Method for identifying algae in water body and measuring chlorophyll content
By establishing a fluorescent fingerprint database of algae and training a neural network model, the problem of accuracy in algae identification and chlorophyll content measurement in water bodies was solved, achieving rapid identification and precise measurement.
Patent Information
- Application Number
- CN202511749232.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-25
- Publication Date
- 2026-02-27
AI Technical Summary
Existing technologies cannot accurately identify the composition of different algae in water bodies and have low accuracy in chlorophyll content measurement. Single-channel fluorescence measurement methods cannot distinguish algae and are greatly affected by environmental factors. Three-dimensional fluorescence spectroscopy methods have complex hardware and high costs, and cannot adapt to complex environments.
A fluorescence fingerprint database of algae was established, a neural network model was trained, and the effects of temperature and light were corrected by multi-channel fluorescence measurement and neural network model, so as to achieve accurate algae identification and chlorophyll content measurement.
It enables rapid and accurate identification of different algae in water bodies and precise measurement of chlorophyll content, and can output the relative abundance of each algae and chlorophyll a concentration, achieving dual characterization of algal community structure and biomass in water bodies.
Smart Images
Figure CN121577593A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of water environment detection technology, specifically relating to a method for identifying algae and measuring chlorophyll content in water bodies. Background Technology
[0002] Phytoplankton species and abundance can be used to evaluate water quality and reflect the ecological status of aquatic bodies. According to the evolutionary classification method proposed by Falkowski et al. (2004), phytoplankton are divided into the green algae lineage (including Chlorophyta), the red algae lineage (including Diatoms, Cryptophyta, Chrysophyta, and Chrysophyta), and the cyanobacteria lineage (including Cyanobacteria). Algae within the same lineage exhibit similar fluorescence responses (fluorescence ratios), while algae from different lineages show significant differences in pigment composition, resulting in marked differences in their fluorescence response characteristics under the same excitation wavelength. Light is one of the most important environmental factors affecting phytoplankton photosynthesis and physiological state. Whether under long-term light exposure (photoacclimation) or short-term light changes caused by diurnal variations, ocean currents, and tides, phytoplankton will physiologically adapt to different light conditions. These physiological adaptations affect the apparent fluorescence efficiency of phytoplankton. Temperature is another important environmental factor affecting the photoluminescence efficiency of phytoplankton. When a fluorophore is excited by incident light, electrons transition to higher energy levels. When these electrons return to the ground state, energy is released as fluorescent photons or through non-radiative transitions. Temperature changes alter the energy balance during these transitions, causing the proportion of non-radiative transitions to increase with rising temperature, leading to a decrease in fluorescence intensity. Therefore, to accurately measure the chlorophyll content of algae in water, the effects of algae, light, and temperature must be considered.
[0003] Currently, single-channel fluorescence measurement is widely used for monitoring algae in water bodies. This method uses a single excitation wavelength to excite chlorophyll molecules in algae, estimates chlorophyll content by measuring fluorescence intensity, and then corrects the calibration slope through temperature response correction and light response correction (i.e., fluorescence quenching correction). However, this method has the following limitations: (1) Unable to distinguish algae: Single-channel fluorescence measurement only captures fluorescence information of a single excitation-emission wavelength pair. The fluorescence signals of different algae at the same wavelength overlap, and only a rough estimate of the total chlorophyll content can be obtained. It is impossible to accurately identify the specific algae composition in the water (such as blue algae, green algae, red algae, etc.), and it is difficult to achieve algae identification and fluorescence signal separation. (2) Low measurement accuracy: Single-channel fluorescence measurement is significantly affected by changes in light intensity, temperature, and pigment ratio. Although temperature and light response corrections are performed, these corrections are usually isolated, linear, and post-compensation, and cannot compensate for the coupling effect of multiple factors such as temperature and light in the actual water environment. As a result, the chlorophyll content estimation error is large, affecting the accuracy of the measurement results. (3) Poor model adaptability: The signal dimension of single-channel fluorescence measurement is singular and cannot distinguish algae. The calibration process depends on preset empirical parameters. Its calibration and correction parameters (such as slope and correction coefficient) are determined by experiments under specific water bodies, specific dominant algae and specific environmental conditions. It is difficult to adapt to mixed algae systems in water bodies under complex natural environments, resulting in model rigidity and weak generalization ability.
