganic carbon, are all related to the particles in the seawater. Theoretically, it is possible to obtain seawater carbon parameters by measuring data such as the types and concentrations of particulate matter in seawater.

[0004] Existing methods for measuring ocean carbon content include satellite remote sensing, laboratory chemical analysis, and incineration. However, these methods have drawbacks. For example, satellite remote sensing cannot detect carbon parameters in deep seawater, and chemical analysis or incineration cannot rapidly measure them on-site. Therefore, developing a method for rapid, on-site measurement of seawater carbon parameters is crucial for advancing ocean carbon sink research.

[0005] Currently, it has been reported that detection instruments based on polarization technology can collect and analyze information on single marine particles. Based on the changes in polarization state before and after the scattering process and the characteristics of fluorescence generation, these instruments can obtain information such as particle type and concentration, thereby enabling the monitoring of marine particle concentration and type. Ocean carbon parameters, particularly particulate carbon and organic carbon, are inextricably linked to information about tiny particles in the ocean. Current technologies for on-site and online detection of marine carbon parameters still have shortcomings, primarily in terms of the accuracy and stability of on-site detection equipment, as well as its adaptability to complex marine environments.

[0006] It should be noted that the information disclosed in the above background technology section is only used to understand the background of this application, and therefore may include information that does not constitute prior art known to ordinary technicians in this field. Summary of the Invention

[0007] The main purpose of the present invention is to solve the problems existing in the above-mentioned background technology and provide a method for on-site detection of seawater carbon parameters using particulate matter polarized light data.

[0008] To achieve the above object, the present invention adopts the following technical solutions:

[0009] A method for on-site detection of seawater carbon parameters using particulate matter polarized light data comprises the following steps:

[0010] S1: Use polarized light technology to classify and extract features of seawater particles in seawater samples and build a polarized scattering database;

[0011] S2: Analyze the correlation between the polarization data of seawater particles and the seawater carbon parameters using statistical methods to obtain an analytical formula that includes the correlation between the polarization parameters of seawater particles and the seawater carbon parameters; or, using the data set of the polarization parameters of seawater particles and the corresponding seawater carbon parameters, establish a digital model that includes the correlation between the polarization parameters of seawater particles and the seawater carbon parameters;

[0012] S3: Utilizing the correlation relationship and taking into account the one-to-one correspondence between the output carbon parameter and the input seawater particulate matter polarization data, the applicable scope, and error control, a detection algorithm based on the analytical formula or the digital model is constructed for each carbon parameter;

[0013] S4: Execute the detection algorithm for the polarized light data of seawater particles detected by the instrument, wherein the seawater carbon parameter is obtained by solving the equation using the analytical formula, or input the polarized light data of the particles into the digital model and calculate the seawater carbon parameter using the digital model.

[0014] Furthermore, the statistical method includes:

[0015] Obtain polarized light data of seawater particles and related seawater carbon parameters and water quality parameters from seawater samples;

[0016] Through correlation analysis, we can identify parameters with high and low correlation between polarized light data of seawater particles and seawater carbon parameters.

[0017] Based on the correlation analysis results, for the highly correlated parameters, analytical formulas are constructed, wherein the formulas describe the relationship between the highly correlated parameters and the seawater carbon parameters;

[0018] Based on the correlation analysis results, the analytical formula is auxiliary constructed for low-correlation parameters to enhance the prediction ability of the analytical formula.

[0019] Furthermore, the construction of the digital model includes:

[0020] Obtain polarized light data of seawater particles and related seawater carbon parameters and water quality parameters from seawater samples;

[0021] Select features that have a significant impact on the prediction of seawater carbon parameters from the pre-processed seawater particle polarization data;

[0022] Using a large amount of seawater sample data from the same source, combined with selected features and known carbon parameters, a numerical model is trained to learn the mapping relationship between seawater particle polarization data and seawater carbon parameters;

[0023] Validate the trained numerical model to assess its accuracy and reliability in predicting seawater carbon parameters;

[0024] According to the results of model verification, adjust the model parameters and structure to optimize the model performance;

[0025] The final optimized digital model is saved for use in on-site testing to ensure the repeatability and consistency of the model.

[0026] Furthermore, the characteristics that have a significant impact on the prediction of seawater carbon parameters include size, shape and polarized light intensity.

[0027] Furthermore, the carbon parameter includes one or more of the following carbon content indicators: total carbon (TC), inorganic carbon (IC), non-purgeable organic carbon (NPOC), and total carbon and non-purgeable organic carbon of seawater after filtration using a filter membrane.

