Fulvic acid fertilizer identification method, device, equipment, medium and product

Abstract

The invention discloses a fulvic acid fertilizer identification method, device, equipment, medium and product, and relates to the field of fulvic acid fertilizer identification, the method comprises: obtaining a to-be-detected solution, the to-be-detected solution comprising fulvic acid fertilizer solutions with different base sources and fulvic acid fertilizer solutions with different doping ratios; and performing three-dimensional fluorescence spectrum scanning on the extracting solution of the solution to be detected to obtain three-dimensional fluorescence spectrum matrix data. Preprocessing the three-dimensional fluorescence spectrum matrix data to obtain a data set, and dividing the data set into a training set and a verification set; and training a preset model based on the training set to obtain a trained prediction model. Acquiring a detection data set based on a to-be-identified fulvic acid fertilizer sample, inputting the detection data set into the prediction model, outputting a prediction result, and judging the base source and the doping proportion of the to-be-identified fulvic acid fertilizer sample according to the prediction result. The detection accuracy is improved.

Description

Technical Field

[0001] The present application relates to the field of fulvic acid fertilizer identification, and in particular to a fulvic acid fertilizer identification method, device, equipment, medium and product. Background Art

[0002] Humus is an important component of soil organic matter. It is a complex and relatively stable organic substance formed by the decomposition and synthesis of animal and plant residues by microorganisms. Fulvic acid is an amorphous organic weak acid mixture with a relatively small molecular weight in humus. It is soluble in dilute alkali, acid and water. It has aromatic, aliphatic and multiple functional group structural characteristics, has high biological activity, strong permeability, is easily absorbed by plants, and can also adjust the pH of the soil according to the needs of crops to create a suitable soil environment.

[0003] According to the different sources, fulvic acid is divided into two categories: mineral fulvic acid and biological fulvic acid. Mineral fulvic acid is usually used in the high-end agricultural market due to its high efficiency and high value, especially in soil remediation and crop quality improvement. However, due to its high cost, there are adulteration and inferior products in the market. Therefore, the fertilizer industry is in urgent need of a detection method that can accurately identify the source and type of fulvic acid.

[0004] At present, the identification and content determination methods of mineral-based and biological-based fulvic acid fertilizers are mainly carried out through two methods: fluorescence spectroscopy and potassium dichromate titration (HG / T 5334-2018, GB / T 34766-2017). Among them, the fluorescence spectroscopy method obtains the fluorescence excitation spectrum / emission spectrum of the substance through a fluorescence spectrophotometer, and qualitatively identifies and analyzes the mineral-based fulvic acid and biological-based fulvic acid based on the difference in the fluorescence excitation wavelength and emission wavelength of the two basic-based fulvic acids. However, as a high-molecular polymer organic molecule with multiple functional groups and complex structures, the fluorescence spectrum of fulvic acid overlaps seriously and is difficult to analyze. It is impossible to accurately distinguish fulvic acid from different basic sources by fluorescence spectroscopy alone. The potassium dichromate titration method consumes a lot of reagents, is not environmentally friendly, the analysis process is time-consuming and labor-intensive, and the reliability of the results is poor. Summary of the invention

[0005] The purpose of this application is to provide a method, device, equipment, medium and product for identifying fulvic acid fertilizer, which can improve the detection accuracy.

[0006] To achieve the above objectives, this application provides the following solutions:

[0007] In a first aspect, the present application provides a method for identifying fulvic acid fertilizer, comprising:

[0008] Obtaining a solution to be tested, wherein the solution to be tested includes fulvic acid fertilizer solutions of different origins and fulvic acid fertilizer solutions of different doping ratios;

[0009] Performing a three-dimensional fluorescence spectrum scan on the extract of the solution to be tested to obtain three-dimensional fluorescence spectrum matrix data;

[0010] Preprocessing the three-dimensional fluorescence spectrum matrix data to obtain a data set, and dividing the data set into a training set and a validation set;

[0011] Training a preset model based on the training set to obtain a trained prediction model;

[0012] A detection data set is obtained based on the fulvic acid fertilizer sample to be identified, the detection data set is input into the prediction model, a prediction result is output, and the base source and doping ratio of the fulvic acid fertilizer sample to be identified are determined by the prediction result.

[0013] In a second aspect, the present application provides a fulvic acid fertilizer identification device, comprising:

[0014] An acquisition module, used for acquiring a solution to be tested, wherein the solution to be tested includes fulvic acid fertilizer solutions of different origins and fulvic acid fertilizer solutions of different doping ratios;

[0015] A scanning module, used for performing a three-dimensional fluorescence spectrum scan on the extract of the solution to be tested to obtain three-dimensional fluorescence spectrum matrix data;

[0016] A processing module, used for preprocessing the three-dimensional fluorescence spectrum matrix data to obtain a data set, and dividing the data set into a training set and a validation set;

[0017] A training module, used to train a preset model based on the training set to obtain a trained prediction model;

[0018] The determination module is used to obtain a detection data set based on the fulvic acid fertilizer sample to be identified, input the detection data set into the prediction model, output a prediction result, and determine the base source and doping ratio of the fulvic acid fertilizer sample to be identified based on the prediction result.

[0019] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any one of the above-described methods for identifying fulvic acid fertilizers.

[0020] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for identifying fulvic acid fertilizers.

[0021] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for identifying fulvic acid fertilizers.

[0022] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0023] The present application provides a method, device, equipment, medium and product for identifying fulvic acid fertilizers. The method for identifying fulvic acid fertilizers obtains a solution to be tested, wherein the solution to be tested includes fulvic acid fertilizer solutions of different base sources and fulvic acid fertilizer solutions of different doping ratios. The extract of the solution to be tested is subjected to a three-dimensional fluorescence spectrum scan to obtain three-dimensional fluorescence spectrum matrix data. The three-dimensional fluorescence spectrum exclusive to fulvic acid fertilizers is established by a three-dimensional fluorescence spectrometer, which is simple to operate, shortens the detection time, improves the detection accuracy, and is safe and environmentally friendly. Then, the three-dimensional fluorescence spectrum matrix data is preprocessed to obtain a data set, and the data set is divided into a training set and a validation set. The preset model is trained by the above training set to obtain a trained prediction model. The present application replaces the traditional "physical separation" and "chemical separation" with a green and efficient "mathematical separation" method, saving experimental costs. A detection data set is obtained based on the fulvic acid fertilizer sample to be identified, the detection data set is input into the prediction model, and the prediction result is output. The base source and doping ratio of the fulvic acid fertilizer sample to be identified are determined by the prediction result. Through data preprocessing and model training, integrated identification of fulvic acid fertilizers from different sources and adulterated fulvic acid fertilizers is achieved. The identification process is simple and environmentally friendly, with a high identification accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 This is an application environment diagram of a method for identifying fulvic acid fertilizer in one embodiment of the present application;

[0026] Figure 2 A schematic diagram of a process for identifying fulvic acid fertilizer provided in one embodiment of the present application;

[0027] Figure 3 A schematic diagram of a flow chart of a method for identifying fulvic acid fertilizer provided in another embodiment of the present application;

[0028] Figure 4A schematic diagram of a flow chart of a method for identifying fulvic acid fertilizer provided in another embodiment of the present application;

[0029] Figure 5 A schematic diagram of a flow chart of a method for identifying fulvic acid fertilizer provided in another embodiment of the present application;

