Cotton canopy nitrogen monitoring method based on optimized nitrogen vertical attenuation coefficient

Abstract

The invention discloses a cotton canopy nitrogen monitoring method based on an optimized nitrogen vertical attenuation coefficient, which comprises the following steps: collecting spectral reflection information of a cotton canopy, synchronously obtaining photosynthetic parameters and different nitrogen components of different vertical layers of the cotton canopy, and establishing a cotton canopy vertical scale nitrogen nutrition monitoring spectrum database; constructing a nitrogen distribution model by combining photosynthetic nitrogen and non-photosynthetic nitrogen; providing an optimized nitrogen vertical attenuation coefficient K based on a nitrogen distribution model; constructing an optimized nitrogen vertical attenuation coefficient K spectrum estimation model; inverting nitrogen contents of different vertical layers of the cotton canopy based on the attenuation coefficient K; according to the method, the limitation that vertical heterogeneity distribution of different nitrogen components of the crop canopy is difficult to accurately invert by using a traditional spectrum technology is broken through, and a new thought is provided for improving estimation precision of vertical scale nitrogen nutrition of the cotton canopy.

Description

Technical Field

[0001] The invention relates to the technical field of crop nitrogen nutrition monitoring, in particular to a cotton canopy vertical scale nitrogen nutrition monitoring method based on optimizing the nitrogen vertical attenuation coefficient. Background Art

[0002] Nitrogen is one of the main nutrients that limit crop growth and yield formation, and is closely related to crop photosynthesis and total primary productivity. Therefore, timely and accurate estimation of crop canopy nitrogen content is the key to achieving nitrogen nutrition diagnosis and scientific fertilization plan formulation. In recent years, the vertical distribution of nitrogen in the crop canopy has attracted much attention from scholars. Optimization theory (OT) shows that nitrogen, as one of the main elements for synthesizing chlorophyll, is closely related to crop photosynthesis. Crops allocate more nitrogen absorbed by the root system to the upper leaves directly exposed to sunlight, so the upper leaves have higher photosynthetic rate and nitrogen use efficiency. When the nitrogen supply is insufficient, nitrogen is transferred from the middle and lower old leaves to the upper young leaves to support the growth and development of the apical tissue and high photosynthetic activity. The leaves in the middle and lower parts of the canopy preferentially show nitrogen deficiency symptoms. Therefore, obtaining information on the vertical distribution of nitrogen in the crop canopy can accurately diagnose the nitrogen abundance and deficiency of crops and provide a decision-making basis for adjusting and optimizing scientific fertilization strategies in farmland.

[0003] Compared with traditional destructive sampling chemical tests to obtain crop nitrogen content, remote sensing technology has been widely used in assessing nitrogen nutrition of crops such as wheat, corn, and rice due to its advantages of high spectral resolution, strong continuity, and rich spectral information. However, since the sensor cannot fully capture the spectral characteristics of the crop canopy, especially the dense leaves in the middle and lower canopy and the complex light energy transmission, the response of nitrogen content in different vertical layers to the in-situ spectral reflectance varies greatly, which greatly limits the accuracy and practicality of diagnosing nitrogen abundance and deficiency in the crop canopy.

[0004] The nitrogen content at the top of the canopy and the cumulative leaf area index follow an exponential function, and the nitrogen vertical attenuation coefficient Kt can well characterize the vertical distribution characteristics of nitrogen in the canopy. When the nitrogen supply is insufficient, the nitrogen content of the leaves in the lower part of the crop canopy decreases sharply, resulting in a larger difference in the vertical scale distribution of nitrogen, and the coefficient Kt is more sensitive to the deficiency of nitrogen nutrition. With the advancement of the crop growth process, especially in the late growth period, the nitrogen in the middle and lower leaves is transported to reproductive organs such as grains, the difference in the vertical scale nitrogen distribution of the canopy becomes larger, and the coefficient Kt gradually increases. Therefore, the vertical attenuation coefficient Kt of nitrogen in different growth periods of crops can accurately diagnose the abundance and deficiency of nitrogen nutrition in the vertical scale of the crop canopy, so as to formulate and optimize the nitrogen fertilizer application strategy. The coefficient Kt can be estimated using spectral technology, which can achieve high-precision remote sensing monitoring of the vertical distribution of nitrogen in the crop canopy.

[0005] It is worth noting that the coefficient Kt is currently constructed based on the nitrogen content per unit area. Due to the high correlation between nitrogen content and chlorophyll, the coefficient Kt can only characterize nitrogen related to photosynthesis. Although studies have found that 75-80% of the total nitrogen in plant leaves is distributed in the photosynthetic chloroplasts, some nitrogen is also distributed in other compounds, such as cell wall proteins and antioxidant enzymes. Therefore, although crop nitrogen nutrition monitoring based on spectral technology has made great progress, the research on the use of spectroscopy to monitor the vertical heterogeneity of different nitrogen components in the crop canopy is still insufficient. In addition, since the traditional spectral estimation coefficient Kt only considers the characteristic bands of photosynthesis-related nitrogen and ignores the spectral characteristics of non-photosynthetic nitrogen, it will limit the stability and universality of the model.