[0004] Currently, some studies propose using three-dimensional fluorescence data or discrete three-dimensional fluorescence spectra for algal identification and quantification. This method involves: measuring the three-dimensional fluorescence spectra of different algae in the laboratory, establishing a standard three-dimensional fluorescence spectrum library for different algae using chlorophyll a concentration as a unified standard; measuring the discrete three-dimensional fluorescence spectra of mixed phytoplankton samples; and performing multiple linear regression analysis using the standard three-dimensional fluorescence spectra of each phytoplankton phylum as independent variables and the fluorescence spectrum of the mixed phytoplankton sample as the dependent variable to obtain the chlorophyll a concentration of each phytoplankton phylum. This method provides relatively rich algal fluorescence fingerprint information, allowing for more detailed identification of phytoplankton, but it has the following limitations: (1) This method requires setting up an LED array and a multi-band filter group, designing a more complex optical path and more optical components, and requires stronger data processing capabilities, resulting in high hardware complexity and cost; in order to obtain a three-dimensional fluorescence spectrum, a large amount of excitation-emission wavelength pair data needs to be collected, resulting in slow measurement speed and high instrument power consumption; (2) This method mainly focuses on spectral identification and does not consider the influence of environmental factors (temperature, light, etc.) on the quantitative results in the process of chlorophyll content estimation, resulting in low accuracy of chlorophyll content estimation. Summary of the Invention
[0005] In view of the shortcomings of related technologies, the present invention provides a method for identifying algae and measuring chlorophyll content in water, aiming to achieve rapid identification of different algae and accurate measurement of chlorophyll content in water.
[0006] This invention provides a method for identifying algae and measuring chlorophyll content in water, comprising the following steps: S1. Establish a fluorescent fingerprint database for algae, specifically including: S11. Obtain samples of multiple algae; S12. Measure the fluorescence intensity of each algal sample at a preset receiving wavelength under various preset excitation wavelengths, and calculate the fluorescence intensity ratio of each algal sample under different preset excitation wavelengths to form the fluorescence fingerprint spectrum of each algae, thereby establishing a fluorescence fingerprint database of algae. S2. Training the neural network model, specifically including: S21. Obtain multiple experimental water samples for each type of algae; S22. Measure the fluorescence intensity of algae in each experimental water sample at a preset receiving wavelength under various preset excitation wavelengths, and measure the temperature, photosynthetically active radiation and chlorophyll a concentration of each experimental water sample. S23. Using the fluorescence intensity, temperature, and photosynthetically active radiation of all experimental water samples of each algae under various preset excitation wavelengths as input parameters and chlorophyll a concentration as output parameter, train a multilayer perceptron model to obtain a chlorophyll a concentration prediction model for each algae. S3. Perform in-situ measurements, specifically including... S31. Measure the fluorescence intensity of mixed algae in the water at a preset receiving wavelength under various preset excitation wavelengths, and measure the water temperature and photosynthetically active radiation. S32. Calculate the ratio of fluorescence intensity of mixed algae in the water body under different preset excitation wavelengths, match it with the fluorescence fingerprint spectrum in the fluorescence fingerprint database of algae, and use the least squares method to obtain the proportion of each algae in the water body. S33. Using the fluorescence fingerprint database of algae, the fluorescence intensity of mixed algae in the water body at each preset excitation wavelength is allocated to the fluorescence intensity of each algae at each preset excitation wavelength. S34. Input the fluorescence intensity, temperature and photosynthetically active radiation of each algae in the water at each preset excitation wavelength into the corresponding chlorophyll a concentration prediction model to obtain the chlorophyll a concentration of each algae in the water.
[0007] In some embodiments, the number of algae is three, namely, cyanobacteria, green algae and red algae; in step S11, obtaining algae samples includes inoculating algae in a sterile culture medium made of artificial seawater and culturing them under preset conditions in a constant temperature incubator.
[0008] In some embodiments, in step S12, the fluorescence intensity of each algal sample at a preset receiving wavelength is measured at three preset excitation wavelengths. , , ,in, Represents algae; when respectively , and When indicated, they represent cyanobacteria, green algae, and red algae, respectively. The fluorescence intensity ratio of each algal sample under different preset excitation wavelengths was calculated according to equation (1). , , To form the fluorescence fingerprint spectrum of each algae This will lead to the establishment of a fluorescent fingerprint database for algae. , expressed as equation (2); (1); (2).
[0009] In some embodiments, in step S21, obtaining multiple experimental water samples for each algae includes diluting each algae sample by different factors to obtain multiple water samples, and conducting stress experiments on each water sample under different temperatures and light conditions to obtain multiple experimental water samples with different chlorophyll a concentrations, temperatures, and photosynthetically active radiation.