[0030] Figure 6 (a) is a three-dimensional fluorescence spectrum diagram of a three-dimensional fluorescence spectrum matrix data of a biological fulvic acid solution after background subtraction and scattering subtraction in an embodiment of the present application;

[0031] Figure 6 (b) is a factor decomposition diagram of the three-dimensional fluorescence spectrum after preprocessing the three-dimensional fluorescence spectrum matrix data of the biological fulvic acid solution in one embodiment of the present application;

[0032] Figure 7 (a) is a three-dimensional fluorescence spectrum diagram of a three-dimensional fluorescence spectrum matrix data of a mineral source fulvic acid solution after background subtraction and scattering subtraction in an embodiment of the present application;

[0033] Figure 7 (b) is a factor decomposition diagram of the three-dimensional fluorescence spectrum of the mineral source fulvic acid solution after preprocessing the three-dimensional fluorescence spectrum matrix data in one embodiment of the present application;

[0034] Figure 8 (a) is a three-dimensional fluorescence spectrum diagram after background subtraction and scattering subtraction of three-dimensional fluorescence spectrum matrix data of fulvic acid fertilizer solution with different doping ratios according to an embodiment of the present application;

[0035] Figure 8 (b) is a factor decomposition diagram of the three-dimensional fluorescence spectrum after preprocessing the three-dimensional fluorescence spectrum matrix data of fulvic acid fertilizer solutions with different doping ratios according to an embodiment of the present application;

[0036] Fig. 9 A PLS-DA model score graph provided in one embodiment of the present application;

[0037] Fig.10 A schematic diagram of functional modules of a fulvic acid fertilizer identification device provided in one embodiment of the present application;

[0038] Fig.11 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0040] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0041] Fulvic acid is a mixture of weak organic acids found in humus with a small relative molecular mass and an amorphous structure. This substance has unique solubility properties and can be dissolved in dilute alkaline solutions as well as easily in acidic and aqueous environments. The chemical structure of fulvic acid is complex and diverse, containing aromatic and aliphatic components, as well as a variety of functional groups, which give it strong biological activity and significant penetration capacity. Due to these properties, fulvic acid can be easily absorbed by plant cells, thereby exerting its effects in the plant body. In addition, fulvic acid has the ability to regulate soil pH, which can improve the soil environment according to the specific needs of crops and make it more suitable for plant growth. This regulatory effect is essential to creating a suitable soil environment, which helps crops better absorb nutrients and improve yield and quality.

[0042] In agricultural production, fulvic acid can be combined with other inorganic fertilizers such as nitrogen, phosphorus, potassium, etc. to produce fulvic acid compound fertilizer. This compound fertilizer has multiple functions of nitrogen fixation, phosphorus dissolution and potassium release, and can provide the nutrients needed by plants more effectively. At the same time, fulvic acid can also be directly processed into non-flocculated clear liquid fertilizer, which can significantly promote plant growth and enhance plant resistance to adversity. Another significant advantage of fulvic acid is that it can increase the organic matter content in the soil and provide a good growth environment for beneficial microorganisms. In this way, fulvic acid helps to inhibit the growth of harmful pathogens, thereby enhancing the natural disease and pest resistance of plants. This is of great significance for reducing the use of chemical pesticides and protecting the ecological environment. As a natural organic substance, fulvic acid is environmentally friendly and will not pollute soil and water sources. Its application helps to achieve sustainable development of agricultural production, reduce negative impacts on the environment, and improve the yield and quality of crops. Therefore, fulvic acid has a wide range of application prospects in modern agriculture and is a green fertilizer additive with great potential. As bulk organic fertilizers and feed additives, solid and liquid humic acid and its compound fertilizers have been widely used in my country's agricultural production, with output and application amount second only to traditional chemical fertilizers such as nitrogen, phosphorus and potassium.

[0043] According to the different sources, fulvic acid is divided into two categories: mineral fulvic acid and biological fulvic acid. Among them, mineral fulvic acid is mainly extracted from organic minerals such as weathered coal, lignite, and peat. It contains rich carboxyl, phenolic hydroxyl, carbonyl, quinone, methoxy and other groups, among which carboxyl and phenolic hydroxyl are its main functional groups. Biological fulvic acid is mainly produced by reacting biomass resources such as plant straw, sawdust, and bagasse as raw materials. The raw materials are abundant and renewable resources. At the same time, the raw materials are stable, the production process is easy to standardize and control, and the product quality is stable.

[0044] Compared with biological fulvic acid, mineral fulvic acid has the following advantages: (1) It contains more functional groups such as hydroxyl, phenolic, and carboxyl groups, so it has higher activity and stronger complexing ability; (2) It has a larger exchange capacity and pore structure, which helps to form soil aggregate structure, improve fertilizer utilization, and is not easy to absorb moisture; (3) Mineral fulvic acid contains about 60-70 mineral elements, which are needed for crop soil, while the mineral content of biological fulvic acid is usually low; (4) Mineral fulvic acid has high activity, and its dosage is usually only 10% of biological fulvic acid, which can reduce agricultural costs. Mineral fulvic acid is usually used in the high-end agricultural market due to its high efficiency and high value, especially in soil remediation and crop quality improvement. However, due to the high cost of mineral fulvic acid, there are adulteration and inferior products on the market. Therefore, the fertilizer industry is in urgent need of a detection method that can accurately identify the source and type of fulvic acid.

[0045] At present, the identification of mineral-based and biological-based fulvic acid fertilizers and the determination of their content are mainly carried out by fluorescence spectroscopy and potassium dichromate titration (HG / T 5334-2018, GB / T 34766-2017). Among them, the former obtains the fluorescence excitation spectrum / emission spectrum of the substance by a fluorescence spectrophotometer, and qualitatively distinguishes and analyzes the mineral-based fulvic acid and biological-based fulvic acid according to the difference in the fluorescence excitation wavelength and emission wavelength of the two basic-based fulvic acid. However, as a high-molecular polymer organic molecule with multiple functional groups and complex structures, the fulvic acid fluorescence spectrum overlaps seriously and is difficult to analyze. It is impossible to accurately distinguish fulvic acid from different basic sources by fluorescence spectroscopy alone. The potassium dichromate titration method not only consumes a lot of reagents, is unfriendly to the environment, and the analysis process is time-consuming and labor-intensive, and the reliability of the results is poor. Especially when there is doping of biological-based fulvic acid in the mineral-based fulvic acid, it is more difficult to give a reliable judgment by the above-mentioned existing detection technology. At the same time, the quality of the analysis results depends on the operating level and analysis experience, which is highly subjective and is not conducive to the standardization and objectivity of the detection method in practice.

[0046] In order to solve the quality analysis and authenticity problems of fulvic acid fertilizers in the above-mentioned agricultural fertilizer industry, the present application provides a method for quickly identifying the authenticity of fulvic acid fertilizers, based on the differences in components and component contents of biological and mineral fulvic acid fertilizers, combined with a three-dimensional fluorescence spectrum matrix and a factor decomposition algorithm, and through an artificial intelligence machine learning algorithm, accurate, rapid, and green identification and analysis of different sources of fulvic acid and possible adulteration is performed.

