Ammonia nitrogen detection method in water based on polarization spectral feature fusion

Abstract

The present invention provides a method for detecting ammonia nitrogen in water by fusion of polarization spectrum characteristics, which is used in a polarization spectrum characteristic fusion system, wherein the system includes a light source, a reflector, a quarter wave plate, a linear polarizer, a fiber lens and a fiber spectrometer; the method includes: C1: collecting polarization modulation spectrum based on a target water sample through the system; C2: demodulating four Stokes parameter spectra of the reflected light wave of the target water sample; C3: selecting different numbers of polarization modulation spectra and Stokes parameter spectra for spectral characteristic fusion respectively; C4: performing spectral classification based on characteristic fusion spectra of target water samples with different ammonia nitrogen concentrations. The present invention realizes the detection of ammonia nitrogen in water based on polarization characteristic fusion spectrum, and accurately detects the ammonia nitrogen concentration in water through the interaction between light and matter without using auxiliary reagents.

Description

Technical Field

[0001] The present invention relates to the technical field of polarization spectrum water quality detection, and in particular to a method for detecting ammonia nitrogen in water by integrating polarization spectrum characteristics. Background Art

[0002] Water is not only the source of life, but also a necessity for life. However, water shortage is an increasingly serious environmental problem worldwide. As we all know, water shortage is mainly caused by water pollution. As expected, water pollution usually comes from our daily lives, including but not limited to garbage dumps and unsanitary landfills, swine wastewater, agriculture and mining activities. Therefore, these conscious behaviors and other unconscious behaviors bring various pollutants into the water. However, more and more research is focused on some major pollutants, such as microplastics, heavy metals, petroleum hydrocarbons and antibiotics. In terms of chemical composition, ammonia nitrogen has attracted continuous attention from all over the world due to its ubiquitous presence and high toxicity in surface water. The oxidation of ammonia nitrogen can reduce the concentration of dissolved oxygen in water, making the water black and smelly, and the deterioration of water quality will further affect the survival of aquatic plants and animals. The sustainable use of water resources is inseparable from water pollution detection.

[0003] In order to improve water quality and save lives, it is crucial and urgent to detect pollutants in water for further treatment. In recent years, the detection of ammonia nitrogen mainly includes optical, electrochemical and bioenzymatic methods. In terms of optical methods focusing on the optical properties of ammonia nitrogen, sensing strategies based on spectrophotometry, fluorescence and optical fiber have been widely developed and applied worldwide.

[0004] Initially, spectrophotometry used Berthelot's reagent, an alkaline solution of phenol and hypochlorite, with a mixed green indicator to detect ammonia. Berthelot's reaction was then improved to optimize the reaction conditions, mainly by replacing toxic phenol with salicylate, changing the substituents of salicylate, and evaporating dissolved ammonia. In addition, ammonia can react with fluorescent reagents to form fluorescent compounds, which absorb external energy and emit light with a certain wavelength and concentration-dependent intensity. However, both spectrophotometry and fluorescence methods require auxiliary reagents to quantify ammonia, which are complicated to operate, have limited reaction conditions, and may cause secondary pollution.

[0005] As a novel approach, fiber optic detection utilizes optical fibers to transmit light that is sensitive to the external environment. Usually ammonia reacts with the coating material of the optical fiber, first changing the refractive index and then modulating the transmitted light. The modulated light is detected and further demodulated to quantify the ammonia. However, changes in the refractive index or poor adsorption of the coating material may reduce the detection sensitivity at low concentrations. At the same time, these existing optical methods require ammonia to react with a specified substance, which limits their application in practical environments and long-term real-time monitoring.