[0010] In some embodiments, in step S22, the fluorescence intensity of algae in each experimental water sample at a preset receiving wavelength is measured at three preset excitation wavelengths. , , And measure the temperature of each experimental water sample. Photosynthetically active radiation And chlorophyll a concentration, thus forming the original dataset; In step S23, before training the multilayer perceptron model, the input parameters and output parameters are normalized according to equation (3) and mapped to... The interval, where, This represents the original data value of an input or output parameter before normalization. and These represent the maximum and minimum values of the input or output parameter in the original dataset, respectively. This represents the normalized data value of the input or output parameter. and These represent the upper and lower limits of the mapping interval, respectively. (3); In step S23, the chlorophyll a concentration prediction model for each algae obtained after training is expressed as equation (4), where, For the predicted normalized chlorophyll a concentration, For the output layer linear activation function of the multilayer perceptron model, This is the Sigmoid activation function for the hidden layers of a multilayer perceptron model. These are the normalized input parameters. and These are the weight matrix and bias vector from the input layer to the hidden layer of the multilayer perceptron model, respectively. and These are the weight matrix and bias vector from the hidden layer to the output layer of the multilayer perceptron model, respectively. (4).
[0011] In some embodiments, in step S31, the fluorescence intensity of mixed algae in the water at a preset receiving wavelength is measured at three preset excitation wavelengths. , , and measure water temperature and photosynthetically active radiation ; In step S32, the fluorescence intensity ratio of mixed algae in the water at three preset excitation wavelengths is calculated according to equation (5). According to equation (6), the proportion of cyanobacteria in the water body is obtained by the least squares method. The proportion of green algae The proportion of red algae in the spectrum ; (5); (6).
[0012] In some embodiments, in step S33, the fluorescence intensity of cyanobacteria, green algae, and red algae in the water at a third preset excitation wavelength is calculated according to equation (7). , , ; Then, the fluorescence intensity of cyanobacteria in the water at the first and second preset excitation wavelengths is calculated according to equation (8). and Fluorescence intensity of green algae at the first and second preset excitation wavelengths and Fluorescence intensity of red algae at the first and second preset excitation wavelengths and ; (7); (8).
[0013] In some embodiments, in step S34, the fluorescence intensity of each algae in the water body at each preset excitation wavelength obtained in step S33, the water temperature and photosynthetically active radiation measured in step S31 are first normalized, and then input into the corresponding chlorophyll a concentration prediction model to obtain the normalized chlorophyll a concentration of the algae. Then, the normalized chlorophyll a concentration is reverse normalized to obtain the chlorophyll a concentration of the algae in the water body.
[0014] In some embodiments, a multi-channel fluorescence sensor is used to measure fluorescence intensity at various preset excitation wavelengths in steps S12, S22, and S31; the excitation wavelength range of the multi-channel fluorescence sensor is 410nm to 580nm, and the receiving wavelength of the multi-channel fluorescence sensor is 685nm.
[0015] Based on the above technical solutions, the method for identifying algae and measuring chlorophyll content in water in this embodiment of the invention achieves rapid and accurate identification of different algae and precise measurement of chlorophyll content in water by employing multi-channel fluorescence measurement, constructing an algal fluorescence fingerprint database, training a neural network model, and processing the fluorescence signals of mixed algae in the water measured in situ using the algal fluorescence fingerprint database. It can also output the relative abundance of each algae in the water and the chlorophyll a concentration of each algae, thus achieving a dual characterization of algal community structure and algal biomass in the water. Attached Figure Description
[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings: Figure 1 This is a flowchart of the method for identifying algae and measuring chlorophyll content in water bodies according to the present invention. Detailed Implementation
[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0018] In the description of this invention, it should be understood that the terms "center", "lateral", "longitudinal", "upper", "lower", "top", "bottom", "inner", "outer", "left", "right", "front", "rear", "vertical", "horizontal", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.
[0019] The terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature.
[0020] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0021] refer to Figure 1 As shown, the present invention provides a method for identifying algae and measuring chlorophyll content in water bodies. The method mainly includes three steps: establishing a fluorescent fingerprint database of algae (step S1), training a neural network model (step S2), and performing in-situ measurements (step S3). Steps S1 and S2 are performed in the laboratory modeling stage, and step S3 is performed in the field in-situ measurement stage.
[0022] Step S1 of establishing a fluorescent fingerprint database of algae specifically includes the following steps: Step S11: Obtain multiple algal samples. Specifically, obtaining algal samples includes inoculating algae in a sterile culture medium made from artificial seawater and culturing them under preset conditions in a constant temperature incubator. The preset conditions include light intensity, light-dark time ratio, culture temperature, and culture period; specifically, a white fluorescent lamp is used as the light source, and the light intensity range of the white fluorescent lamp is controllable from 5000 lx to 50000 lx. During the culture of the algal samples, the light intensity is 5600 lx, the light-dark time ratio is 12h:12h, the culture temperature is 25℃, and the culture time is more than three days to ensure that the algal samples reach a good activity state.