[0047] In order to solve the problem of quality analysis and authenticity verification of fulvic acid fertilizers in the agricultural fertilizer industry, this application proposes an efficient and reliable method for quickly identifying the authenticity of fulvic acid fertilizers. The core of this method is to utilize the differences in components and component contents between biological fulvic acid fertilizers and mineral fulvic acid fertilizers. By combining a three-dimensional fluorescence spectrum matrix, the base sources of fulvic acid fertilizers are distinguished, and possible adulteration is further identified. In addition, by introducing artificial intelligence and machine learning algorithms, accurate, rapid and environmentally friendly identification and analysis of fulvic acid fertilizers is achieved, thereby ensuring the quality and authenticity of agricultural fertilizers and providing a safer and more reliable fertilizer option for agricultural production.

[0048] The main structural formula of fulvic acid is as follows:

[0049]

[0050] The fulvic acid fertilizer identification method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the three-dimensional fluorescence spectrum matrix data to be processed to the server 104. After the server 104 receives the three-dimensional fluorescence spectrum matrix data to be processed, the server 104 pre-processes the three-dimensional fluorescence spectrum matrix data for the three-dimensional fluorescence spectrum matrix data to be processed, and obtains a data set, and then divides the data set into a training set and a verification set. The preset model is trained based on the training set to obtain a trained prediction model. Based on the fulvic acid fertilizer sample to be identified, a detection data set is obtained, the detection data set is input into the prediction model, and the prediction result is output. The base source and doping ratio of the fulvic acid fertilizer sample to be identified are determined by the prediction result. The server 104 can feedback the obtained base source and doping ratio determination results to the terminal 102. In addition, in some embodiments, the humic acid fertilizer identification method can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly perform data processing on the three-dimensional fluorescence spectrum matrix data to be processed, or the server 104 can obtain the three-dimensional fluorescence spectrum matrix data to be processed from the data storage system and perform data processing on the three-dimensional fluorescence spectrum matrix data to be processed.

[0051] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0052] In an exemplary embodiment, Figure 2 As shown, a method for identifying fulvic acid fertilizer is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate, including the following steps 201 to 205. Among them:

[0053] Step 201, obtaining a solution to be tested, wherein the solution to be tested includes fulvic acid fertilizer solutions of different origins and fulvic acid fertilizer solutions of different doping ratios.

[0054] Specifically, it is necessary to first prepare solution samples to be tested. These solution samples include two types: one is a fulvic acid fertilizer solution from different sources, and the other is a fulvic acid fertilizer solution mixed with biological fulvic acid and mineral fulvic acid in different proportions. The preparation of these two solution samples is for further related experimental analysis and research, so as to better understand and evaluate the performance and application effect of fulvic acid fertilizer.

[0055] Step 202: Perform a three-dimensional fluorescence spectrum scan on the extract of the solution to be tested to obtain three-dimensional fluorescence spectrum matrix data.

[0056] Specifically, there are two solutions to be tested in this embodiment: different base source fulvic acid fertilizer solutions and different doping ratio fulvic acid fertilizer solutions. Since different base source fulvic acid solutions include biological source fulvic acid solutions and mineral source fulvic acid solutions, and different doping ratio fulvic acid fertilizer solutions are solutions mixed with different proportions of biological source fulvic acid and mineral source fulvic acid, the extracts of the three solutions are actually subjected to three-dimensional fluorescence spectrum scanning. This embodiment adopts a sensitive and efficient three-dimensional fluorescence detection technology. This three-dimensional fluorescence spectrum detection method can comprehensively obtain the complete fluorescence spectrum information of biological and mineral source fulvic acid fertilizers. Through this method, not only can the fluorescence intensity change of the sample under different excitation wavelengths λex and different emission wavelengths λem be obtained, but also the fluorescence characteristics of the sample that change with wavelength can be described in detail. This spectrum is represented by a three-dimensional coordinate system, in which λem is used as the x-axis, λex is used as the y-axis, and the fluorescence intensity is used as the z-axis. In this way, a matrix spectrum can be constructed to show the fluorescence characteristics of the sample. In the three-dimensional fluorescence spectrum, we connect points with the same fluorescence intensity with lines to form a contour map. Through the contour map, we can not only determine the specific excitation wavelength λex and emission wavelength λem, but also accurately understand the fluorescence intensity at these specific wavelengths. By observing the change in fluorescence intensity when the wavelength changes, we can further analyze the fluorescence characteristics of the sample and obtain more comprehensive and in-depth analysis results.

[0057] It should be noted that the fluorescence excitation spectrum and emission spectrum of a substance are closely related to many factors such as its rigid structure, substituents, large π bonds, chromophore functional groups, etc. These factors work together to determine the specific wavelength range of the substance during fluorescence excitation and emission. Due to the differences in the raw material structure and properties of fulvic acid products extracted from different raw materials, their three-dimensional fluorescence peaks will appear in different excitation-emission wavelength ranges. These differences mainly come from the different number and position of component peaks in different parent core structures. Each fulvic acid product with a parent core structure has its own unique three-dimensional fluorescence fingerprint spectrum, which is specific and can reflect the unique characteristics of different fulvic acid products.

[0058] For example, the three-dimensional fluorescence spectrometer used is UltraFLUOR 1000 When setting the instrument parameters, the excitation wavelength λex is set between 200 and 700 nanometers, and the bandwidth is ensured to be 5 nanometers to obtain the best resolution and sensitivity. Similarly, the emission wavelength λem is also set between 200 and 700 nanometers, and the bandwidth is also 5 nanometers to ensure that the fluorescence signal of the sample at different wavelengths can be captured. In addition, in order to improve the detection sensitivity, the voltage of the photomultiplier diode (PMT) is set to 600 volts, which can ensure a higher signal-to-noise ratio when collecting the three-dimensional fluorescence spectrum data of the sample. Before the sample analysis, the sample solution must first be transferred to the cuvette. Place the cuvette containing the sample solution in the sample slot of the three-dimensional fluorescence spectrometer to ensure that its position is accurate for measurement. During the sample measurement process, the instrument automatically scans the excitation wavelength and emission wavelength to obtain the three-dimensional fluorescence spectrum data of the sample.

[0059] Specifically, the three-dimensional fluorescence spectrum matrix data includes emission wavelength, excitation wavelength and fluorescence intensity value.

[0060] For example, under the condition of excitation wavelength I and emission wavelength J, the extracts of K samples of the test solution are subjected to three-dimensional fluorescence spectrum scanning to obtain a series of fluorescence spectrum response matrices and collect them in a three-dimensional array of size I×J×K. X That is, in the three-dimensional fluorescence spectrum matrix, the matrix is ​​composed of the relative excitation spectrum matrix A, the relative emission spectrum matrix B, the relative concentration matrix C, and the residual matrix E.

[0061] According to the linear addition principle, each element x in X ijk The expression is as follows:

[0062]

[0063] Among them, x ijk is a matrix X The element (i, j, k) in is the fluorescence intensity of the kth sample when the excitation wavelength is i and the emission wavelength is j. in represents the element (i, n) in the relative excitation spectrum matrix A(I×N), that is, the relative excitation intensity of component n when the excitation wavelength is i; b jn represents the element (j, n) in the relative emission spectrum matrix B(J×N), that is, the relative emission intensity of component n when the emission wavelength is j; c kn represents the element (k, n) in the relative concentration matrix C(K×N), that is, the relative concentration of component n in the kth sample, e ijkis the element (i, j, k) in the residual matrix E, N represents the number of components, that is, the number of different components that may exist in the sample. ijk It represents the residual error that cannot be explained by the model in the measurement of the kth sample with excitation wavelength i and emission wavelength j.