[0006] The present invention proposes a polarization-inspired feature fusion spectroscopy (PIFFS) method to simplify ammonia nitrogen detection and avoid secondary pollution; the PIFFS system is mainly composed of a light source, a reflector, a quarter wave plate (QWP), a linear polarizer (LP), a fiber lens and a fiber spectrometer; in the spectral domain, the PIFFS system can effectively acquire the visible spectrum. Summary of the invention

[0007] To solve the above problems, the present invention provides a method for detecting ammonia nitrogen in water by fusing polarization spectral features. In the polarization domain, the light emitted by the light source is first changed by the water sample, then modulated by the polarization optical element, and finally received by the fiber spectrometer. Obviously, the critical ammonia concentration is reflected in the light changed by the water sample; the target light is modulated four times by adjusting the QWP and LP angles, and its four Stokes parameters are calculated. Then, the modulated or calculated spectra of ammonia samples of different concentrations can be classified by a machine learning algorithm. The present invention opens up a new way for optical detection of ammonia nitrogen in water by combining polarization and spectroscopy.

[0008] The present invention provides a method for detecting ammonia nitrogen in water by integrating polarization spectrum characteristics. The specific technical scheme is as follows:

[0009] The method is used for a polarization spectrum feature fusion system, which includes a light source, a reflector, a quarter wave plate, a linear polarizer, a fiber lens and a fiber spectrometer;

[0010] The method comprises the following steps:

[0011] C1: Collect polarization modulation spectrum based on target water sample through the polarization spectrum feature fusion system;

[0012] C2: demodulate the four Stokes parameter spectra of the reflected light waves of the target water sample;

[0013] C3: Select different numbers of polarization modulation spectra and Stokes parameter spectra to perform spectral feature fusion respectively;

[0014] C4: Based on the characteristic fusion spectra of target water samples with different ammonia nitrogen concentrations, spectral classification is performed through machine learning algorithm.

[0015] Furthermore, in step C1, collecting polarization modulation spectrum includes the following steps:

[0016] C101: Make the light waves emitted by the set light source irradiate the target water sample to obtain the reflected light waves carrying the ammonia nitrogen information in the water;

[0017] C102: The reflected light wave is sequentially passed through a quarter wave plate and a linear polarizer for polarization modulation;

[0018] C103: The polarization modulated light wave information is focused by the fiber lens into the fiber spectrometer to obtain the polarization modulated spectrum intensity.

[0019] Furthermore, the specific method of polarization modulation is as follows:

[0020] P m =M LP ×M QWP ×P=M total ×P

[0021] Where P = (S0, S1, S2, S3) T is the vector composed of the four Stokes parameters of the light wave incident on the quarter-wave plate, P m =(S′0, s′1, S′2, S′3) T is the vector composed of the four Stokes parameters of the light wave emitted by the linear polarizer, M QWP is the Mueller matrix of the quarter-wave plate, M LP is the Mueller matrix of the linear polarizer, M total The Mueller matrix of the system is a combination of a quarter wave plate and a linear polarizer.

[0022] Furthermore, in step C102, the reflected light wave is polarized four times;

[0023] For the four-order polarization modulation, the corresponding quarter-wave plate fast axis angles are 0 degrees, 0 degrees, 0 degrees and 45 degrees, and the linear polarizer transmission axis angles are 45 degrees, 135 degrees, 0 degrees and 45 degrees, respectively.

[0024] Furthermore, the four Stokes parameters and polarization degree of the target water sample reflected light wave are calculated as follows:

[0025]

[0026] Among them, S0, S1, S2, S3 represent four Stokes parameters, DoP represents the degree of polarization, Represent the light intensity values ​​of four polarization modulations respectively.

[0027] Furthermore, the target water samples include standard ammonia nitrogen water samples and environmental ammonia nitrogen water samples.

[0028] Furthermore, in step C3, the spectral feature fusion selects at least one polarization modulation spectrum, at least one Stokes parameter spectrum, and a combination of at least one polarization modulation spectrum and at least one Stokes parameter spectrum to perform feature fusion.

[0029] Furthermore, the spectral feature fusion also includes fusing at least one light intensity.

[0030] Furthermore, the polarization spectrum feature fusion system has the following specific structure:

[0031] The light source projects the light source onto the sample through a fiber optic ring illuminator;

[0032] The reflector is used to receive the reflected light wave and reflect it onto the quarter wave plate;

[0033] The quarter wave plate re-projects the light wave onto the linear polarizer;

[0034] The linear polarizer emits the polarization-modulated light wave to the optical fiber lens;

[0035] The optical fiber lens is connected to the optical fiber spectrometer.