[0023] Step S12: Measure the fluorescence intensity of each algal sample at a preset receiving wavelength under various preset excitation wavelengths, and calculate the fluorescence intensity ratio of each algal sample under different preset excitation wavelengths to form the fluorescence fingerprint spectrum of each algae, thereby establishing a fluorescence fingerprint database of algae.
[0024] Step S2 in training the neural network model specifically includes the following steps: Step S21: Obtain multiple experimental water samples for each type of algae. Specifically, the experimental water samples can be obtained using the algae samples after completing step S1; each algae sample is diluted by different multiples to obtain multiple water samples, showing that the chlorophyll content in the multiple water samples is different; then the water samples are placed in a constant temperature water bath, and stress experiments are conducted on the water samples under different temperatures and light irradiance using the constant temperature water bath and light source to obtain multiple experimental water samples with different chlorophyll a concentrations, temperatures, and photosynthetically active radiation.
[0025] To further explain, in the temperature and light stress experiments, white fluorescent lamps were used as the light source, and the light intensity could be controlled at 5600 lx, 11200 lx, 16800 lx, 22400 lx, 28000 lx, 33600 lx, 39200 lx, and 44800 lx, for a total of eight culture light intensities. A constant-temperature water bath was used to control the culture temperature, with six temperature gradients: 5℃, 10℃, 15℃, 25℃, 35℃, and 45℃. Temperature was increased or decreased at each light intensity, and stress was applied for 2 hours at each light intensity-temperature point. Those skilled in the art will understand that the set light intensity and temperature are not the actual light intensity and temperature of the water sample; their purpose is to generate randomized scatter points for calibration. The temperature and light intensity data used for calibration need to be measured using temperature sensors and irradiance sensors.
[0026] Step S22: Measure the fluorescence intensity of algae in each experimental water sample at a preset receiving wavelength under various preset excitation wavelengths, and measure the temperature, photosynthetically active radiation and chlorophyll a concentration of each experimental water sample, thereby forming the original dataset of all experimental water samples for each algae.
[0027] Step S23: Using the fluorescence intensity, temperature, and photosynthetically active radiation of all experimental water samples of each algae under various preset excitation wavelengths as input parameters and chlorophyll a concentration as output parameter, train a multilayer perceptron model to obtain a chlorophyll a concentration prediction model for each algae.
[0028] To further explain, the Multilayer Perceptron (MLP) model is a neural network model used to map multiple input datasets to a single output dataset. During model training, the gradient descent algorithm combined with the backpropagation mechanism is used to iteratively optimize the network weights and biases until the model converges or reaches the preset number of training rounds.
[0029] Step S3 for in-situ measurement specifically includes the following steps: Step S31: Collect data at the test site in the lake, river or ocean. It is understood that the actual test water contains multiple algae. Measure the fluorescence intensity of the mixed algae in the water at the preset receiving wavelength under various preset excitation wavelengths, and measure the water temperature and photosynthetically active radiation.
[0030] Step S32: Calculate the ratio of fluorescence intensity of mixed algae in the water body under different preset excitation wavelengths, match it with the fluorescence fingerprint spectrum in the fluorescence fingerprint database of algae, and use the least squares method to obtain the composition ratio of algae in the water body, that is, the proportion of each type of algae.
[0031] Step S33: Using the fluorescence fingerprint database of algae, the fluorescence intensity of mixed algae in the water at each preset excitation wavelength is allocated to the fluorescence intensity of each algae at each preset excitation wavelength.
[0032] Step S34: Input the fluorescence intensity, temperature and photosynthetically active radiation of each algae in the water at each preset excitation wavelength into the corresponding chlorophyll a concentration prediction model to obtain the chlorophyll a concentration of each algae in the water; the sum of the chlorophyll a concentrations of each algae in the water is the chlorophyll content in the water.
[0033] The above illustrative embodiment calculates the fluorescence intensity ratio of each algae at different excitation wavelengths to construct a fluorescence fingerprint spectrum for each algae, thereby establishing a fluorescence fingerprint database for algae. Through training a neural network model, the influence of temperature and light intensity on the measurement results can be corrected, improving the accuracy of chlorophyll a concentration prediction and obtaining a chlorophyll a concentration prediction model for each algae. During in-situ measurement, the proportion of each algae in the water body is obtained by combining the fluorescence fingerprint database of algae, and the fluorescence intensity of mixed algae in the water body is allocated to each algae. Then, the chlorophyll a concentration of each algae in the water body is obtained using the chlorophyll a concentration prediction model of each algae. This enables rapid and accurate identification of different algae in the water body and accurate measurement of the chlorophyll a concentration of each algae in the water body and the total chlorophyll content in the water body.