[0064] Step 203: pre-process the three-dimensional fluorescence spectrum matrix data to obtain a data set, and divide the data set into a training set and a validation set.

[0065] Specifically, when processing three-dimensional fluorescence spectrum matrix data, it is first necessary to preprocess the data to ensure the quality and availability of the data. In order to train and verify the machine learning model, the data set needs to be further divided into a training set and a validation set. The training set is used to train the preset model so that the preset model can learn the characteristics and rules in the data. The validation set is used to evaluate the performance of the model to ensure that the model can also show good generalization ability on unknown data. Usually, the data set is divided into a training set and a validation set according to a certain ratio. For example, 80% of the data is used for training and the remaining 20% ​​is used for validation. Through this division, the model parameters are continuously adjusted during the training process to achieve the best training effect.

[0066] Step 204: Train the preset model based on the training set to obtain a trained prediction model.

[0067] In one embodiment, Figure 5 As shown, step 204 is replaced by the following steps 2041 to 2043:

[0068] Step 2041: train the preset PLS-DA model using the first training set to obtain a pattern recognition classification model, and the pattern recognition classification model is used to determine the source of the humic acid fertilizer sample.

[0069] Step 2042: Train the preset BP-ANN model using the second training set to obtain a regression prediction quantitative model, which is used to determine the doping ratio of biological fulvic acid and mineral fulvic acid in the fulvic acid fertilizer sample.

[0070] Specifically, since the solution to be tested includes fulvic acid fertilizer solutions from different origins and fulvic acid fertilizer solutions with different doping ratios, it can be understood that in step 202, the extract of the solution to be tested is subjected to a three-dimensional fluorescence spectrum scan to obtain three-dimensional fluorescence spectrum matrix data of fulvic acid fertilizer solutions from different origins and three-dimensional fluorescence spectrum matrix data of fulvic acid fertilizer solutions with different doping ratios, respectively. The three-dimensional fluorescence spectrum matrix data of fulvic acid fertilizer solutions from different origins also includes three-dimensional fluorescence spectrum matrix data of biological sources and three-dimensional fluorescence spectrum matrix data of mineral sources. In step 203, the above-mentioned biological source three-dimensional fluorescence spectrum matrix data, mineral source three-dimensional fluorescence spectrum matrix data and three-dimensional fluorescence spectrum matrix data of fulvic acid fertilizer solution with different doping ratios are preprocessed to obtain two sets of data sets, specifically: different source fulvic acid data sets and different doping ratio fulvic acid data sets, wherein the different source fulvic acid data sets include biological source fulvic acid data sets and mineral source fulvic acid data sets, the different source fulvic acid data sets are the first data sets, and the different doping ratio fulvic acid data sets are the second data sets, the first data set is divided into a first training set and a first validation set, and the second data set is divided into a second training set and a second validation set.

[0071] The prediction model includes a pattern recognition classification model and a regression prediction quantitative model. The pre-set partial least squares discriminant analysis (PLS-DA) model is trained, and a large amount of data is input into the model so that the model can identify the characteristic differences between different samples by learning and analyzing these data. After repeated iterations and optimization processes, the performance of the model is gradually improved, and finally a pattern recognition classification model is obtained, which can effectively identify the source of humic acid fertilizer. In this embodiment, the sources of humic acid fertilizer include biological humic acid and mineral humic acid. The supervised partial least squares discriminant analysis (PLSDA) method is used to discriminate and model humic acid fertilizers with different sources and perform cross-validation analysis, in order to distinguish and identify humic acid fertilizers with different sources and establish a pattern recognition classification model. Fig. 9 As shown, Fig. 9 is the PLS-DA model score graph, from Fig. 9 It can be seen that the grouping boundaries of humic acid fertilizers from different sources are obvious, indicating that there are obvious differences between the two. The results show that the PLS-DA model can be used to identify the source of humic acid fertilizers.

[0072] For example, the basic model of PLS-DA is:

[0073] X = TP + E;

[0074] Y = UQ + F;

[0075] Among them, X represents the independent variable matrix, usually refers to the fluorescence intensity, that is, the fluorescence intensity of each component obtained by factor decomposition, and Y represents the dependent variable matrix, that is, the category matrix of fulvic acid source. T and U correspond to the score matrix of the principal component decomposition of X and Y, respectively, reflecting the main components of the respective variables and dependent variables. P and Q represent the loading matrix of the principal component decomposition of X and Y, respectively, and E and F are error matrices.

[0076] To further analyze and predict, T and U were subjected to multiple linear regression analysis:

[0077] U = TB;

[0078] Among them, B is the regression coefficient matrix, which can be calculated by the following formula:

[0079] B=(X T X) -1 X T Y;

[0080] When predicting the source of the fulvic acid sample to be tested, it is necessary to calculate the score matrix T' of the fluorescence intensity matrix X' of the sample to be tested according to the load matrix P. T' can be obtained by the following formula:

[0081] T' = X'P;

[0082] Wherein, X' represents the fluorescence intensity matrix of the sample to be tested.

[0083] The predicted category Y' of the fulvic acid sample to be tested is further obtained, and Y' satisfies the following formula:

[0084] Y'=T'BQ;

[0085] The above formula is used to predict the category of fulvic acid source.

[0086] Optionally, the pattern recognition classification model can also be used to identify the classification of fertilizers doped with fulvic acid at different ratios.

[0087] In addition, in order to further verify the stability and reliability of the pattern recognition classification model, the pattern recognition classification results were statistically analyzed, and the classification of fertilizers doped with humic acid in different proportions was obtained. The statistical results of the pattern recognition classification are shown in Table 1. Through statistical analysis, it can be concluded that the classification accuracy of the model also reached 100%. Specifically, the rejection rate of each grouping is 0, which means that the model does not mistakenly exclude any samples that should be recognized. Similarly, the false recognition rate is 0, which indicates that the model does not mistakenly classify any samples that should not be recognized as positive. The sensitivity and specificity of the model are both 100%, which further verifies the high efficiency of the PLS-DA type in identifying and distinguishing humic acid fertilizers from different sources and doped humic acid fertilizers.

[0088] Table 1

[0089]

[0090] A systematic training process was carried out on the pre-set BP-ANN (back propagation neural network regression) model, which involved multiple iterations and parameter adjustments to ensure that the model could accurately capture the complex relationships in the data. After sufficient training, a regression prediction quantitative model was obtained, which can accurately quantitatively analyze the ratio of biogenic fulvic acid and mineral fulvic acid in fulvic acid fertilizers.

[0091] The BP-ANN (back propagation neural network regression) model is a complex structure consisting of multiple neurons (also called activation units) connected to each other through weights W and bias values ​​b. These neurons are organized into different layers, where the layer at the front end of the entire network is called the input layer, and the layer at the end is called the output layer. In addition, there are several hidden layers in the network, which are located between the input layer and the output layer and are responsible for processing and transmitting information. In the training process of the BP-ANN model, the forward propagation and backward propagation of the network error are mainly involved. In the forward propagation stage, the input data is sent to the input layer of the network, and after the weighted summation of neurons in each layer and the processing of the activation function, the prediction result is finally obtained in the output layer. Then, the prediction result is compared with the expected value to calculate the error of the network. Next, the back propagation stage begins, and the error signal is transmitted from the output layer to the input layer layer by layer, and the gradient of the weights and bias values ​​of each layer is calculated by the chain rule. According to these gradients, the weights and bias values ​​are adjusted by optimization algorithms (such as gradient descent) to reduce the network error.