[0006] Therefore, how to provide a cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient that can improve the stability and universality of nitrogen estimation is a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the invention

[0007] In response to the above research status and existing problems, the present invention provides a cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient, combining the basic physiology of the cotton canopy with remote sensing technology, breaking through the limitation of traditional spectral technology that is difficult to accurately invert the vertical heterogeneous distribution of different nitrogen components in the crop canopy, and providing a theoretical basis for improving the estimation accuracy of nitrogen nutrition in the vertical scale of the cotton canopy.

[0008] The present invention provides a cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient, comprising the following steps:

[0009] S1: Collect the spectral reflectance information of cotton canopy, as well as the photosynthetic parameters and physiological parameters of different vertical layers of cotton canopy as a sample database;

[0010] S2: calculating the photosynthetic nitrogen content and the non-photosynthetic nitrogen content of the corresponding vertical layer using the photosynthetic parameters and the physiological parameters, and constructing a nitrogen allocation model based on the photosynthetic nitrogen content and the non-photosynthetic nitrogen content;

[0011] S3: constructing an optimization model based on the nitrogen distribution model, the photosynthetic parameters and the physiological parameters, for optimizing the nitrogen vertical attenuation coefficient K of the corresponding vertical layer by combining photosynthetic nitrogen and non-photosynthetic nitrogen;

[0012] S4: Screening the spectral characteristic band corresponding to the attenuation coefficient K, and constructing an attenuation coefficient K estimation model based on a machine learning algorithm, taking the spectral reflectance information corresponding to the spectral characteristic band as input and the attenuation coefficient K as output;

[0013] S5: Inputting the currently collected cotton canopy spectral reflectance information into the attenuation coefficient K estimation model to obtain the real-time attenuation coefficient K, and inverting the nitrogen content of different vertical layers of the cotton canopy based on the real-time attenuation coefficient K.

[0014] Preferably, in S1, spectral reflectance information of the cotton canopy is collected at a fixed distance from the cotton canopy, and the spectral reflectance information includes in-situ spectral reflectance.

[0015] Preferably, in S1:

[0016] The photosynthetic parameters include: net photosynthetic rate A n , intercellular CO2 concentration C i 、Rubisco maximum carboxylation efficiency V c,max , maximum electron transfer rate J max , carboxylation efficiency CE and CO2 compensation point CCP;

[0017] The physiological parameters include: nitrogen content N, chlorophyll LCC, soluble protein SP, weighted antioxidant enzyme W AE .

[0018] Preferably, the weighted antioxidant enzyme W in the physiological parameters AE for:

[0019]

[0020] Where: Antioxidant enzyme indicators include: peroxidase POD, superoxide dismutase SOD, malondialdehyde MDA, catalase CAT, W POD , W SOD , W MDA , W CAT are the weights of the corresponding antioxidant enzyme indicators.

[0021] Preferably, S2 comprises the following steps:

[0022] S21: Obtaining the carboxylation system nitrogen PN of the current vertical layer of the cotton canopy based on the photosynthetic parameters and physiological parameters C , electron transfer nitrogen component PN B and light harvesting system nitrogen PN L , calculate the sum of the three to obtain the measured value of photosynthetic nitrogen PN;

[0023] S22: Characterizing photosynthetic nitrogen with chlorophyll content and characterizing non-photosynthetic nitrogen with weighted antioxidant enzymes, and constructing the nitrogen allocation model:

[0024] N=PN+NPN;

[0025] PN=LCC×S PN ;

[0026] NPN=W AE ×S NPN ;

[0027] Where: N is the measured value of nitrogen content in the current vertical layer of the cotton canopy, PN and NPN represent photosynthetic nitrogen and non-photosynthetic nitrogen, respectively, S PN is the ratio coefficient of photosynthetic nitrogen to chlorophyll, S NPN is the ratio coefficient of non-photosynthetic nitrogen to weighted antioxidant enzymes, LCC is the chlorophyll content, W AE for weighted antioxidant enzymes;

[0028] S23: Calculate the photosynthetic nitrogen to chlorophyll ratio coefficient S based on the measured value of photosynthetic nitrogen PN, the chlorophyll content LCC and the nitrogen distribution model PN ; Based on the measured value of photosynthetic nitrogen PN and weighted antioxidant enzyme W AE The photosynthetic nitrogen and weighted antioxidant enzyme ratio coefficient S is calculated using the nitrogen distribution model NPN .