[0034] In some embodiments, algae classification can adopt the evolutionary lineage classification method proposed by Falkowski et al. (2004), setting the number of algae to three: cyanobacteria, green algae, and red algae. These three lineages exhibit significant differences in pigment composition. It should be noted that one or more phyla can be selected within each lineage to represent it, and the selection of phyla can refer to the target algae species in the water body being tested. This embodiment clearly defines the algal group classification method to achieve the differentiation of the three main algal groups: cyanobacteria, green algae, and red algae.
[0035] In some embodiments, in step S12, three preset excitation wavelengths are set, and the fluorescence intensity of each algal sample at a preset receiving wavelength is measured at each of the three preset excitation wavelengths, and recorded as follows: , , ; where the subscript is Represents algae; when respectively , and When indicated, they represent cyanobacteria, green algae, and red algae, respectively; the subscripts 1, 2, and 3 represent the numbers of the three preset excitation wavelengths.
[0036] The fluorescence intensity ratio of each algal sample under different preset excitation wavelengths was calculated according to equation (1). , , These three ratio characteristics are combined to form the fluorescence fingerprint spectrum of each algae. The fluorescence fingerprint spectrum is independent of temperature, light intensity, and chlorophyll concentration. Based on this, characteristic matrices of various algae were plotted, and a fluorescence fingerprint database of algae was established. , expressed as equation (2); (1); (2).
[0037] The above illustrative embodiments refine the method for constructing a fluorescent fingerprint database of algae, thereby distinguishing the three major groups: cyanobacteria, green algae, and red algae.
[0038] In some embodiments, in step S22, the fluorescence intensity of algae in each experimental water sample at a preset receiving wavelength is measured at three preset excitation wavelengths. , , And measure the temperature of each experimental water sample. Photosynthetically active radiation And chlorophyll a concentration, thereby forming the original dataset of all experimental water samples for each algae; it is understood that the settings of the three preset excitation wavelengths and one preset receiving wavelength in step S22 are consistent with the settings in step S12.
[0039] In step S23, before training the multilayer perceptron model, the data in the original dataset needs to be normalized. Specifically, the input parameters () are normalized according to equation (3). , , , , The output parameters (chlorophyll a concentration) and other parameters were normalized and mapped to... The interval, where, This represents the original data value of a certain input or output parameter before normalization. and These represent the maximum and minimum values of the input or output parameter in the original dataset, respectively. This represents the normalized data value of the input or output parameter. and These represent the upper and lower limits of the mapping interval, respectively; the normalized input and output parameters will be used to train the multilayer perceptron model. (3).
[0040] In step S23, the chlorophyll a concentration prediction model for each algae obtained after training is expressed as equation (4), where, For the predicted normalized chlorophyll a concentration, For the output layer linear activation function of the multilayer perceptron model, This is the Sigmoid activation function for the hidden layers of a multilayer perceptron model. These are the normalized input parameters. and These are the weight matrix and bias vector from the input layer to the hidden layer of the multilayer perceptron model, respectively. and These are the weight matrix and bias vector from the hidden layer to the output layer of the multilayer perceptron model, respectively; from this, chlorophyll a concentration prediction models for cyanobacteria, green algae, and red algae are obtained, respectively. (4).
[0041] Further explaining step S23, the multilayer perceptron model includes an input layer, a hidden layer, and an output layer; in this embodiment, the input layer has 5 nodes, corresponding to... , , , , The hidden layer is set to one layer to balance model complexity and training efficiency; the hidden layer can have eight nodes to ensure good fitting ability while avoiding overfitting; the hidden layer nodes use the Sigmoid function, which has good nonlinear mapping ability and is suitable for modeling complex relationships of fluorescence signals; the output layer has one node, corresponding to the chlorophyll a concentration. A linear activation function is used to ensure continuous real-valued outputs, making it suitable for regression tasks. When training the multilayer perceptron model, the original dataset is divided into three parts: 70% for model training, 15% for model validation, and 15% for model testing. Gradient descent combined with backpropagation is used for iterative optimization of network weights and biases. Mean squared error (MSE) is used as the loss function. Hyperparameters are set, with the initial learning rate set to 0.001, which can be dynamically adjusted based on the validation set error during training. The batch size is set to 16. The number of training epochs is set to 500, and early stopping is used to prevent overfitting; training is terminated when the validation set error does not decrease for 10 consecutive epochs. After model training, the test set is used to evaluate model performance, with key metrics including the coefficient of determination (COP). The methods used include root mean square error (RMSE) and mean absolute error (MAE). Therefore, by employing machine learning techniques such as multilayer perceptron, gradient descent algorithm, data normalization, and model evaluation, a neural network model is trained to achieve nonlinear fitting of chlorophyll a concentration, thus forming a complete method for accurately retrieving chlorophyll a concentration.