[0092] When building a neural network model for classification tasks, we first need to construct a model training dataset (x (i) ,y (i) ), where the input data x represents the three-dimensional fluorescence matrix and y represents the predicted classification value of the model.

[0093] Through the processing of the activation function, the forward propagation output calculation of the network is performed layer by layer. The forward propagation output calculation formula is as follows:

[0094] a l =σ(w l a l-1 +b l );

[0095] Among them, a l represents the output of the lth layer of neurons, a l-1 represents the output of the l-1th layer of neurons, σ is the activation function, such as the Sigmoid function, which is defined as follows:

[0096]

[0097] By calculating layer by layer, the output value of the network is obtained, and then the total forward propagation error of the network is calculated, that is, the loss function, whose formula is as follows:

[0098]

[0099] Among them, f(x) represents the predicted classification value (expected value) of a sample, y represents the actual output value (prediction result) of the network for the sample, and δ is a hyperparameter used to control the turning point of the loss function.

[0100] After calculating the total error, the back propagation phase begins, and the weights W and bias values ​​b are updated using the gradient descent method. The partial derivative calculation formulas corresponding to the updated weights W and bias values ​​b are as follows:

[0101]

[0102] Among them, E total represents the overall loss, μ represents the learning rate, denoted as partial derivative.

[0103] Repeat the above forward propagation and back propagation process, and gradually adjust the weights and bias values ​​in the network until the total error between the network output value and the expected value reaches the preset value target.

[0104] It can be understood that the main function of the regression prediction quantitative model is to determine and distinguish the two main components in fulvic acid fertilizer: biological fulvic acid fertilizer and mineral fulvic acid fertilizer. Through in-depth analysis of sample data, the model can accurately calculate the proportion of each of these two components in the fertilizer. This function can be used to better understand and control the proportion of the product ingredients, thereby ensuring the quality of the fertilizer. The BP-ANN model was used to analyze fulvic acid fertilizer samples with unknown doping ratios and predict the doping ratio of unknown samples. The prediction results of the BP-ANN model are shown in Table 2.

[0105] Table 2

[0106]

[0107] According to the data analysis results in Table 2, it can be concluded that when the BP-ANN (back propagation artificial neural network) regression model is used to predict the unknown doping ratio of fulvic acid fertilizer, its prediction deviation is less than 3%, and the average prediction deviation is 1.4301%. This result further verifies the accuracy and reliability of the model in predicting the doping ratio of fulvic acid fertilizer, and is used to effectively estimate the doping ratio of various components in fulvic acid fertilizer.

[0108] The present application discloses an artificial intelligence identification method for samples of fulvic acid fertilizers of different origins and fulvic acid fertilizers doped with different proportions based on three-dimensional fluorescence spectra and high-order matrix decomposition technology. According to the differences in fluorescent components contained in fulvic acid fertilizers of different origins and fulvic acid fertilizers doped with different proportions, they can be quickly identified and distinguished with or without standard products, and supervised pattern recognition algorithms are used for modeling and cross-validation analysis. The results show that the model classification accuracy of samples of fulvic acid fertilizers of different origins and fulvic acid fertilizers doped with different proportions is 100%, and the prediction deviation for the doping ratio is <3%. The method is simple, rapid, highly sensitive, safe and environmentally friendly.

[0109] The present application proposes an innovative method based on three-dimensional fluorescence spectrum and high-order matrix decomposition technology, aiming to realize artificial intelligence identification of different base-source fulvic acid fertilizers and samples of fulvic acid fertilizers doped with different proportions. By analyzing and comparing samples of fulvic acid fertilizers from different base sources and samples doped with fulvic acid fertilizers in different proportions, it is found that there are significant differences in the fluorescent components contained in these samples. Based on these differences, the artificial intelligence identification method can quickly and accurately distinguish fulvic acid fertilizer samples of different base sources and doping ratios in the presence or absence of standard products. The method uses a supervised pattern recognition algorithm for modeling and cross-validation analysis to ensure the accuracy and reliability of the identification process. The model classification accuracy of the identification method for different base-source fulvic acid fertilizers and samples of fulvic acid fertilizers doped with different proportions reached 100%. In addition, the prediction deviation of the doping ratio is also controlled at an extremely low level, i.e., less than 3%. This result fully demonstrates the efficiency and accuracy of the method in identifying fulvic acid fertilizers. In addition, the identification method also has significant advantages such as simple operation, fast response speed, high sensitivity, safety and environmental protection. The simple operation process makes it easy for non-professionals to master it, and the quick response feature greatly shortens the time required for identification and improves work efficiency.

[0110] Step 2043: Evaluate the pattern recognition classification model using the first validation set, and evaluate the regression prediction quantitative model using the second validation set.

[0111] Optionally, the above step 2043 includes the following steps:

[0112] A1. Input the three-dimensional fluorescence spectrum matrix data in the first validation set and the second validation set into the pattern recognition classification model and the regression prediction quantitative model respectively to obtain the corresponding model judgment results, wherein the model judgment result of the pattern recognition classification model is used for the source discrimination of humic acid fertilizer, and the model judgment result of the regression prediction quantitative model is used for the doping discrimination of humic acid fertilizer.

[0113] A2. Compare the model judgment results with the measured results. If the model judgment results are completely consistent with the measured results, the model judgment is recorded as correct;

[0114] A3. Count the total number of correct model judgments and the total number of model judgments. Calculate the accuracy of humic acid fertilizer source discrimination and humic acid fertilizer adulteration discrimination based on the total number of correct model judgments and the total number of model judgments. Compare the predicted ratio of the regression model with the actual adulteration ratio and calculate the prediction deviation.

[0115] A4. The prediction model was evaluated based on the accuracy and prediction deviation of humic acid fertilizer source identification and humic acid fertilizer adulteration identification.

[0116] Specifically, the three-dimensional fluorescence spectrum matrix data in the first validation set and the second validation set need to be input into two different models for processing. The first model is a pattern recognition classification model, which identifies the source of fulvic acid fertilizer by analyzing data, that is, qualitatively analyzing its source or type. The second model is a regression prediction quantitative model, which determines the proportion of biological fulvic acid and mineral fulvic acid in fulvic acid fertilizer by analyzing data, that is, quantitatively analyzing its component content. Through the processing of these two models, we can get two groups of model judgment results, one for source discrimination and the other for doping ratio discrimination. Then, the model judgment results are compared and analyzed with the actual measurement results. In order to evaluate the accuracy of the model, a standard needs to be set. Under this standard, if the model judgment result is completely consistent with the measured result or within the allowable deviation range, then the model judgment is considered to be correct. This consistency is one of the key indicators for evaluating model performance.