[0029] Preferably, the nitrogen vertical attenuation coefficient K is:

[0030]

[0031]

[0032] Where: N0 is the nitrogen content in the vertical layer at the top of the canopy, F i is the cumulative leaf area index from the i-th layer to the top of the canopy, LCC0 is the chlorophyll content of the vertical layer at the top of the canopy, and LCC i To determine the chlorophyll content of the layer, Weighted antioxidant enzymes for the top vertical layer of the canopy, is the weighted antioxidant enzyme of the determination layer, n b The minimum value of nitrogen content in the determination layer, i.e., the light saturation photosynthetic rate A n The intercept in the regression with N0, K LCC and They are the chlorophyll vertical characteristic coefficient and the weighted antioxidant enzyme vertical characteristic coefficient respectively.

[0033] Preferably, the step of screening the spectral characteristic band corresponding to the attenuation coefficient K in S4 comprises:

[0034] S41: preprocessing the cotton canopy spectral reflectance information by using the radiation transfer model PROSAIL and continuous wavelet transform CWT;

[0035] S42: Obtaining the chlorophyll LCC and weighted antioxidant enzyme W of the current vertical layer in the sample database AEThe corresponding spectral characteristic bands, the chlorophyll LCC of the current vertical layer and the weighted antioxidant enzyme W calculated by S3 AE The corresponding nitrogen vertical attenuation coefficient K is used to establish a corresponding relationship and construct a spectral characteristic band screening algorithm model.

[0036] Preferably, the spectral feature band screening algorithm model is a multi-feature integration algorithm MFE.

[0037] Preferably, the machine learning algorithm in S4 includes: an estimation algorithm based on random forest RF, artificial neural network ANN or extreme gradient boosting tree XGBoost.

[0038] Preferably, the step of inverting the nitrogen content of different vertical layers of the cotton canopy based on the real-time attenuation coefficient K in S5 comprises:

[0039] Substitute the real-time attenuation coefficient K into the attenuation coefficient K estimation model to determine the layer nitrogen content N i As an unknown quantity, the nitrogen content in different vertical layers of the cotton canopy was inferred.

[0040] The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient proposed in the present invention has the following beneficial effects compared with the prior art:

[0041] The present invention photographs cotton canopy spectral data, synchronously obtains photosynthetic parameters of different vertical layers and parameters of different nitrogen components, analyzes the vertical distribution patterns of different nitrogen components in the cotton canopy, divides total nitrogen into photosynthetic nitrogen and non-photosynthetic nitrogen, characterizes photosynthetic nitrogen with chlorophyll, characterizes non-photosynthetic nitrogen with weighted antioxidant enzymes, and optimizes the coefficient K by comprehensively considering photosynthetic nitrogen and non-photosynthetic nitrogen on the basis of the traditional nitrogen vertical attenuation coefficient Kt which only considers photosynthetic nitrogen. The original spectral data are further preprocessed based on PROSAIL-CWT, and the K characteristic bands are screened in combination with MFE. The RF, ANN and XGBoost models are used to evaluate the potential of K to estimate canopy nitrogen nutrition, and the nitrogen content of different vertical layers of the cotton canopy is inverted based on the attenuation coefficient K.

[0042] The present invention combines different nitrogen components in the vertical scale of the cotton canopy with remote sensing technology. While utilizing spectral technology, it fully considers the vertical heterogeneity of different nitrogen components in the crop canopy during the inversion process. While ensuring the accuracy of nitrogen estimation, it provides a new method for improving the stability and universality of the cotton nitrogen nutrition estimation model. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without creative work.

[0044] Figure 1 A flowchart of a cotton canopy nitrogen nutrition spectral estimation method based on optimizing the nitrogen vertical attenuation coefficient provided by the present invention;

[0045] Figure 2 A linear fitting diagram of cotton photosynthetic nitrogen and chlorophyll content, non-photosynthetic nitrogen and weighted antioxidant enzymes provided in an embodiment of the present invention;

[0046] The corresponding figure (a) is the linear fitting diagram of cotton photosynthetic nitrogen and chlorophyll content;

[0047] The corresponding figure (b) is the linear fitting diagram of cotton non-photosynthetic nitrogen and weighted antioxidant enzymes;

[0048] Figure 3 The cotton leaf area index LAI provided in the embodiment of the present invention is respectively related to the chlorophyll content LCC, the weighted antioxidant enzyme W AE the relevant relationship;

[0049] The corresponding figure (a) is the correlation diagram between cotton leaf area index LAI and chlorophyll content LCC;

[0050] The corresponding figure (b) is the relationship between cotton leaf area index LAI and weighted antioxidant enzyme W AE Correlation diagram of

[0051] Figure 4 A distribution diagram of the spectral characteristic bands of the cotton canopy nitrogen vertical attenuation coefficient K and Kt provided in an embodiment of the present invention;