[0028] Furthermore, the construction of the digital model also includes:

[0029] Based on the polarized light data of seawater particles measured by the instrument, a pulse information extraction method is used to extract the scattered pulse information of single particles and generate multi-dimensional data of particles for particle feature identification; the number of particles detected at a given flow rate and within a given time is counted to reflect the relative number of particles in the seawater; among them, the single pulse characteristics of the single particle information and the total number of specific types of particles in a statistical sense are characterized to reflect the information of seawater from multiple levels; the obtained data are used for training and optimization of the digital model.

[0030] Furthermore, the instrument uses four polarized light detection channels and one fluorescence detection channel to collect one or more information of the amplitude, distribution interval, full width at half maximum, protrusion, etc. of the single particle scattered pulse.

[0031] Furthermore, the statistical method adopts linear fitting, and the concentration of different types of particles is c n , the corresponding model coefficient is a n , construct the functional relationship of Equation 1-Equation 3:

[0032] NPOC=f1(a1c1,a2c2...,a n c n ) (1)

[0033] TC=f2(a1c1,a2c2...,a n c n ) (2)

[0034] IC=f3(a1c1,a2c2...,a n c n ) (3)

[0035] The concentration of particles c is obtained by using the instrument to identify the particles n , based on a large number of real seawater carbon parameters, the coefficient a is calculated using the least squares method n .

[0036] Furthermore, a machine learning method is used to construct a digital model based on a regression model, and the regression model is a random forest regression model, a Lasso regression model or a GBDT model.

[0037] A computer-readable storage medium stores a computer program, which implements the method when executed by a processor.

[0038] The present invention has the following beneficial effects:

[0039] The present invention provides a novel on-site detection method for seawater carbon parameters. This method measures polarized light data from particulate matter in seawater and utilizes a detection algorithm combined with statistical analytical formulas or machine learning digital models to achieve on-site detection of seawater carbon parameters. This method, based on the phenomenon that particulate matter in seawater participates in the ocean carbon cycle, leverages the ability of polarized light technology to detect and classify tiny marine particles, leveraging the advantages of big data models in correlation analysis to establish a relationship between particulate matter polarized light data and seawater carbon parameters. The detection algorithm then derives seawater carbon parameters from the field-measured particulate matter polarized light data, enabling efficient and reliable on-site, online detection of the latter. This method provides strong technical support for ocean carbon cycle research and environmental monitoring.

[0040] Other beneficial effects of the embodiments of the present invention will be further described below. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1Schematic diagram of the statistical method according to an embodiment of the present invention.

[0042] Figure 2 A schematic diagram is constructed for a digital model of an embodiment of the present invention.

[0043] Figure 3 A schematic diagram illustrating a detection algorithm according to an embodiment of the present invention.

[0044] Figure 4 This is a physical picture of the polarization fluorescence instrument according to an embodiment of the present invention.

[0045] Figure 5 This is a diagram showing the classification effect of typical particles in seawater according to an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The following is a detailed description of the embodiments of the present invention. It should be emphasized that the following description is only exemplary and is not intended to limit the scope of the present invention and its application.

[0047] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0048] See Figures 1 to 3 The embodiment of the present invention provides a method for on-site detection of seawater carbon parameters using particulate matter polarization data, comprising the following steps:

[0049] S1: Use polarized light technology to classify and extract features of seawater particles (such as seaweed, microplastics, and sediment) in seawater samples and build a polarized scattering database;

[0050] S2: The correlation between the polarization data of seawater particles and the carbon parameters of seawater is analyzed by statistical methods, and the analytical formula containing the correlation between the polarization parameters of seawater particles and the carbon parameters of seawater is obtained (see Figure 1 ); or, using regression analysis or other methods to analyze the polarization parameters of seawater particles and the corresponding true values ​​of seawater carbon parameters, and construct a digital model that includes the correlation between the polarization parameters of seawater particles and seawater carbon parameters (see Figure 2 );

[0051] S3: Using the correlation relationship, and considering the one-to-one correspondence between the output carbon parameter and the input seawater particle polarization data, the applicable scope and error control, a detection algorithm based on the analytical formula or the digital model is constructed for each carbon parameter (see Figure 3 );

[0052] S4: Execute the detection algorithm for the polarized light data of seawater particles detected by the instrument, wherein the seawater carbon parameter is obtained by solving the equation using the analytical formula, or input the polarized light data of the particles into the digital model and calculate the seawater carbon parameter using the digital model.