[0117] It can be understood that in order to more comprehensively evaluate the performance of the pattern recognition classification model and the regression prediction quantitative model, statistical analysis is required. The total number of correct model judgments and the total number of model judgments are counted. Through these two statistical data, the accuracy of the source discrimination of humic acid fertilizer and the accuracy of the doping discrimination of humic acid fertilizer are calculated. Accuracy = (total number of correct model judgments / total number of model judgments) × 100%, and the prediction model is comprehensively evaluated according to the calculated accuracy. And the statistical model judges the predicted ratio and the actual doping ratio, and the deviation of the doping ratio of doped humic acid fertilizer is calculated through these two statistical data, prediction deviation = (predicted ratio-actual ratio) / actual ratio × 100%. Accuracy and prediction deviation are important indicators for measuring the prediction ability of the model, which reflects the reliability and effectiveness of the model in practical applications. By comparing the accuracy and prediction deviation of the source discrimination and doping discrimination of humic acid fertilizer, the model is optimized and adjusted to improve its prediction performance.

[0118] In one embodiment, the similarity between the spectrum of the sample to be tested and the standard spectrum can also be evaluated by using the angle cosine supervised learning algorithm, and the origin of the sample to be tested can be determined by similarity discrimination.

[0119] Specifically, when evaluating the origin of the sample to be tested, the angle cosine supervised learning algorithm can be used to make a judgment by calculating the similarity between the spectrum of the sample to be tested and the standard spectrum. The spectrum of the sample to be tested should show basically consistent peaks at the corresponding positions of the standard spectrum, and under naked eye inspection, there should not be any significant peaks, except those that already exist in the standard spectrum. This method can quantitatively measure the similarity between the spectrum of the sample to be tested and the standard spectrum by calculating the cosine value of the angle between the two. The closer the angle cosine value is to 1, the higher the similarity between the two; conversely, the greater the difference between the angle cosine value and 1, the lower the similarity between the two. Through this quantitative similarity assessment, we can more accurately determine the origin of the sample to be tested.

[0120] Step 205: acquiring a test data set based on the humic acid fertilizer sample to be identified, inputting the test data set into a prediction model, outputting a prediction result, and determining the base source and doping ratio of the humic acid fertilizer sample to be identified based on the prediction result.

[0121] In one embodiment, the above step 205 includes the following steps:

[0122] B1. Take a certain amount of fulvic acid fertilizer as a sample to be identified, add water to the fulvic acid fertilizer to be identified, and perform ultrasonic extraction and centrifugation in sequence to obtain a fulvic acid fertilizer solution.

[0123] B2. Add water to the humic acid fertilizer solution and dilute it to the predicted concentration to obtain the solution to be identified.

[0124] B3. Perform a three-dimensional fluorescence spectrum scan on the extract of the solution to be identified to obtain three-dimensional fluorescence spectrum matrix data of the solution to be identified.

[0125] B4. Preprocess the three-dimensional fluorescence spectrum matrix data of the solution to be identified to obtain a data set to be identified.

[0126] B5. Input the data set to be identified into the pattern recognition classification model, the source of the humic acid fertilizer to be identified can be discriminated, and the prediction results of the pattern recognition classification model are output. The prediction classification includes biological humic acid and mineral humic acid. If the prediction classification has no attribution, the humic acid fertilizer to be identified is neither biological humic acid nor mineral humic acid. Input the data set to be identified into the regression prediction quantitative model, the doping ratio of biological humic acid and mineral humic acid in the humic acid fertilizer to be identified can be discriminated, and the prediction results of the regression prediction quantitative model are output to obtain the doping ratio of biological humic acid and mineral humic acid in the humic acid fertilizer to be identified.

[0127] In another exemplary embodiment of the present application, Figure 3As shown, the above step 201 is replaced by the following steps 2011 to 2014:

[0128] Step 2011: Obtain samples to be tested, wherein the samples to be tested include biological fulvic acid fertilizer and mineral fulvic acid fertilizer stored separately.

[0129] Step 2012: adding water to the biological fulvic acid fertilizer and the mineral fulvic acid fertilizer, respectively, and performing ultrasonic extraction and centrifugation treatment on each of them in sequence to obtain a biological fulvic acid solution and a mineral fulvic acid solution.

[0130] Specifically, the biogenic fulvic acid fertilizer and the mineral fulvic acid fertilizer are placed in two different containers, respectively, and then an appropriate amount of pure water is added to the two containers. Then, the two fertilizers mixed with water are subjected to ultrasonic extraction treatment to ensure that the effective ingredients in the fertilizers can be fully dissolved in the water. After the ultrasonic extraction is completed, the biogenic fulvic acid mixed solution and the mineral fulvic acid mixed solution are centrifuged respectively to separate the solid residue and the liquid effective ingredients. Through the high-speed rotation of the centrifuge, the solid part will be thrown to the bottom of the container, while the liquid part will remain in the upper layer. Finally, the upper liquid is collected separately, so that a biogenic fulvic acid solution and a mineral fulvic acid solution are obtained. These two solutions contain the effective ingredients extracted from the biogenic fulvic acid fertilizer and the mineral fulvic acid fertilizer, respectively.

[0131] For example, weigh 0.0250g of each of the biological source fulvic acid fertilizer and mineral source fulvic acid fertilizer to be detected, and then place them in 25mL volumetric flasks respectively, and add appropriate amount of Wahaha pure water to the two volumetric flasks until the water level reaches the scale line of the volumetric flask. Afterwards, the two volumetric flasks are placed in an ultrasonic extractor for ultrasonic extraction, and the extraction time is 10 minutes. After the extraction is completed, the two volumetric flasks are taken out and placed in a room temperature environment, and appropriate amount of solution is taken from the two volumetric flasks respectively in a centrifuge tube and the solution is centrifuged using a centrifuge. After the centrifugation is completed, the supernatant is carefully transferred to a clean container and used as a mother liquor for subsequent experiments.

[0132] Step 2013, dilute the biological fulvic acid solution with water to a first preset concentration to obtain a first test solution, and dilute the mineral fulvic acid solution with water to a second preset concentration to obtain a second test solution, wherein the first test solution and the second test solution constitute fulvic acid fertilizer solutions from different sources.

[0133] Illustratively, the stock solution is diluted with water to a concentration of 0.1 mg / mL.

[0134] Step 2014: Mix the biological fulvic acid solution and the mineral fulvic acid solution in multiple groups of different proportions, and add water to dilute them to a third preset concentration to obtain fulvic acid fertilizer solutions with different doping ratios.

[0135] Specifically, it is necessary to dilute the biological source humic acid solution, add an appropriate amount of water according to a certain ratio, so that the concentration of the solution reaches a preset first concentration value, thereby obtaining a first solution to be tested. Similarly, for the mineral source humic acid solution, a similar dilution operation is also required, that is, it is diluted with water to a second preset concentration, thereby obtaining a second solution to be tested. It should be noted that the two solutions to be tested represent humic acid fertilizer solutions from different sources, namely biological sources and mineral sources, which will be used for comparison and analysis in subsequent experiments or applications. The values ​​of the first preset concentration, the second preset concentration and the third preset concentration can be selected to be the same or different according to experimental needs. In this embodiment, the first preset concentration, the second preset concentration and the third preset concentration are all set to 0.1 mg / mL.

[0136] The biological source fulvic acid solution and the mineral source fulvic acid solution are mixed in a variety of different proportions. During the mixing process, a plurality of different proportions can be used, such as 1:0, 0:1, 9:1, etc., to ensure that a variety of fulvic acid fertilizer solutions with different doping ratios can be obtained. After the mixing is completed, the mixed solution needs to be diluted with water to a third preset concentration, so that fulvic acid fertilizer solutions with different doping ratios are obtained. It should be noted that the values ​​of the first preset concentration, the second preset concentration, and the third preset concentration are exactly the same, so that the concentration variable can be ensured to remain consistent, thereby making the test results more accurate and reliable.