[0052] Figure 5 A distribution diagram of spectral characteristic bands of nitrogen components in different vertical layers of a cotton canopy provided by an embodiment of the present invention;

[0053] Figure 6 A schematic diagram of constructing a model for estimating the nitrogen vertical attenuation coefficient K and Kt based on spectral characteristic bands provided in an embodiment of the present invention;

[0054] The corresponding figure (a) is the optimized nitrogen vertical attenuation coefficient K estimation model;

[0055] The corresponding figure (b) is the traditional nitrogen vertical attenuation coefficient Kt estimation model;

[0056] Figure 7Model verification of the nitrogen vertical attenuation coefficient K and Kt provided in the embodiment of the present invention;

[0057] The corresponding figure (a) is a schematic diagram for verifying the optimized nitrogen vertical attenuation coefficient K;

[0058] The corresponding figure (b) is a schematic diagram of the verification of the traditional nitrogen vertical attenuation coefficient Kt;

[0059] Figure 8 A schematic diagram of linear fitting of the nitrogen content inverted based on the nitrogen vertical attenuation coefficient and the measured nitrogen content provided in an embodiment of the present invention;

[0060] The corresponding figure (a) is a schematic diagram of the estimation accuracy of nitrogen content in the upper layer of cotton canopy based on the optimized nitrogen vertical coefficient K;

[0061] The corresponding figure (b) is a schematic diagram of the estimation accuracy of nitrogen content in the middle layer of cotton canopy based on the optimized nitrogen vertical coefficient K;

[0062] The corresponding figure (c) is a schematic diagram of the estimation accuracy of nitrogen content in the lower layer of cotton canopy based on the optimized nitrogen vertical coefficient K;

[0063] The corresponding figure (d) is a schematic diagram of the estimation accuracy of nitrogen content in the upper layer of cotton canopy based on the traditional nitrogen vertical coefficient Kt;

[0064] The corresponding figure (e) is a schematic diagram of the estimation accuracy of nitrogen content in the middle layer of cotton canopy based on the traditional nitrogen vertical coefficient Kt;

[0065] The corresponding figure (f) is a schematic diagram of the estimation accuracy of nitrogen content in the lower layer of the cotton canopy based on the traditional nitrogen vertical coefficient Kt. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe 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.

[0067] By comprehensively analyzing the changing patterns of different nitrogen components in the vertical scale of the canopy, optimizing the attenuation coefficient K that characterizes the dynamic changes of different nitrogen components in crops is a difficult problem that needs to be solved in current research. The present invention divides different nitrogen components into photosynthetic nitrogen and non-photosynthetic nitrogen, introduces the expression of non-photosynthetic nitrogen components on the basis of the traditional canopy nitrogen vertical attenuation coefficient Kt, and further improves the sensitivity and estimation accuracy of the attenuation coefficient K to the vertical heterogeneity of crop canopy nitrogen, which deserves further discussion.

[0068] Photosynthetic nitrogen mainly includes nitrogen components involved in light energy capture and carbon assimilation, which is closely related to chlorophyll; non-photosynthetic nitrogen mainly includes nitrogen components in peroxisomes, cytoplasms and cell walls, which are related to antioxidant enzymes, among which SOD in the cytoplasm can resist the damage to cells caused by oxygen free radicals and repair damaged cells in time; POD and CAT in peroxisomes can not only prevent H2O2 from reacting with O2 to generate O2·- and -OH, but also participate in the secondary metabolic processes of plants, such as the synthesis of lignin and phenols; MDA is related to the activity of cell wall degrading enzymes and reactive oxygen proteins.

[0069] like Figure 1 As shown, a cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient provided by an embodiment of the present invention includes the following steps: S1: collecting cotton canopy spectral reflectance information, as well as photosynthetic parameters and physiological parameters of different vertical layers of the cotton canopy, and obtaining a cotton canopy vertical scale nitrogen nutrition monitoring spectral database as a sample database; S2: using photosynthetic parameters and physiological parameters to calculate the photosynthetic nitrogen content and non-photosynthetic nitrogen content of the corresponding vertical layer, and constructing a nitrogen allocation model based on the photosynthetic nitrogen content and the non-photosynthetic nitrogen content; S3: constructing an optimization model based on the nitrogen allocation model, photosynthetic parameters and physiological parameters, which is used to optimize the nitrogen vertical attenuation coefficient K of the corresponding vertical layer in combination with photosynthetic nitrogen and non-photosynthetic nitrogen; S4: screening the spectral characteristic band corresponding to the attenuation coefficient K, and based on a machine learning algorithm, using the spectral reflectance information corresponding to the spectral characteristic band as input and the attenuation coefficient K as output, constructing an attenuation coefficient K estimation model; S5: inputting the cotton canopy spectral reflectance information currently collected in real time into the attenuation coefficient K estimation model to obtain the real-time attenuation coefficient K, and inverting the nitrogen content of different vertical layers of the cotton canopy based on the real-time attenuation coefficient K.