[0053] like Figure 1 As shown, in some preferred embodiments, the statistical method includes: obtaining seawater particle polarization data and related seawater carbon parameters and water quality parameters of seawater samples; identifying high-correlation parameters and low-correlation parameters between seawater particle polarization data and seawater carbon parameters through correlation analysis; based on the correlation analysis results, for high-correlation parameters, constructing an analytical formula, which describes the relationship between the high-correlation parameters and the seawater carbon parameters; based on the correlation analysis results, for low-correlation parameters, auxiliary construction of the analytical formula is performed to enhance the predictive ability of the analytical formula.

[0054] like Figure 2 As shown, in other preferred embodiments, the construction of the digital model includes: obtaining seawater particle polarization data and related seawater carbon parameters and water quality parameters of seawater samples; selecting features that have a significant impact on the prediction of seawater carbon parameters from the pre-processed seawater particle polarization data, including but not limited to the size, shape, polarization intensity, etc. of the particles; using a large amount of seawater sample data from the same source, combined with the selected features and known carbon parameters, to train the digital model to learn the mapping relationship between seawater particle polarization data and seawater carbon parameters; validating the trained digital model to evaluate its accuracy and reliability in predicting seawater carbon parameters; adjusting the model parameters and structure according to the results of the model validation to optimize the model performance; and saving the final optimized digital model for use in field testing to ensure the repeatability and consistency of the model.

[0055] In some preferred embodiments, the construction of the digital model also includes: extracting single-particle scattering pulse information using a pulse information extraction method based on the polarized light data of seawater particles measured by the instrument, preferably including using four polarized light detection channels and one fluorescence detection channel to collect the amplitude, distribution range, half-maximum full width, protrusion and other information of the single-particle scattering pulse to generate multi-dimensional data of particles for particle feature identification; counting the number of particles detected within a certain time at a certain flow rate to reflect the relative number of particles in seawater; wherein, the single pulse characteristics of the single particle information and the total number of specific types of particles in a statistical sense are characterized to reflect the information of seawater from multiple levels; and the obtained data are used for training and optimization of the digital model.

[0056] Specific embodiments of the present invention are further described below.

[0057] In order to establish the relationship between the particle polarization data and the seawater carbon parameters, a statistical method can be used to perform correlation analysis to find specific parameters with a high correlation with the carbon parameters, so as to construct an analytical formula for the carbon parameters with respect to the particle polarization data. The principle is as follows: Figure 1 As shown in Figure 2, the polarized light data of seawater particles can be used to obtain information about the type and quantity of seawater particles. Correlation analysis is performed on seawater quality parameters and seawater carbon parameters. Parameters with high correlations can be used to directly construct analytical formulas, while parameters with low correlations can be used to construct or optimize analytical formulas, serving as a reference and auxiliary. Using statistical methods, an analytical formula was obtained that reflects the correlation between the polarized light parameters of seawater particles and the seawater carbon parameters.

[0058] The composition of seawater particles is complex and has many characteristics. There are more than ten common water quality parameters. The polarization data of seawater particles contains many dimensions. In addition to statistical methods, artificial intelligence technology can also be used to construct a digital model of seawater particle polarization data and carbon parameters. The principle is as follows Figure 2 As shown in the figure, a particle classification instrument can be used to detect single particles in seawater, distinguishing them based on their characteristics. A regression model is constructed using the large amount of data generated by the instrument and the true carbon parameter values ​​of the corresponding seawater samples. This digital model is then saved for future use. This method produces a digital model that reflects the complex correlation between the polarization parameters of seawater particles and the carbon parameters of seawater.

[0059] By using the correlation relationship, taking into account the one-to-one correspondence between the output carbon parameter and the input seawater particle polarization data, the applicable scope and error control, the best detection algorithm is developed for each carbon parameter. Using this detection algorithm, within the applicable scope, the seawater carbon parameter can be obtained from the seawater particle polarization data. The construction principle and use process of the detection algorithm are as follows: Figure 3 As shown in the figure, by utilizing the existing correlations and screening through factors such as one-to-one correspondence, scope of application, and error control, a detection algorithm with better performance was ultimately obtained. When this detection algorithm is used to detect unknown seawater, the carbon parameters can be calculated using the polarization data of seawater particles.

[0060] For the statistical method, the polarized light data of seawater particles detected by the instrument is input into the analytical formula, and the equation is solved to obtain the carbon parameters of the unknown seawater; for the artificial intelligence digital model, the method of use is to input the polarized light data of various seawater particles of the unknown seawater into the existing digital model, and use the model calculation output to obtain the seawater carbon parameters.