[0137] For example, biogenic fulvic acid and mineral fulvic acid are prepared in 11 ratios (biogenic fulvic acid: mineral fulvic acid = 1:0, 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8, 1:9, 0:1) to prepare biogenic and mineral mixed fulvic acid fertilizer (doped fulvic acid fertilizer) with a concentration of 0.1 mg / mL.

[0138] In another exemplary embodiment of the present application, Figure 4 As shown, the above step 203 is replaced by the following steps 2031 to 2032:

[0139] Step 2031: perform background subtraction and scatter subtraction on the three-dimensional fluorescence spectrum matrix data.

[0140] Specifically, when processing the three-dimensional fluorescence spectrum matrix data, all collected data are imported into the ModelLabSpecman3D chemometric modeling software in batches to preprocess the three-dimensional fluorescence spectrum data. First, background subtraction and scattering subtraction are required to ensure the accuracy and reliability of the data. Background subtraction is mainly to eliminate non-fluorescent background signals that may be introduced during the experiment. Scattering subtraction is to remove various scattering effects, including Raman scattering, Rayleigh scattering, and secondary Rayleigh scattering. In order to achieve this goal, an efficient interpolation algorithm is used to accurately subtract scattering and background through interpolation methods. This method can not only effectively remove Raman scattering, that is, scattering signals caused by molecular vibration, but also effectively remove Rayleigh scattering, that is, signals caused by molecular elastic scattering. In addition, secondary Rayleigh scattering, that is, signals caused by multiple scattering, are also effectively subtracted. Through the above method, purer and more accurate three-dimensional fluorescence spectrum matrix data can be obtained.

[0141] For example, when performing three-dimensional fluorescence spectroscopy analysis of water solvents, background subtraction is first required to ensure the accuracy of the data. Utilizing the characteristics of Rayleigh scattering, the Rayleigh scattering bandwidth is set to 65 nanometers to ensure that the Rayleigh scattering signal can be effectively identified and subtracted. In addition, in order to further improve the accuracy of the data, considering the influence of secondary Rayleigh scattering, the bandwidth of secondary Rayleigh scattering is set to 75 nanometers. As a common interference signal, Raman scattering has a bandwidth set to 0 nanometers to ensure that it will not affect the results during data processing. In order to process the scattered signal more finely, the cubic Spline spline interpolation method is used for scattering subtraction. By constructing a smooth curve to fit the data points, the influence of the scattered signal on the fluorescence spectrum is effectively eliminated, ensuring the accuracy and reliability of the final result.

[0142] Step 2032: Decompose and correct the three-dimensional fluorescence spectrum matrix data based on a factor decomposition algorithm to obtain a data set, wherein the data set includes a first data set and a second data set, the first data set is divided into a first training set and a first validation set, and the second data set is divided into a second training set and a second validation set.

[0143] Specifically, in order to accurately decompose and correct the three-dimensional fluorescence spectrum matrix data, a method based on a factor decomposition algorithm is used to obtain a corrected data set. Specifically, the SWATLD factor decomposition algorithm is used to decompose and correct the three-dimensional fluorescence spectrum. Through this method, EX / EM spectra of various components in fulvic acid fertilizer extracts from different sources and doped fulvic acid fertilizer extracts are obtained.

[0144] SWATLD uses the alternating least squares principle and Moore-Penrose generalized inverse calculation based on T-SVD (truncation singular value decomposition) to calculate the three-dimensional fluorescence spectrum matrix data. X Decompose. Let a (i) 、b (j) 、c (k) Respectively represent the row vectors composed of the i-th, j-th, and k-th row elements in matrices A, B, and C. For three-dimensional fluorescence spectrum matrix data X When decomposing, the iterative formulas of A, B, and C are respectively composed of the following formulas:

[0145]

[0146] Three-dimensional fluorescence spectrum matrix data of the sample X Also expressed as:

[0147] X i.. =Bdiag(a (i) )C T +E i.. (i=1,……,I);

[0148] X .j. =Cdiag(b (j) )A T +E .j. (j=1, ..., J);

[0149] X ..k =Adiag(c (k) )B T +E ..k (k=1,……,K);

[0150] Among them, X i.. , X .j. , X ..k Represents three-dimensional fluorescence spectrum matrix data X The i-th horizontal plane matrix (J×K), the j-th side matrix (K×I), and the k-th frontal matrix (I×J) of the vector are represented by diag, T represents transpose, diagm represents taking the matrix diagonal elements, and + represents the Moore-Penrose generalized inverse. In the above three formulas, E i.. , E .j. , E ..k Represents the error term or residual, which is a random variable used to represent the variability that cannot be explained by the model. Through the above decomposition method, the spectral information of each component in the sample can be obtained more clearly. in 、b jn 、c kn .

[0151] The three-dimensional fluorescence spectrum matrix data of biological fulvic acid solution, mineral fulvic acid solution, and fulvic acid fertilizer solution with different doping ratios were subjected to background subtraction, scattering subtraction, and factor decomposition. Figure 6 (a) Figure 6 (b) Figure 7 (a) Figure 7 (b) Figure 8 (a) Figure 8 As shown in (b), by carefully observing the above 6 images, we can clearly identify that whether it is biological fulvic acid solution, mineral fulvic acid solution, or fulvic acid fertilizer solution with different doping ratios, their three-dimensional fluorescence spectrum matrix data after the preprocessing step shows significant differences compared with the original data before preprocessing. Specifically, Figure 6 (b) Figure 7 (b) and Figure 8 In (b), it can be intuitively observed that the distribution and intensity of each component of the preprocessed three-dimensional fluorescence spectrum matrix data become clearer, making the analysis and interpretation of the data clearer and more accurate. Through this preprocessing method, the background noise can be effectively removed, the contrast of the signal can be enhanced, and a basis for subsequent data analysis can be provided.

[0152] Based on the same inventive concept, the embodiment of the present application also provides a fulvic acid fertilizer identification device for implementing the above-mentioned fulvic acid fertilizer identification method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more fulvic acid fertilizer identification device embodiments provided below can refer to the limitations of the fulvic acid fertilizer identification method above, and will not be repeated here.

[0153] In an exemplary embodiment, Fig.10 As shown, a fulvic acid fertilizer identification device is provided, comprising:

[0154] An acquisition module 1010 is used to acquire a solution to be tested, wherein the solution to be tested includes fulvic acid fertilizer solutions of different origins and fulvic acid fertilizer solutions of different doping ratios;

[0155] The scanning module 1020 is used to perform a three-dimensional fluorescence spectrum scan on the extract of the solution to be tested to obtain three-dimensional fluorescence spectrum matrix data;

[0156] The processing module 1030 is used to pre-process the three-dimensional fluorescence spectrum matrix data to obtain a data set, and divide the data set into a training set and a validation set;

[0157] A training module 1040 is used to train a preset model based on a training set to obtain a trained prediction model;

[0158] The first determination module 1050 is used to obtain a detection data set based on the fulvic acid fertilizer sample to be identified, input the detection data set into the prediction model, output the prediction result, and determine the base source and doping ratio of the fulvic acid fertilizer sample to be identified based on the prediction result.