[0036] The beneficial effects of the present invention are as follows:

[0037] A feature fusion spectroscopy (PIFFS) method was proposed to detect dissolved ammonia in water. The proposed PIFFS system detects dissolved ammonia through non-contact surface detection without the need for auxiliary reagents.

[0038] The PIFFS system has a simple structure and is mainly composed of a light source, a reflector, a quarter wave plate (QWP), a linear polarizer (LP), a fiber lens and a fiber spectrometer. The polarization information is introduced into the ammonia spectrum detection through the modulation of QWP and LP, and the polarization modulated light wave information is focused by the fiber lens into the fiber spectrometer to obtain the polarization modulated spectral intensity, and then the Stokes parameter spectrum of the reflected light wave of the target water sample is demodulated, and the spectral features are fused. Finally, the most accurate machine learning algorithm is selected for classification and identification, ensuring the accuracy and effectiveness of the detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 The figure is a schematic diagram of the PIFFS process for detecting ammonia nitrogen in water according to the present invention.

[0040] Figure 2 This is a schematic diagram of the structure of the PIFFS system for detecting ammonia nitrogen in water according to the present invention.

[0041] Figure 3 This is the average spectrum of four polarization modulations of 29 standard ammonia nitrogen water samples of the present invention.

[0042] Figure 4 This is the average spectrum of four polarization modulations of 11 environmental ammonia nitrogen water samples of the present invention.

[0043] Figure 5 This is a diagram showing the classification results of 29 standard ammonia nitrogen water samples using the fusion of 33 features of the present invention. DETAILED DESCRIPTION

[0044] The following description clearly and completely describes the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] The present invention proposes a method for detecting ammonia nitrogen in water by integrating polarization spectrum characteristics to simplify ammonia nitrogen detection and avoid secondary pollution. The method is applied to the PIFFS system, which mainly consists of a light source, a reflector, a quarter wave plate (QWP), a linear polarizer (LP), a fiber lens and a fiber spectrometer.

[0046] In the spectral domain, the PIFFS system can effectively obtain the visible spectrum. In the polarization domain, the target light is modulated four times by adjusting the QWP and LP angles, and its four Stokes parameters are calculated. In addition, the modulated light intensity and the calculated Stokes parameters are superimposed respectively to obtain 33 fusion features. Referring to the national standard on the limit of ammonia nitrogen in water, 29 concentrations are selected from 0 mg / L to 100 mg / L to classify the ammonia nitrogen in the standard water sample. Various experimental results fully show that superimposing the modulated light intensity inspired by the first Stokes parameter can obtain higher classification accuracy. At the same time, compared with the nearest neighbor discriminant method (k-NN) and random forest (RF), the support vector machine (SVM) shows an overall advantage in classification accuracy and time. When the training-test ratio is as low as 1:1, the overall classification accuracy of 29 concentrations can be close to 100%. The PIFFS method can successfully detect ammonia nitrogen as low as 0.1 mg / L in standard water samples and environmental water samples. Due to the special reaction between light and matter, the PIFFS method can be further applied to the detection of other water pollution. The overall implementation of the method is as follows.

[0047] Example 1

[0048] Embodiment 1 of the present invention discloses a method for detecting ammonia nitrogen in water by fusion of polarization spectrum features, the method being based on the PIFFS system (polarization spectrum feature fusion system);

[0049] In this embodiment, the PIFFS system includes a light source, a reflector, a quarter wave plate, a linear polarizer, a fiber lens and a fiber spectrometer;

[0050] The target water sample to be tested is placed on the sample stage. In this embodiment, the sample stage is composed of a linear translation stage (Thorlabs, LTS300 / M), a YZ translation stage (Thorlabs, LX20YZ / M) and a right-angle bracket (Thorlabs, MT402);

[0051] The sample stage provides free movement of the sample in three-dimensional space.