[0061] In order to achieve on-site and rapid measurement of seawater carbon parameters, a portable and intelligent device is needed to accurately detect and analyze the composition of seawater, so as to detect the carbon parameters of seawater. Currently, there are instruments that meet the above conditions. The appearance of the instrument is as follows: Figure 4 In order to distinguish the types of particles, a polarization scattering database containing typical marine particles was constructed for the instrument. Typical particles in seawater include seaweed, microplastics, and mud. The instrument's classification effect on different particles is as follows: Figure 5 After the instrument database was built, the instrument was used to monitor seawater at 76 locations near Guangdong Province for several months, covering a wide range of offshore waters.

[0062] Each sample was also tested for a variety of carbon parameters and water quality indicators. These carbon parameters were determined under laboratory conditions according to the National Environmental Protection Standard HJ 501-2009: total carbon (TC), inorganic carbon (IC), non-purgeable organic carbon (NPOC), and total carbon and non-purgeable organic carbon in seawater filtered through a 0.45-micron PES membrane.

[0063] In order to make the model more accurate and have a wider range of applicability, a large number of carbon parameters of real seawater samples and corresponding particulate matter information models are invested in training. Preferably, for the data measured by the instrument, a pulse information extraction method that can reflect the multi-dimensional information of particulate matter is adopted. This helps to fully tap the pulse information detected by the instrument, extract as many pulse parameters as possible, perform particle feature identification, and improve the accuracy of particle type identification. This will generate multi-dimensional data containing a large amount of particulate matter information, which can be used to build artificial intelligence digital models. Preferably, the sources of this particulate matter information include: the amplitude, distribution range, half-maximum full width, protrusion and other information of single-particle scattered pulses collected by four polarized light detection channels and one fluorescence detection channel. At a certain flow rate, the number of particles detected within a certain period of time reflects the relative number of particles in seawater. A series of information such as the single pulse characteristics of single particle information and the total number of specific types of particles in a statistical sense are characterized, reflecting the information of seawater from multiple levels.

[0064] Preferably, the statistical method model can adopt linear fitting, assuming that the concentration of different types of particles is c n , the corresponding model coefficient is a n , then we can construct the functional relationship as Equation 1-Equation 3. Using the instrument's ability to identify the type of particles, we can get the concentration of particles c n Based on a large number of carbon parameters of real seawater, the coefficient a is calculated using the least square method. n Perform the calculation:

[0065] NPOC=f1(a1c1,a2c2...,a n c n ) (1)

[0066] TC=f2(a1c1,a2c2...,a n c n ) (2)

[0067] IC=f3(a1c1,a2c2...,a n c n ) (3)

[0068] When the amount of data is large, especially when the instrument extracts a large number of features, it is preferred to use artificial intelligence methods to construct a digital model. The instrument's carbon parameter detection model can use an appropriate regression model, such as random forest regression, Lasso regression, GBDT model, etc. Similar to the aforementioned correlation analysis, characteristic coefficient analysis of the preliminarily constructed model can be performed to study the particulate matter parameters that significantly affect parameter carbon, thereby constructing a regression model more scientifically and efficiently. Current research shows that pulse parameters such as peak size and full width at half maximum, as well as certain particulate matter parameters such as seaweed, microplastics, and sediment content, have the most significant impact on the regression model.

[0069] The model was trained using no fewer than 200 data sets. This is because a larger sample size minimizes overfitting, while seawater data from a wider range of locations allows the model to adapt to a wider range of marine scenarios. This process completes the model's construction. During model application, instruments can obtain on-site particulate matter information. This data, when input into the model, allows for on-site and rapid detection of seawater carbon parameters.

[0070] An embodiment of the present invention further provides a storage medium for storing a computer program, which at least performs the above method when executed.

[0071] An embodiment of the present invention further provides a control device, comprising a processor and a storage medium for storing a computer program; wherein the processor is configured to execute at least the method described above when executing the computer program.

[0072] An embodiment of the present invention further provides a processor, which executes a computer program and at least performs the method described above.

[0073] The storage medium can be implemented by any type of non-volatile storage device, or a combination thereof. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); the magnetic surface memory can be a magnetic disk memory or a magnetic tape memory. The storage medium described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0074] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0075] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0077] Those skilled in the art will appreciate that all or part of the steps of the above-mentioned method embodiments may be implemented by hardware associated with program instructions, and the aforementioned program may be stored in a computer-readable storage medium. When the program is executed, the program executes the steps of the above-mentioned method embodiments. The aforementioned storage medium includes various media that can store program codes, such as mobile storage devices, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0078] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0079] The methods disclosed in the several method embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new method embodiments.

[0080] The features disclosed in several product embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new product embodiments.

[0081] The features disclosed in several method or device embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new method embodiments or device embodiments.