[0159] As an optional implementation, the acquisition module 1010 is specifically configured to:

[0160] Obtaining a sample to be tested, wherein the sample to be tested includes a biological source fulvic acid fertilizer and a mineral source fulvic acid fertilizer that are stored separately;

[0161] Adding water to the biological fulvic acid fertilizer and the mineral fulvic acid fertilizer, respectively, and performing ultrasonic extraction and centrifugation treatment on each of them in sequence to obtain a biological fulvic acid solution and a mineral fulvic acid solution;

[0162] The biological fulvic acid solution is diluted with water to a first preset concentration to obtain a first test solution, and the mineral fulvic acid solution is diluted with water to a second preset concentration to obtain a second test solution, wherein the first test solution and the second test solution constitute different-source fulvic acid fertilizer solutions;

[0163] The biological source fulvic acid solution and the mineral source fulvic acid solution are mixed in multiple groups according to different proportions, and water is added to dilute them to a third preset concentration to obtain fulvic acid fertilizer solutions with different doping proportions.

[0164] As an optional implementation manner, the processing module 1030 is specifically configured to:

[0165] Perform background subtraction and scatter subtraction on the three-dimensional fluorescence spectrum matrix data;

[0166] The three-dimensional fluorescence spectrum matrix data is decomposed and corrected based on a factor decomposition algorithm to obtain a data set, wherein the data set includes a first data set and a second data set, the first data set is divided into a first training set and a first validation set, and the second data set is divided into a second training set and a second validation set.

[0167] As an optional implementation, the training module 1040 is specifically configured to:

[0168] The preset PLS-DA model is trained by the first training set to obtain a pattern recognition classification model, and the pattern recognition classification model is used to determine the source of the humic acid fertilizer sample;

[0169] The preset BP-ANN model is trained by the second training set to obtain a regression prediction quantitative model, and the regression prediction quantitative model is used to determine the doping ratio of biological source fulvic acid and mineral source fulvic acid in the fulvic acid fertilizer sample;

[0170] The pattern recognition classification model is evaluated by the first validation set, and the regression prediction quantitative model is evaluated by the second validation set.

[0171] As an optional embodiment, a fulvic acid fertilizer identification device further includes:

[0172] The second determination module 1060 is used to evaluate the similarity between the spectrum of the sample to be tested and the standard spectrum based on the angle cosine algorithm, and determine the origin of the sample to be tested by similarity determination.

[0173] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Fig.11 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store three-dimensional fluorescence spectrum matrix data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. The computer program is executed by the processor to implement a method for identifying fulvic acid fertilizer.

[0174] Those skilled in the art will understand that Fig.11 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0175] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0176] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0177] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0178] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0179] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0180] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0181] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0182] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for identifying fulvic acid fertilizer, characterized in that: The fulvic acid fertilizer identification method comprises: Obtaining a solution to be tested, wherein the solution to be tested includes fulvic acid fertilizer solutions of different origins and fulvic acid fertilizer solutions of different doping ratios; Performing a three-dimensional fluorescence spectrum scan on the extract of the solution to be tested to obtain three-dimensional fluorescence spectrum matrix data; Preprocessing the three-dimensional fluorescence spectrum matrix data to obtain a data set, and dividing the data set into a training set and a validation set; Training a preset model based on the training set to obtain a trained prediction model; A detection data set is obtained based on the fulvic acid fertilizer sample to be identified, the detection data set is input into the prediction model, a prediction result is output, and the base source and doping ratio of the fulvic acid fertilizer sample to be identified are determined by the prediction result.

2. The method for identifying fulvic acid fertilizer according to claim 1, characterized in that: The obtaining of the solution to be tested specifically comprises: Obtaining a sample to be tested, wherein the sample to be tested includes a biological source fulvic acid fertilizer and a mineral source fulvic acid fertilizer that are stored separately; Adding water to the biological fulvic acid fertilizer and the mineral fulvic acid fertilizer, respectively, and performing ultrasonic extraction and centrifugation treatment on each of them in sequence to obtain a biological fulvic acid solution and a mineral fulvic acid solution; The biological fulvic acid solution is diluted with water to a first preset concentration to obtain a first test solution, and the mineral fulvic acid solution is diluted with water to a second preset concentration to obtain a second test solution, wherein the first test solution and the second test solution constitute the different-source fulvic acid fertilizer solution; The biological source fulvic acid solution and the mineral source fulvic acid solution are mixed in a plurality of different proportions, and diluted with water to a third preset concentration to obtain the fulvic acid fertilizer solutions with different doping proportions.

3. The method for identifying fulvic acid fertilizer according to claim 1, characterized in that: The three-dimensional fluorescence spectrum matrix data includes emission wavelength, excitation wavelength and fluorescence intensity value.

4. The method for identifying fulvic acid fertilizer according to claim 1, characterized in that: The preprocessing of the three-dimensional fluorescence spectrum matrix data to obtain a data set, and dividing the data set into a training set and a validation set specifically includes: Performing background subtraction and scattering subtraction on the three-dimensional fluorescence spectrum matrix data to obtain processed three-dimensional fluorescence spectrum matrix data; The processed three-dimensional fluorescence spectral matrix data is decomposed and corrected based on a factor decomposition algorithm to obtain the data set, wherein the data set includes a first data set and a second data set, the first data set is divided into a first training set and a first validation set, and the second data set is divided into a second training set and a second validation set.

5. The method for identifying fulvic acid fertilizer according to claim 4, characterized in that: The training of the preset model based on the training set to obtain a trained prediction model specifically includes: Training a preset PLS-DA model using the first training set to obtain a pattern recognition classification model, wherein the pattern recognition classification model is used to determine the source of the fulvic acid fertilizer sample; The preset BP-ANN model is trained by the second training set to obtain a regression prediction quantitative model, wherein the regression prediction quantitative model is used to determine the doping ratio of biological source fulvic acid and mineral source fulvic acid in the fulvic acid fertilizer sample; The pattern recognition classification model is evaluated using the first validation set, and the regression prediction quantitative model is evaluated using the second validation set.

6. The method for identifying fulvic acid fertilizer according to claim 2, characterized in that: The fulvic acid fertilizer identification method also includes: The similarity between the spectrum of the sample to be tested and the standard spectrum is evaluated based on the angle cosine algorithm, and the origin of the sample to be tested is determined according to the similarity.

7. A fulvic acid fertilizer identification device, characterized in that: The fulvic acid fertilizer identification device comprises: An acquisition module, used for acquiring a solution to be tested, wherein the solution to be tested includes fulvic acid fertilizer solutions of different origins and fulvic acid fertilizer solutions of different doping ratios; A scanning module, used for performing a three-dimensional fluorescence spectrum scan on the extract of the solution to be tested to obtain three-dimensional fluorescence spectrum matrix data; A processing module, used for preprocessing the three-dimensional fluorescence spectrum matrix data to obtain a data set, and dividing the data set into a training set and a validation set; A training module, used to train a preset model based on the training set to obtain a trained prediction model; The determination module is used to obtain a detection data set based on the fulvic acid fertilizer sample to be identified, input the detection data set into the prediction model, output a prediction result, and determine the base source and doping ratio of the fulvic acid fertilizer sample to be identified based on the prediction result.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the fulvic acid fertilizer identification method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fulvic acid fertilizer identification method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the fulvic acid fertilizer identification method according to any one of claims 1 to 6 are implemented.

Images