[0070] In one embodiment, in S1, spectral reflectance information of the cotton canopy is collected at a fixed distance from the cotton canopy, and the spectral reflectance information includes in-situ spectral reflectance.

[0071] In one embodiment, in S1, the height of the cotton plant (excluding the roots) is evenly divided into three vertical layers of the cotton canopy: the upper layer, the middle layer, and the lower layer, and the photosynthetic parameters and the physiological parameters are collected respectively.

[0072] Photosynthetic parameters include: Net photosynthetic rate A n , intercellular CO2 concentration C i 、Rubisco maximum carboxylation efficiency V c,max , maximum electron transfer rate J max , carboxylation efficiency CE and CO2 compensation point CCP;

[0073] Physiological parameters include: nitrogen content N, chlorophyll LCC, soluble protein SP, weighted antioxidant enzyme W AE .

[0074] In one embodiment, the physiological parameters are weighted antioxidant enzymes W AE for:

[0075]

[0076] Where: Antioxidant enzyme indicators include: peroxidase POD, superoxide dismutase SOD, malondialdehyde MDA, catalase CAT, W POD , W SOD , W MDA , W CAT are the weights of the corresponding antioxidant enzyme indicators.

[0077] In one embodiment, S2 includes the following steps:

[0078] S21: Obtaining carboxylation system nitrogen PN of the current vertical layer of the cotton canopy based on photosynthetic parameters and physiological parameters C , electron transfer nitrogen component PN B and light harvesting system nitrogen PN L , calculate the sum of the three to obtain the measured value of photosynthetic nitrogen PN;

[0079] In this embodiment, the carboxylation system nitrogen PNC, the electron transfer nitrogen component PNB and the light capture system nitrogen PNL can be expressed as follows:

[0080]

[0081] Where: V c,max is the maximum carboxylation efficiency of Rubisco, V cr 20.78 μmol CO2 g at 25 °C -1 Rubisco's -1 , J max is the maximum electron transfer rate, J mc 155.65 μmol e at 25 °C -1 μmol of cytochrome f complex protein Cytfs -1 , LCC is the chlorophyll content, CB is the chlorophyll in the light-harvesting antenna LHCII of the combined photosystem IPSI, photosystem IIPSII, and photosystem II, and the value is 2.15mmol·g -1 N.

[0082] S22: Use chlorophyll content to characterize photosynthetic nitrogen, weighted antioxidant enzymes to characterize non-photosynthetic nitrogen, and construct a nitrogen distribution model:

[0083] N=PN+NPN;

[0084] PN=LCC×S PN ;

[0085] NPN=W AE ×S NPN ;

[0086] Where: N is the measured value of nitrogen content in the current vertical layer of the cotton canopy, PN and NPN represent photosynthetic nitrogen and non-photosynthetic nitrogen, respectively, S PN is the ratio coefficient of photosynthetic nitrogen to chlorophyll, S NPN is the ratio coefficient of non-photosynthetic nitrogen to weighted antioxidant enzymes, LCC is the chlorophyll content, W AE for weighted antioxidant enzymes;

[0087] S23: The photosynthetic nitrogen to chlorophyll ratio coefficient S is calculated based on the measured value of photosynthetic nitrogen PN, chlorophyll content LCC and nitrogen distribution model PN ; Based on the measured value of photosynthetic nitrogen PN and weighted antioxidant enzyme W AE The photosynthetic nitrogen and weighted antioxidant enzyme ratio coefficient S was calculated using the nitrogen distribution model NPN .

[0088] In one embodiment, the traditional nitrogen vertical attenuation coefficient Kt is calculated as:

[0089] (N i -n b )=(N0-n b )×exp(-KtF i );

[0090] The construction of the optimized nitrogen vertical attenuation coefficient K includes: on the basis of the traditional nitrogen vertical attenuation coefficient Kt, the optimized nitrogen vertical attenuation coefficient K is proposed by integrating photosynthetic nitrogen and non-photosynthetic nitrogen. The specific calculation is:

[0091]

[0092] The optimized nitrogen vertical attenuation coefficient K is:

[0093]

[0094] Where: Ni is the nitrogen content of the measured layer, N0 is the nitrogen content of the vertical layer at the top of the canopy, F i is the cumulative leaf area index from the i-th layer to the top of the canopy, LCC0 is the chlorophyll content of the vertical layer at the top of the canopy, and LCC i To determine the chlorophyll content of the layer, Weighted antioxidant enzymes for the top vertical layer of the canopy, is the weighted antioxidant enzyme of the determination layer, n b The minimum value of nitrogen content in the determination layer, i.e., the light saturation photosynthetic rate A n The intercept in the regression with N0, K LCC and They are the chlorophyll vertical characteristic coefficient and the weighted antioxidant enzyme vertical characteristic coefficient respectively.