[0042] In some embodiments, in step S31, the fluorescence intensity of mixed algae in the water at a preset receiving wavelength is measured at three preset excitation wavelengths. , , and measure water temperature and photosynthetically active radiation It is understandable that the settings of the three preset excitation wavelengths and one preset receiving wavelength in step S31 are consistent with those in steps S12 and S22. The mixed algae in the water body consists of cyanobacteria, green algae and red algae.
[0043] In step S32, the fluorescence intensity ratio of mixed algae in the water at three preset excitation wavelengths is calculated according to equation (5). According to equation (6), the fluorescence intensity ratio is... Fluorescent fingerprint database of algae By combining these methods, the proportion of cyanobacterial species in the water body can be obtained using the least squares method. The proportion of green algae The proportion of red algae in the spectrum Using the least squares method to find , , The specific method for calculating the optimal solution is well known to those skilled in the art and will not be elaborated here. (5); (6).
[0044] The above illustrative embodiment obtains the proportion of each algae in the water by matching the fluorescence intensity ratio of mixed algae in the water with the fluorescence fingerprint spectrum in the algal fluorescence fingerprint database and using the least squares method, that is, the relative abundance data of each algae in the water.
[0045] In some embodiments, the fluorescence signal generated by the mixed algae in the water at the excitation wavelength is the result of the combined action of cyanobacteria, green algae, and red algae in the water. Therefore, the fluorescence intensity of the mixed algae in the water at each preset excitation wavelength is the sum of the fluorescence intensities of the cyanobacteria, green algae, and red algae in the water at that preset excitation wavelength. In step S33, the fluorescence intensity of the cyanobacteria, green algae, and red algae in the water at the third preset excitation wavelength is calculated according to equation (7). , , ; make , , Then equation (7) can be expressed as equation (71); by solving the matrix That is, to obtain , , ; (7); (71); Then, using the fluorescence fingerprint spectrum of each algae, the fluorescence intensity of the cyanobacterial species in the water at the first and second preset excitation wavelengths is calculated according to Equation (8). and Fluorescence intensity of green algae at the first and second preset excitation wavelengths and Fluorescence intensity of red algae at the first and second preset excitation wavelengths and This allows us to obtain the fluorescence intensity of cyanobacteria in the water at three preset excitation wavelengths. , , ), fluorescence intensity of green algae in water at three preset excitation wavelengths ( , , ), fluorescence intensity of red algae in water at three preset excitation wavelengths ( , , ); (8).
[0046] It should be noted that this embodiment first calculates the fluorescence intensity of each algae in the water at the third preset excitation wavelength, but it is not limited to this. Based on the same idea, the fluorescence intensity of each algae in the water at the first or second preset excitation wavelength can also be calculated first.
[0047] The above illustrative embodiment utilizes an established fluorescence fingerprint database of algae to allocate the fluorescence intensity of mixed algae in the water at each preset excitation wavelength to the fluorescence intensity of each algae at each preset excitation wavelength, thereby achieving rapid and accurate allocation of the fluorescence signal of mixed algae among the algae.
[0048] In some embodiments, in step S34, the fluorescence intensity of each algae in the water body at each preset excitation wavelength obtained in step S33, the water temperature and photosynthetically active radiation obtained in step S31 are first normalized, and then input into the chlorophyll a concentration prediction model corresponding to the algae to obtain the normalized chlorophyll a concentration of the algae. Then, the normalized chlorophyll a concentration is reverse normalized to obtain the chlorophyll a concentration of the algae in the water body, thereby achieving accurate inversion of chlorophyll a concentration. In this way, the chlorophyll a concentrations of cyanobacteria, green algae and red algae in the water body are obtained respectively, and the sum of the chlorophyll a concentrations of each algae is the chlorophyll content in the water body.
[0049] Taking the calculation of chlorophyll a concentration of cyanobacteria in water as an example, this further illustrates the point: First, use equation (3) to calculate the fluorescence intensity of cyanobacteria in water at three preset excitation wavelengths. , , Water temperature and photosynthetically active radiation Perform normalization to obtain the normalized result. , , , and The chlorophyll a concentration of the cyanobacteria is input into the trained chlorophyll a concentration prediction model of the cyanobacteria series shown in equation (4) to obtain the normalized chlorophyll a concentration of the cyanobacteria series; then, the normalized chlorophyll a concentration of the cyanobacteria series is reversed using equation (3) to restore it to the actual chlorophyll a concentration, thereby obtaining the chlorophyll a concentration of the cyanobacteria series in the water body. Following this step, the chlorophyll a concentration of the green algae series and the red algae series in the water body are obtained respectively, thereby accurately obtaining the chlorophyll a concentration of each algae in the water body.