[0082] The above is a further detailed description of the present invention in conjunction with specific preferred embodiments, and the specific implementation of the present invention should not be considered to be limited to these descriptions. Those skilled in the art will recognize that, without departing from the scope of the present invention, several equivalent substitutions or obvious variations can be made, and the performance or use of the same should be considered to fall within the scope of protection of the present invention.

Claims

1. A method for on-site detection of seawater carbon parameters using particulate matter polarized light data, characterized in that: The steps include: S1: Use polarized light technology to classify and extract features of seawater particles in seawater samples and build a polarized scattering database; S2: performing a correlation analysis on the polarized light data of seawater particles and the seawater carbon parameter using a statistical method to obtain an analytical formula containing the correlation between the polarized light parameters of seawater particles and the seawater carbon parameter; or, using a data set of the polarized light data of seawater particles and the corresponding seawater carbon parameter, establishing a digital model containing the correlation between the polarized light data of seawater particles and the seawater carbon parameter; the carbon parameter includes one or more of the following carbon content indicators: total carbon (TC), inorganic carbon (IC), non-purgeable organic carbon (NPOC), and total carbon and non-purgeable organic carbon in seawater after filtration using a filter membrane; The statistical methods include: Obtain polarized light data of seawater particles and related seawater carbon parameters and water quality parameters from seawater samples; Through correlation analysis, we can identify parameters with high and low correlation between polarized light data of seawater particles and seawater carbon parameters. Based on the correlation analysis results, for the highly correlated parameters, an analytical formula is constructed, wherein the analytical formula describes the relationship between the highly correlated parameters and the seawater carbon parameter; Based on the correlation analysis results, for low-correlation parameters, the analytical formula is assisted in construction to enhance the predictive ability of the analytical formula; The construction of the digital model includes: Obtain polarized light data of seawater particles and related seawater carbon parameters and water quality parameters from seawater samples; Select features that have a significant impact on the prediction of seawater carbon parameters from the pre-processed seawater particle polarization data; Using seawater sample data from the same source, combined with selected features and known carbon parameters, a numerical model is trained to learn the mapping relationship between seawater particle polarization data and seawater carbon parameters; Validate the trained numerical model to assess its accuracy and reliability in predicting seawater carbon parameters; According to the results of model verification, adjust the model parameters and structure to optimize the model performance; Save the final optimized digital model for use in on-site testing to ensure the repeatability and consistency of the model; S3: Utilizing the correlation relationship and taking into account the one-to-one correspondence between the output carbon parameter and the input seawater particulate matter polarization data, the applicable scope, and error control, a detection algorithm based on the analytical formula or the digital model is constructed for each carbon parameter; S4: Execute the detection algorithm for the polarized light data of seawater particles detected by the instrument, wherein the seawater carbon parameter is obtained by solving the equation using the analytical formula, or input the polarized light data of the particles into the digital model and calculate the seawater carbon parameter using the digital model.

2. The method for on-site detection of seawater carbon parameters according to claim 1, wherein: The characteristics that have a significant impact on the prediction of seawater carbon parameters include size, shape and polarized light intensity.

3. The method for on-site detection of seawater carbon parameters according to any one of claims 1 to 2, characterized in that: The construction of the digital model also includes: Based on the polarized light data of seawater particles measured by the instrument, a pulse information extraction method is used to extract the scattered pulse information of single particles and generate multi-dimensional data of particles for particle feature identification; the number of particles detected at a given flow rate and within a given time is counted to reflect the relative number of particles in the seawater; among them, the single pulse characteristics of the single particle information and the total number of specific types of particles in a statistical sense are characterized to reflect the information of seawater from multiple levels; the obtained data are used for training and optimization of the digital model.

4. The method for on-site detection of seawater carbon parameters according to claim 3, wherein: The instrument uses four polarized light detection channels and one fluorescence detection channel to collect one or more of the amplitude, distribution interval, full width at half maximum, and protrusion information of a single particle scattered pulse.

5. The method for on-site detection of seawater carbon parameters according to any one of claims 1 to 2, characterized in that: The statistical method uses linear fitting, and the concentration of different types of particles is c n , the corresponding model coefficient is a n , construct the functional relationship of Equation 1-Equation 3: ； The concentration of particles c is obtained by using the instrument to identify the particles n , based on a large number of real seawater carbon parameters, the coefficient a is calculated using the least squares method n .

6. The method for on-site detection of seawater carbon parameters according to any one of claims 1 to 2, characterized in that: A machine learning method is used to construct a digital model based on a regression model, and the regression model is a random forest regression model, a Lasso regression model or a GBDT model.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images