[0095] In one embodiment, the step of selecting the spectral characteristic band corresponding to the attenuation coefficient K in S4 includes:

[0096] S41: Preprocessing of cotton canopy spectral reflectance information by using the radiative transfer model PROSAIL and continuous wavelet transform CWT;

[0097] S42: Get the chlorophyll LCC and weighted antioxidant enzyme W of the current vertical layer in the sample database AE The corresponding spectral characteristic bands, the chlorophyll LCC of the current vertical layer and the weighted antioxidant enzyme W calculated by S3 AE The corresponding nitrogen vertical attenuation coefficient K is used to establish a corresponding relationship and construct a spectral characteristic band screening algorithm model.

[0098] In one embodiment, the spectral feature band screening algorithm model is a multi-feature integration algorithm MFE.

[0099] In one embodiment, the machine learning algorithm in S4 includes: an estimation algorithm based on random forest RF, artificial neural network ANN or extreme gradient boosting tree XGBoost.

[0100] In one embodiment, the step of inverting the nitrogen content of different vertical layers of the cotton canopy based on the real-time attenuation coefficient K in S5 includes:

[0101] Substitute the real-time attenuation coefficient K into the attenuation coefficient K estimation model to determine the layer nitrogen content N i As an unknown quantity, the nitrogen content in different vertical layers of the cotton canopy was inferred.

[0102] The specific experimental data of the embodiment of the present invention are given below:

[0103] The present invention sets up a cotton test plot with different nitrogen application levels. The specific test design and sampling date are shown in Table 1. By collecting cotton canopy spectral reflectance information, the spectral data is spectrally calibrated, spatial radiation and spectral radiation calibration are processed to convert into spectral reflectance as the original reflectance (R); the cotton plant height (excluding the root) is evenly divided into upper layer, middle layer and lower layer, and the collected photosynthetic parameters include net photosynthetic rate An, intercellular CO2 concentration value C i 、Rubisco maximum carboxylation efficiency V c,max , maximum electron transfer rate J max, carboxylation efficiency CE and CO2 compensation point CCP; the collected physiological parameters include leaf area index LAI, dry matter per unit area DRY, nitrogen content N, chlorophyll LCC, soluble protein SP, antioxidant enzyme indicators: malondialdehyde MDA, superoxide dismutase SOD, peroxidase POD, catalase CAT. 45 data sets were collected in each period in 2021 and 2022 (5 nitrogen gradients, 3 replicate plots for each nitrogen gradient, 3 copies for each plot), and 30 data sets were collected in each period in 2023 (5 nitrogen gradients, 3 replicate plots for each nitrogen gradient, 2 copies for each plot), for a total of 480 data sets, in order to establish a spectral database for nitrogen nutrition monitoring at the vertical scale of the cotton canopy.

[0104] Table 1 Experimental design and data collection information

[0105]

[0106] In order to verify the rationality of the nitrogen allocation model, the present invention uses 135 sample data from the seedling stage in 2021 (45 samples each in the upper, middle and lower layers) to perform linear fitting on photosynthetic nitrogen components and chlorophyll content, non-photosynthetic nitrogen components and weighted antioxidant enzymes, as shown in the following figure: Figure 2 As shown, the correlation coefficient R 2 They are 0.787 and 0.667 respectively, which have a good correlation and confirm the rationality of our hypothesis.

[0107] Based on the traditional nitrogen vertical attenuation coefficient Kt, the present invention combines the nitrogen distribution model to construct an optimized nitrogen vertical attenuation coefficient K. The specific details of the invention are as follows: The present invention uses 270 sample data from the early boll period and the peak boll period in 2021 (45 samples in each period, upper, middle and lower layers) to analyze the relationship between different nitrogen components and leaf area index, among which chlorophyll and leaf area index are exponentially related, and weighted antioxidant enzymes are linearly related to leaf area index, such as Figure 3 shown.

[0108] The coefficient K and Kt estimation model based on the spectral characteristic band is constructed. The specific details are as follows: The present invention performs continuous wavelet transform CWT preprocessing on the spectral data, and constructs the partial least squares PLSR estimation model of the nitrogen vertical attenuation coefficient K and Kt of CWT based on the full band to screen the optimal spectral preprocessing, as shown in Table 2. The optimal preprocessing of coefficients K and Kt is CWT-4, and the model estimation accuracy wR 2 They are 0.689 and 0.677 respectively.