[0052] Specifically, Figure 2 As shown, the system structure is as follows:

[0053] The light source (Thorlabs, OSL2, 400–1600 nm) illuminated the sample through a fiber optic ring illuminator (Thorlabs, FRI61F50);

[0054] The mirror (Thorlabs, BB2-E02, 400-750 nm) was mounted on a right-angle kinematic mirror frame (Thorlabs, KCB2 / M);

[0055] The quarter-wave plate QWP (Thorlabs, SAQWP05M-700, 325-1100 nm) was mounted on a motorized rotating platform (Thorlabs, KPRM1E / M), and the linear polarizer LP (Thorlabs, LPVISC100-MP2, 510-800 nm) was mounted on a cage rotating base (Thorlabs, CRM1 / M);

[0056] The optical wave focused by the fiber lens (Daheng Optics, GCX-LF6SMA-532, 400-700 nm) was transmitted by the optical fiber (OceanOptics, QP600-1-SR, 200-1100 nm) and received by the fiber spectrometer (Ocean Optics, HR4000CG-UV-NIR, 200-1100 nm).

[0057] The ammonia nitrogen detection method in water is based on the PIFFS system constructed as above. Figure 1 As shown in the figure, water samples with different ammonia concentrations are first prepared, polarization modulation is performed through QWP and LP, Stokes parameter spectra are calculated, and then spectral feature fusion is performed. Finally, the characteristic fusion spectra of ammonia concentrations are classified based on the machine learning algorithm. The specific process is as follows:

[0058] C1: Collect polarization modulation spectrum based on target water sample through the polarization spectrum feature fusion system;

[0059] In this embodiment, the target water samples include standard ammonia nitrogen and environmental ammonia nitrogen water samples, which are specifically prepared as follows:

[0060] According to the Chinese national standard (GB 5749-2006), the ammonia nitrogen in drinking water is limited to no more than 0.5 mg / L. In addition, according to the Chinese national standard (GB / T 18920-2020), the ammonia nitrogen content in toilet flushing and car washing shall not exceed 5 mg / L, and the ammonia nitrogen in urban beautification, street cleaning, fire fighting, construction sites and concrete production shall not exceed 8 mg / L. In order to meet practical applications, these limits and their adjacent values ​​are therefore included in the ammonia nitrogen concentration selected by the present invention. In addition, according to the Chinese national standard (GB5749-2006), COD, nitrate, chromaticity and turbidity in drinking water are limited to less than 3 mg / L, 10 mg / L, 15 Hazen and 1 NTU, respectively.

[0061] When preparing the standard ammonia nitrogen water sample, ammonium chloride was used as the solute and ultrapure water was used as the solvent. The standard water sample specification for each ammonia nitrogen concentration was 50 mL. A total of 29 ammonia nitrogen concentrations were prepared, which were 0 mg / L, 0.1 mg / L, 0.2 mg / L, 0.3 mg / L, 0.4 mg / L, 0.5 mg / L, 0.6 mg / L, 0.7 mg / L, 0.8 mg / L, 0.9 mg / L, 1 mg / L, 2 mg / L, 3 mg / L, 4 mg / L, 5 mg / L, 6 mg / L, 7 mg / L, 8 mg / L, 9 mg / L, 10 mg / L, 20 mg / L, 30 mg / L, 40 mg / L, 50 mg / L, 60 mg / L, 70 mg / L, 80 mg / L, 90 mg / L, and 100 mg / L. For each experimental measurement, 5 mL of water sample was collected using a pipette (Thermo, 100-1000 μL) and placed in a petri dish.