[0050] In some embodiments, a multi-channel fluorescence sensor is used to measure fluorescence intensity at various preset excitation wavelengths in steps S12, S22, and S31. The excitation wavelength range of the multi-channel fluorescence sensor is 410 nm to 580 nm, and the various preset excitation wavelengths are selected within this range. The receiving wavelength of the multi-channel fluorescence sensor is 685 nm. It can be understood that if there are three preset excitation wavelengths, the fluorescence sensor can be designed as a three-channel fluorescence sensor. Referring to "GB17378.7-2007 Marine Monitoring Standard Part 7 Nearshore Pollution Ecological Investigation and Biological Monitoring", a fluorescence spectrophotometer is used to measure the chlorophyll a concentration in step S22. (unit The photosynthetically active radiation (PAR) in steps S22 and S31 is measured using an irradiance sensor, and the temperature in steps S22 and S31 is measured using a temperature sensor. During measurement, the multi-channel fluorescence sensor, irradiance sensor, and temperature sensor should be at the same detection position. The stability of the irradiance sensor is 0.05. The temperature sensor has an accuracy of 0.01℃.
[0051] Through the description of several embodiments of the method for identifying algae and measuring chlorophyll content in water according to the present invention, it can be seen that the present invention has at least one or more of the following advantages: 1) Construct a unique fluorescence fingerprint spectrum for each algae by using the fluorescence intensity ratio under multiple channels / multiple excitation wavelengths, and establish a fluorescence fingerprint database for algae to achieve accurate differentiation of algae categories; 2) A neural network model is used to correct the influence of temperature and light intensity on the measurement results, thereby improving the accuracy of chlorophyll a concentration prediction; 3) By measuring and fusing fluorescence intensity under multiple channels / multi-excitation wavelengths, errors caused by single excitation wavelength drift can be reduced, and anti-interference ability can be improved; 4) By employing multi-channel fluorescence measurement, constructing an algal fluorescence fingerprint database, training a neural network model, and using the algal fluorescence fingerprint database to process the mixed algal fluorescence signals in the water body measured in situ, rapid and accurate identification of different algae in the water body and precise measurement of chlorophyll content were achieved. 5) This method can be applied to real-time in-situ measurement to realize real-time online monitoring of algal community structure in water bodies, and can output the relative abundance of each algae in the water body and the chlorophyll a concentration of each algae, thus realizing the dual characterization of algal community structure and algal biomass in water bodies.
[0052] Finally, it should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.
[0053] The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications can still be made to the specific implementation of the present invention or equivalent substitutions can be made to some technical features without departing from the spirit of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the technical solutions claimed in the present invention.
Claims
1. A method for algae identification and chlorophyll content measurement in a water body, characterized in that, Includes the following steps: S1. Establish a fluorescent fingerprint database for algae, specifically including: S11. Obtain samples of multiple algae; S12. Measure the fluorescence intensity of each algal sample at a preset receiving wavelength under multiple preset excitation wavelengths, and calculate the fluorescence intensity ratio of each algal sample under different preset excitation wavelengths to form the fluorescence fingerprint spectrum of each algae, thereby establishing a fluorescence fingerprint database of algae. S2. Training the neural network model, specifically including: S21. Obtain multiple experimental water samples for each type of algae; S22. Measure the fluorescence intensity of algae in each experimental water sample at a preset receiving wavelength under the various preset excitation wavelengths, and measure the temperature, photosynthetically active radiation and chlorophyll a concentration of each experimental water sample. S23. Using the fluorescence intensity, temperature, and photosynthetically active radiation of all experimental water samples of each algae under various preset excitation wavelengths as input parameters and chlorophyll a concentration as output parameter, train a multilayer perceptron model to obtain a chlorophyll a concentration prediction model for each algae. S3. Perform in-situ measurements, specifically including... S31. Measure the fluorescence intensity of mixed algae in the water body at the preset receiving wavelength under the various preset excitation wavelengths, and measure the temperature and photosynthetically active radiation of the water body; S32. Calculate the ratio of fluorescence intensity of mixed algae in the water body under different preset excitation wavelengths, match it with the fluorescence fingerprint spectrum in the fluorescence fingerprint database of the algae, and use the least squares method to obtain the proportion of each algae in the water body. S33. Using the fluorescence fingerprint database of the algae, the fluorescence intensity of the mixed algae in the water body at each preset excitation wavelength is allocated to the fluorescence intensity of each algae at each preset excitation wavelength. S34. Input the fluorescence intensity, temperature and photosynthetically active radiation of each algae in the water body at each preset excitation wavelength into the corresponding chlorophyll a concentration prediction model to obtain the chlorophyll a concentration of each algae in the water body.