[0109] The present invention subsequently screens the attenuation coefficients K, Kt and characteristic bands of nitrogen components in different vertical layers based on the optimal pretreatment, such as Figure 4As shown in the figure, the characteristic band of the traditional coefficient Kt based on MFE screening is mainly distributed in the red light region of 600-900nm, and there is also a small amount of characteristic band distribution in 1200-1300nm; the present invention further found that the characteristic band of the optimized nitrogen vertical attenuation coefficient K not only exists in visible light, but also has a characteristic band aggregation trend in the short-wave near-infrared region of 1150-1650nm. The distribution of characteristic bands of LCC, SOD, CAT, MDA and POD in other different layers is shown in the figure. Figure 5 .

[0110] Table 2 Optimal pretreatment screening of nitrogen vertical attenuation coefficient based on full-band PLSR. Bold values ​​indicate wR corresponding to the best pretreatment of different indicators. 2

[0111]

[0112] The accuracy of the estimation model constructed based on PROSAIL simulated reflectivity in the present invention is significantly higher than that of the measured reflectivity estimation model, as shown in Table 3. Therefore, the estimation model is subsequently constructed based on PROSAIL simulated reflectivity.

[0113] The present invention further constructs a nitrogen vertical attenuation coefficient estimation model based on the characteristic bands screened by MFE, such as Figure 6 As shown in the figure, the estimation accuracy wR of the optimized nitrogen vertical attenuation coefficient K model is 2 is 0.886, and the estimation accuracy of the traditional nitrogen vertical attenuation coefficient Kt model is wR 2 It is 0.793.

[0114] Table 3 PLSR estimation model of attenuation coefficients K and Kt based on MFE feature screening method

[0115]

[0116] In order to evaluate the universality and stability of the model, the present invention uses 120 independent sample data from 2023 to test the model. Figure 7 As shown in the figure, the prediction accuracy wR of the optimized nitrogen vertical attenuation coefficient K is 2 is 0.862, compared with the traditional nitrogen vertical attenuation coefficient Kt model, wR 2 An increase of 14.10%.

[0117] Based on the optimized nitrogen vertical attenuation coefficient K and the traditional nitrogen vertical attenuation coefficient Kt, the nitrogen content in different vertical layers is inverted. Figure 8As shown in the figure, the inversion accuracy of the optimized nitrogen vertical attenuation coefficient K for nitrogen content in different vertical layers is higher than that of the traditional nitrogen vertical attenuation coefficient Kt, and the inversion accuracy of nitrogen content in different vertical layers of the canopy is higher than 0.72. The inversion accuracy of the traditional nitrogen vertical attenuation coefficient Kt for nitrogen content is lower than 0.70 for the middle and lower parts except the upper part of the canopy. Among them, the inversion accuracy of the optimized nitrogen vertical attenuation coefficient K for nitrogen content in the upper part of the canopy is the highest, wR 2 is 0.815, and MAE is 0.196. Compared with the traditional nitrogen vertical attenuation coefficient Kt inversion accuracy, wR 2 It increased by 11.49% and MAE decreased by 16.84%.

[0118] It can be seen that the present invention analyzes the changing rules of photosynthetic nitrogen and non-photosynthetic nitrogen components in the vertical scale of the cotton canopy at different growth periods, fits the chlorophyll characterization of photosynthetic nitrogen, and the weighted antioxidant enzyme characterization of non-photosynthetic nitrogen, and optimizes the nitrogen vertical attenuation coefficient K, which can characterize the dynamic distribution of photosynthetic nitrogen and non-photosynthetic nitrogen in the vertical scale of the cotton canopy. The present invention further estimates the optimized nitrogen vertical attenuation coefficient K based on remote sensing technology, clarifies the potential of optimizing the nitrogen vertical attenuation coefficient K to invert the nitrogen content of different vertical layers of the cotton canopy, and provides a new method for improving the estimation accuracy of nitrogen nutrition in the vertical scale of the cotton canopy.

[0119] The above is a detailed introduction to the cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present invention.

[0120] In this article, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

Claims

1. A cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient, characterized in that: The steps include: S1: Collect the spectral reflectance information of cotton canopy, as well as the photosynthetic parameters and physiological parameters of different vertical layers of cotton canopy as a sample database; S2: calculating the photosynthetic nitrogen content and the non-photosynthetic nitrogen content of the corresponding vertical layer using the photosynthetic parameters and the physiological parameters, and constructing a nitrogen allocation model based on the photosynthetic nitrogen content and the non-photosynthetic nitrogen content; S3: constructing an optimization model based on the nitrogen distribution model, the photosynthetic parameters and the physiological parameters, for optimizing the nitrogen vertical attenuation coefficient K of the corresponding vertical layer by combining photosynthetic nitrogen and non-photosynthetic nitrogen; S4: Screening the spectral characteristic band corresponding to the attenuation coefficient K, and constructing an attenuation coefficient K estimation model based on a machine learning algorithm, taking the spectral reflectance information corresponding to the spectral characteristic band as input and the attenuation coefficient K as output; S5: Inputting the currently collected cotton canopy spectral reflectance information into the attenuation coefficient K estimation model to obtain the real-time attenuation coefficient K, and inverting the nitrogen content of different vertical layers of the cotton canopy based on the real-time attenuation coefficient K.

2. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 1 is characterized in that: In S1, spectral reflectance information of the cotton canopy is collected at a fixed distance from the cotton canopy, and the spectral reflectance information includes in-situ spectral reflectance.

3. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 1 is characterized in that: In S1: The photosynthetic parameters include: net photosynthetic rate A n , intercellular CO2 concentration C i 、Rubisco maximum carboxylation efficiency V c,max , maximum electron transfer rate J max , carboxylation efficiency CE and CO2 compensation point CCP; The physiological parameters include: nitrogen content N, chlorophyll LCC, soluble protein SP, weighted antioxidant enzyme W AE .

4. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 1 is characterized in that: Weighted antioxidant enzymes W in the physiological parameters AE for: Where: Antioxidant enzyme indicators include: peroxidase POD, superoxide dismutase SOD, malondialdehyde MDA, catalase CAT, W POD , W SOD , W MDA , W CAT are the weights of the corresponding antioxidant enzyme indicators.

5. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 1, characterized in that: The S2 comprises the following steps: S21: Obtaining the carboxylation system nitrogen PN of the current vertical layer of the cotton canopy based on the photosynthetic parameters and physiological parameters C , electron transfer nitrogen component PN B and light harvesting system nitrogen PN L , calculate the sum of the three to obtain the measured value of photosynthetic nitrogen PN; S22: Characterizing photosynthetic nitrogen with chlorophyll content and characterizing non-photosynthetic nitrogen with weighted antioxidant enzymes, and constructing the nitrogen allocation model: N=PN+NPN; PN=LCC×S PN ; NPN=W AE ×S NPN ; Where: N is the measured value of nitrogen content in the current vertical layer of the cotton canopy, PN and NPN represent photosynthetic nitrogen and non-photosynthetic nitrogen, respectively, S PN is the ratio coefficient of photosynthetic nitrogen to chlorophyll, S NPN is the ratio coefficient of non-photosynthetic nitrogen to weighted antioxidant enzymes, LCC is the chlorophyll content, W AE for weighted antioxidant enzymes; S23: Calculate the photosynthetic nitrogen to chlorophyll ratio coefficient S based on the measured value of photosynthetic nitrogen PN, the chlorophyll content LCC and the nitrogen distribution model PN ; Based on the measured value of photosynthetic nitrogen PN and weighted antioxidant enzyme W AE The photosynthetic nitrogen and weighted antioxidant enzyme ratio coefficient S is calculated using the nitrogen distribution model NPN .

6. A cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 5, characterized in that: The nitrogen vertical attenuation coefficient K is: Where: N0 is the nitrogen content in the vertical layer at the top of the canopy, F i is the cumulative leaf area index from the i-th layer to the top of the canopy, LCC0 is the chlorophyll content of the vertical layer at the top of the canopy, and LCC i To determine the chlorophyll content of the layer, Weighted antioxidant enzymes for the top vertical layer of the canopy, is the weighted antioxidant enzyme of the determination layer, n b The minimum value of nitrogen content in the determination layer, i.e., the light saturation photosynthetic rate A n The intercept in the regression with N0, K LCC and They are the chlorophyll vertical characteristic coefficient and the weighted antioxidant enzyme vertical characteristic coefficient respectively.

7. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 1, characterized in that: The step of selecting the spectral characteristic band corresponding to the attenuation coefficient K in S4 includes: S41: preprocessing the cotton canopy spectral reflectance information by using the radiation transfer model PROSAIL and continuous wavelet transform CWT; S42: Obtaining the chlorophyll LCC and weighted antioxidant enzyme W of the current vertical layer in the sample database AE The corresponding spectral characteristic bands, the chlorophyll LCC of the current vertical layer and the weighted antioxidant enzyme W calculated by S3 AE The corresponding nitrogen vertical attenuation coefficient K is used to establish a corresponding relationship and construct a spectral characteristic band screening algorithm model.

8. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 7, characterized in that: The spectral feature band screening algorithm model is a multi-feature integration algorithm MFE.

9. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 1, characterized in that: The machine learning algorithm in S4 includes: an estimation algorithm based on random forest RF, artificial neural network ANN or extreme gradient boosting tree XGBoost.

10. The cotton canopy nitrogen monitoring method based on optimizing the nitrogen vertical attenuation coefficient according to claim 1, characterized in that: The step of inverting the nitrogen content of different vertical layers of the cotton canopy based on the real-time attenuation coefficient K in S5 includes: Substitute the real-time attenuation coefficient K into the attenuation coefficient K estimation model to determine the layer nitrogen content N i As an unknown quantity, the nitrogen content in different vertical layers of the cotton canopy was inferred.

Images