[0062] When preparing environmental ammonia nitrogen water samples, Nongfu Spring drinking water from natural water sources was selected as the solvent. The Nongfu Spring drinking water (specification 5L) used is rich in a variety of natural mineral elements required by the human body, such as metasilicic acid (≥1.8mg / L), potassium (≥0.35mg / L), sodium (≥0.8mg / L), magnesium (≥0.5mg / L), calcium (≥4mg / L), etc. Using ammonium chloride as the solute and Nongfu Spring drinking water as the solvent, ammonia nitrogen water samples were first prepared. The specification of each water sample was 50mL, and the 11 ammonia nitrogen concentrations were 0mg / L, 0.1mg / L, 0.2mg / L, 0.3mg / L, 0.4mg / L, 0.5mg / L, 0.6mg / L, 0.7mg / L, 0.8mg / L, 0.9mg / L, and 1mg / L. Then, standard water samples characterizing comprehensive water quality indicators such as chemical oxygen demand (COD), nitrate, chromaticity, and turbidity were prepared respectively, among which the COD specification was 50mL and the concentration was 30mg / L, the nitrate specification was 50mL and the concentration was 100mg / L, the chromaticity specification was 20mL and the concentration was 150Hazen, and the turbidity specification was 50mL and the concentration was 10NTU. Finally, the water sample measured in each experiment was also 5 mL. A pipette (Thermo, 100-1000 μL) was used to take 4 mL, 0.25 mL, 0.25 mL and 0.25 mL of ammonia nitrogen water sample, COD water sample, nitrate water sample, color water sample and turbidity water sample in a culture dish to obtain environmental ammonia nitrogen water samples. The concentrations of 11 types of ammonia nitrogen were 0 mg / L, 0.08 mg / L, 0.16 mg / L, 0.24 mg / L, 0.32 mg / L, 0.4 mg / L, 0.48 mg / L, 0.56 mg / L, 0.64 mg / L, 0.72 mg / L and 0.8 mg / L, respectively. The COD concentration, nitrate concentration, color concentration and turbidity concentration were 1.5 mg / L, 5 mg / L, 7.5 Hazen and 0.5 NTU, respectively.

[0063] Based on the target water sample prepared as above, collecting polarization modulation spectrum;

[0064] C101: Make the light waves emitted by the set light source irradiate the target water sample to obtain the reflected light waves carrying the ammonia nitrogen information in the water;

[0065] C102: The reflected light wave is sequentially passed through a quarter wave plate and a linear polarizer for polarization modulation;

[0066] C103: The polarization modulated light wave information is focused by the fiber lens into the fiber spectrometer to obtain the polarization modulated spectrum intensity.

[0067] The specific process is as follows:

[0068] The Mueller matrix of the linear delay is as follows:

[0069]

[0070] Wherein, θ (0°≤θ≤180°) is the fast axis angle, which is defined as the angle between the fast axis of the QWP and the horizontal x-axis. δ is the phase delay, and δ=π / 2 (i.e., p=1) indicates that the QWP is a negative crystal. Conversely, δ=-π / 2 (i.e., p=-1) indicates that the QWP is a positive crystal. The QWP used in the present invention is a positive crystal, so p=-1. Therefore, the Mueller matrix of the QWP is shown as follows.

[0071]

[0072] The Mueller matrix of LP is shown as follows:

[0073]

[0074] Here, α (0°≤α≤180°) represents the transmission axis angle of the linear polarizer.

[0075] Therefore, the system Mueller matrix of the QWP and LP combination is as follows:

[0076]

[0077] Assume that the vector composed of the four Stokes parameters of the light wave incident on the QWP is P = (S0, S1, S2, S3) T , then the vector composed of the four Stokes parameters of the light wave emitted by LP is P m =(S′0,S′1,S′2,S′3) T , the polarization modulation mode is expressed as follows:

[0078] P m =M LP ×M QWP ×P=M total ×P

[0079] The fast axis angles of the quarter-wave plate under the four polarization modulations are 0 degrees, 0 degrees, 0 degrees and 45 degrees, respectively, and the transmission axis angles of the linear polarizer are 45 degrees, 135 degrees, 0 degrees and 45 degrees, respectively. Therefore, the system Mueller matrices corresponding to the four polarization modulation angles A1, A2, A3, and A4 are expressed as follows:

[0080]

[0081]

[0082]

[0083]

[0084] C2: demodulate the four Stokes parameter spectra of the reflected light waves of the target water sample;

[0085] The spectrometer only receives the light intensity information after polarization modulation, which is modulated by the first row of the system Mueller matrix. Therefore, the four Stokes parameters (S0, S1, S2, S3) and the degree of polarization (DoP) representing the reflected light wave of the target water sample can be calculated by the following formula:

[0086]

[0087] For each ammonia nitrogen concentration of the standard ammonia nitrogen water sample, 100 spectra were collected under one polarization modulation. Then, the QWP and LP were rotated to continuously collect spectra under four polarization modulations. The fast axis angles of the quarter-wave plate under the four polarization modulations were 0 degrees, 0 degrees, 0 degrees and 45 degrees, respectively, and the transmission axis angles of the linear polarizer were 45 degrees, 135 degrees, 0 degrees and 45 degrees, respectively. Finally, four polarization modulation spectra of 29 standard ammonia nitrogen water samples were measured. The 100 spectra collected for each polarization modulation were averaged, and the polarization modulation average spectrum obtained was as shown in the figure. Figure 3 As shown. Obviously, the spectral peaks decrease slightly in the order of the third, second, fourth and first polarization modulations. Under the same polarization modulation, there is no significant difference in the spectra of the 29 concentrations of ammonia nitrogen standards. Therefore, an appropriate algorithm is needed to classify these spectra.

[0088] Polarization modulation spectra of environmental ammonia nitrogen water samples were also collected, and each set of polarization spectra contained 100 spectra. The polarization modulation average spectrum of environmental ammonia nitrogen water samples is shown in Figure 4 Obviously, the spectrum changes less under the same polarization modulation, which is because the environmental ammonia nitrogen water sample contains a variety of common pollutants and conventional indicators in water.

[0089] C3: Select different numbers of polarization modulation spectra and Stokes parameter spectra to perform spectral feature fusion respectively;

[0090] The collected polarization modulation spectrum and demodulated Stokes parameter spectrum are fused with spectral features respectively. The 33 feature fusion methods are shown in the table below.

[0091] Feature fusion method and its index

[0092]

[0093] C4: Spectral classification is performed based on the characteristic fusion spectra of target water samples with different ammonia nitrogen concentrations.

[0094] 29 standard ammonia nitrogen water samples were classified by fusion of 33 spectral features. 100 spectra were divided into 50 for training and 50 for testing. The classification test results of random forest (RF), nearest neighbor discriminant method (k-NN) and support vector machine (SVM) are shown in Figure 2. Figure 5 As shown. Although the training-test ratio is as low as 1:1, the highest classification accuracy of different algorithms exceeds 99%. For the 33 classification features, the classification accuracy of the last 11 features is slightly higher than that of the middle 18 features, and significantly higher than that of the first 4 features. In other words, the first Stokes parameter can effectively improve the classification accuracy. Extending the concept of the first Stokes parameter, fusing several light intensities can further improve the classification accuracy. Obviously, by fusing the four measured light intensities together, each algorithm obtains the classification accuracy closest to 100%. Among the classification algorithms, SVM shows superiority in both classification accuracy and classification time. This is because SVM is a model trained by maximizing the distance interval, and it performs better on classification problems with small sample sets. However, k-NN cannot form a model, and the test set is identified by the distance calculation of the training set. It is too dependent on the training set and the model is prone to overfitting. RF is an integrated model that obtains multiple sub-models through sampling. However, for small sample sets, the number of features is much larger than the number of samples, which also leads to overfitting of the model.

[0095] It is crucial to classify the polarization spectra of various ammonia concentrations in environmental ammonia nitrogen water samples. The SVM algorithm that performs best in standard ammonia nitrogen water samples is used. At the same time, among the 33 fused spectral features, the 5th, 23rd, and 33rd fused features with increasing classification accuracy in standard ammonia nitrogen water samples are selected. Similarly, 100 spectra are divided into 50 training and 50 testing. Based on these parameters, 11 ammonia nitrogen concentrations in environmental ammonia nitrogen water samples are classified, and the classification accuracy of the training set of the three fused features is 100%, and the classification accuracy of the test set is 99.82%, 100%, and 100%, respectively. Overall, the classification results of environmental ammonia nitrogen water samples are satisfactory. Therefore, the PIFFS method can effectively classify ammonia nitrogen concentrations in environmental water samples.