2. The method of claim 1, wherein the step of identifying the algae in the water body and measuring the chlorophyll content is performed by using a portable device. The number of algae is three, namely, cyanobacteria, green algae and red algae; in step S11, the algae sample is obtained by inoculating the algae in a sterile culture medium made of artificial seawater and culturing it under preset conditions in a constant temperature incubator.
3. The method for identifying algae and measuring chlorophyll content in water according to claim 2, characterized in that, In step S12, the fluorescence intensity of each algal sample at a preset receiving wavelength is measured at three preset excitation wavelengths. , , ,in, Represents algae; when respectively , and When indicated, they represent cyanobacteria, green algae, and red algae, respectively. The fluorescence intensity ratio of each of the algal samples at different pre-set excitation wavelengths is calculated according to formula (1) , , to form a fluorescence fingerprint of each of the algae , and further to establish a fluorescence fingerprint database of the algae , represented as formula (2); (1)； (2)。 4. The method of claim 3, wherein the step of identifying the algae in the water body and measuring the chlorophyll content is performed by using the method according to any one of claims 1 to 2. In step S21, obtaining multiple experimental water samples for each type of algae includes diluting each algae sample by different factors to obtain multiple water samples, and conducting stress experiments on each water sample under different temperatures and light conditions to obtain multiple experimental water samples with different chlorophyll a concentrations, temperatures, and photosynthetically active radiation.
5. The method for identifying algae and measuring chlorophyll content in water according to claim 4, characterized in that, In step S22, the fluorescence intensity of the algae in each experimental water sample at the preset receiving wavelength is measured under the three preset excitation wavelengths respectively , , , and the temperature , photosynthetically active radiation and chlorophyll a concentration of each of the experimental water samples are measured, thereby forming a raw data set; In step S23, before training the multi-layer perceptron model, the input parameters and the output parameters are normalized according to formula (3) respectively, which are mapped to interval, wherein, represents the original data value of the input parameter or the output parameter before normalization, and respectively represent the maximum value and the minimum value of the input parameter or the output parameter in the original data set, represents the data value of the input parameter or the output parameter after normalization, and respectively represent the upper and lower limit values of the mapping interval; (3)； In step S23, each of the chlorophyll a concentration prediction models of the algae obtained after the training is expressed as formula (4), wherein, is the predicted normalized chlorophyll a concentration, is a linear activation function of the output layer of the multi-layer perception model, is a Sigmoid activation function of the hidden layer of the multi-layer perception model, is the normalized input parameter, and are a weight matrix and a bias vector of the input layer to the hidden layer of the multi-layer perception model, respectively, and are a weight matrix and a bias vector of the hidden layer to the output layer of the multi-layer perception model, respectively. (4)。 6. The method for identifying algae and measuring chlorophyll content in water according to claim 5, characterized in that, In step S31, the fluorescence intensity of the mixed algae in the water body at the preset receiving wavelength is measured under the three preset excitation wavelengths respectively , , , and the water temperature and the photosynthetically active radiation are measured. In step S32, the fluorescence intensity ratio of the mixed algae in the water body at three preset excitation wavelengths is calculated according to equation (5). According to equation (6), the proportion of cyanobacteria in the water body is obtained by the least squares method. The proportion of green algae The proportion of red algae in the spectrum ; (5)； (6)。 7. The method for identifying algae and measuring chlorophyll content in water according to claim 5, characterized in that, In step S33, the fluorescence intensity of cyanobacteria, green algae, and red algae in the water body at the third preset excitation wavelength is calculated according to formula (7). , , ; Then, the fluorescence intensity of the cyanobacteria in the water body at the first and second preset excitation wavelengths is calculated according to equation (8). and Fluorescence intensity of green algae at the first and second preset excitation wavelengths and Fluorescence intensity of red algae at the first and second preset excitation wavelengths and ; (7)； (8)。 8. The method for identifying algae and measuring chlorophyll content in water according to claim 7, characterized in that, In step S34, the fluorescence intensity of each algae in the water body obtained in step S33 under various preset excitation wavelengths, the water temperature and photosynthetically active radiation measured in step S31 are first normalized, and then input into the corresponding chlorophyll a concentration prediction model to obtain the normalized chlorophyll a concentration of the algae. Then, the normalized chlorophyll a concentration is reverse normalized to obtain the chlorophyll a concentration of the algae in the water body.
9. The method for identifying algae and measuring chlorophyll content in water according to claim 1, characterized in that, A multi-channel fluorescence sensor is used to measure fluorescence intensity at various preset excitation wavelengths in steps S12, S22, and S31; the excitation wavelength range of the multi-channel fluorescence sensor is 410nm to 580nm, and the receiving wavelength of the multi-channel fluorescence sensor is 685nm.