[0096] As described above, in this embodiment, random forest (RF), nearest neighbor discriminant method (k-NN) and support vector machine (SVM) are comprehensively compared in ammonia classification. Inspired by the first Stokes parameter, the experimental results show that the classification accuracy can be improved by fusing the spectral features received after multiple modulations. In addition, the classified spectra can be sorted and selected based on the RF algorithm to reduce the computational and time costs without losing the classification accuracy.

[0097] The present invention is not limited to the above-mentioned specific embodiments, but extends to any new features or any new combination disclosed in this specification, as well as any new method or process steps or any new combination disclosed.

Claims

1. A method for detecting ammonia nitrogen in water by integrating polarization spectrum characteristics, characterized in that: The method is used for a polarization spectrum feature fusion system, which includes a light source, a reflector, a quarter wave plate, a linear polarizer, a fiber lens and a fiber spectrometer; The method comprises the following steps: C1: Using the polarization spectrum feature fusion system, based on the target water sample, collect polarization modulation spectra. The specific steps are as follows: C101: Make the light waves emitted by the set light source irradiate the target water sample to obtain the reflected light waves carrying the ammonia nitrogen information in the water; C102: performing polarization modulation on the reflected light wave by sequentially passing through a quarter wave plate and a linear polarizer. Specifically, performing polarization modulation on the reflected light wave four times; For the four-order polarization modulation, the corresponding quarter-wave plate fast axis angles are 0 degrees, 0 degrees, 0 degrees and 45 degrees, and the linear polarizer transmission axis angles are 45 degrees, 135 degrees, 0 degrees and 45 degrees, respectively; The specific method of polarization modulation is as follows: P m =M LP ×M QWP ×P=M total ×P Where P = (S0, S1, S2, S3) T is the vector composed of the four Stokes parameters of the light wave incident on the quarter-wave plate, Pm=(S'0, S'1, S'2, S'3) T is the vector composed of the four Stokes parameters of the light wave emitted by the linear polarizer, M QWP is the Mueller matrix of the quarter-wave plate, M LP is the Mueller matrix of the linear polarizer, M total The Mueller matrix of the system is a combination of a quarter wave plate and a linear polarizer; C103: Focus the polarization modulated light wave information through the fiber lens and enter the fiber spectrometer to obtain the polarization modulated spectrum intensity; C2: demodulate the four Stokes parameter spectra of the reflected light waves of the target water sample; C3: Select different numbers of polarization modulation spectra and Stokes parameter spectra to perform spectral feature fusion respectively; C4: Spectral classification is performed based on the characteristic fusion spectra of target water samples with different ammonia nitrogen concentrations.

2. The method for detecting ammonia nitrogen in water according to claim 1, characterized in that: The four Stokes parameters and polarization degree of the target water sample reflected light wave are calculated as follows: Among them, S0, S1, S2, S3 represent four Stokes parameters, DoP represents the degree of polarization, Represent the light intensity values ​​of four polarization modulations respectively.

3. The method for detecting ammonia nitrogen in water according to claim 1, characterized in that: The target water samples include standard ammonia nitrogen water samples and environmental ammonia nitrogen water samples.

4. The method for detecting ammonia nitrogen in water according to claim 1, characterized in that: In step C3, the spectral feature fusion selects at least one polarization modulation spectrum, at least one Stokes parameter spectrum, and a combination of at least one polarization modulation spectrum and at least one Stokes parameter spectrum to perform feature fusion.

5. The method for detecting ammonia nitrogen in water according to claim 4, characterized in that: The spectral feature fusion also includes fusing at least one light intensity.

6. The method for detecting ammonia nitrogen in water according to any one of claims 1 to 5, characterized in that: The specific structure of the polarization spectrum feature fusion system is as follows: The light source projects the light source onto the sample through a fiber optic ring illuminator; The reflector is used to receive the reflected light wave and reflect it onto the quarter wave plate; The quarter wave plate re-projects the light wave onto the linear polarizer; The linear polarizer emits the polarization-modulated light wave to the optical fiber lens; The optical fiber lens is connected to the optical fiber spectrometer.

Images