A crop stress decoupling discrimination method based on hot water carbon multi-dimensional manifold phase space

Abstract

This invention discloses a crop stress decoupling and discrimination method based on the multidimensional manifold phase space of hot water carbon. First, multimodal remote sensing data of the target area is acquired, and pixel-level three-dimensional physical parameters of canopy temperature, equivalent water thickness, and chlorophyll fluorescence are obtained through alignment and inversion. Then, the temporal change rate of these parameters is calculated to construct a Jacobian matrix, extracting the temporal characteristics of crop physiological stress. The spatial second-order partial derivative of the carbon assimilation parameter is calculated to extract the spatial abrupt change characteristics of stress. Subsequently, a multidimensional partial differential equation system is constructed, and logical decisions are made based on the differences in the spatiotemporal physical characteristics of different stresses. Finally, the decision results are optimized to accurately separate drought, nutrient deficiency, and pathogen infection stresses, outputting a pixel-level diagnostic map of crop compound stress mechanisms. This invention decouples compound stresses at the level of crop physiological mechanisms, enabling early identification of abnormal physiological changes, overcoming static threshold dependence, and exhibiting strong anti-interference capabilities. It can achieve early and accurate diagnosis of compound stresses in large-scale crops, providing data support for precision agricultural intervention.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of agricultural remote sensing and ecological metabolic dynamics, specifically involving a crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space. Background Technology

[0002] Against the backdrop of global climate change and intensive agricultural production, the ability to quickly and accurately determine the type of stress crops are under is a crucial prerequisite for precise irrigation, fertilization, and pesticide application, and is of great significance to ensuring food security.

[0003] Currently, the mainstream remote sensing monitoring methods for crop stress in the industry are mainly divided into two categories: threshold monitoring based on a single optical vegetation index and machine learning classification based on multi-source remote sensing features. However, both of these methods have obvious technical defects.

[0004] The first threshold monitoring method relies on calculating indicators such as the normalized difference in vegetation index (NDVI) and chlorophyll index using multispectral remote sensing data. When these indicators fall below historical thresholds for the same period, crop damage is determined. The biggest problem with this method is the potential for misjudgment due to spectral similarity and distortion of the inversion results. Drought-induced yellowing of leaves, chlorophyll loss due to nutrient deficiency, and cell necrosis due to disease can all cause crops to exhibit highly similar spectral characteristics in optical remote sensing images, making it impossible for the model to distinguish which specific factor caused the crop damage.

[0005] The second type of machine learning classification method involves concatenating multi-source data such as optical reflectivity, thermal infrared temperature, and microwave scattering coefficient into a high-dimensional vector, which is then input into models such as random forests and support vector machines for supervised classification. This method is essentially a purely data-driven algorithm that does not consider the physiological and physical mechanisms by which crops respond to stress. It completely ignores the differences in the speed and process of physiological responses under different stresses. When encountering years with abnormal weather or complex stress conditions not present in the training data, the model's classification criteria will be severely skewed, or even completely ineffective, easily leading to incorrect stress type identification.

[0006] In summary, existing crop stress monitoring methods remain at the level of static, one-dimensional surface feature monitoring, failing to integrate the three core physiological processes of crop temperature change, water metabolism, and carbon assimilation, and thus failing to differentiate different stress types from a dynamic physical change perspective. The field of agricultural remote sensing urgently needs a method that can accurately distinguish complex stresses based on the laws governing crop physiological and physical changes. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of existing technologies in remote sensing monitoring of crop stress. Existing methods, such as threshold monitoring based on a single optical vegetation index, rely on static thresholds to determine crop damage status, resulting in serious misjudgments due to spectral similarity and distortion of inversion results. Crop damage caused by drought, nutrient deficiency, and pathogen infection exhibits highly similar characteristics in optical remote sensing images, making it impossible to effectively distinguish stress types. Machine learning classification methods based on multi-source remote sensing features are purely data-driven algorithms, failing to consider the physiological and physical mechanisms of crop stress response and ignoring the differences in the speed and process of crop physiological responses under different stresses. In years with abnormal weather or when facing complex stresses not included in the training data, the model classification criteria are prone to severe deviation or even failure, easily leading to misjudgments. Both types of methods remain at the level of static, single-dimensional surface feature monitoring, failing to integrate the three core physiological processes of crop temperature change, water metabolism, and carbon assimilation, thus failing to achieve accurate differentiation of complex stresses from the perspective of dynamic physical changes. To this end, this invention proposes a crop stress decoupling and discrimination method based on hot water carbon multidimensional manifold phase space, which can achieve precise decoupling and pixel-level early accurate diagnosis of spectrally similar complex stresses such as drought, nutrient deficiency, and pathogen infection from the perspective of crop physiological mechanisms, and provide reliable data support for precision agricultural intervention in large-scale farmland.

[0008] This invention provides a crop stress decoupling discrimination method based on the multidimensional manifold phase space of hot water carbon. The technical approach is as follows: First, acquire multimodal remote sensing data of the target area, align and invert to obtain pixel-level three-dimensional physical parameters of canopy temperature, equivalent water thickness, and chlorophyll fluorescence; then, calculate the temporal change rate of the parameters to construct the Jacobian matrix, extract the temporal characteristics of crop physiological stress, calculate the spatial second-order partial derivative of the carbon assimilation parameter, and extract the spatial abrupt change characteristics of stress; subsequently, construct a multidimensional partial differential equation system, and perform logical judgment based on the spatiotemporal physical characteristics differences of different stresses; finally, optimize the judgment results, accurately separate drought, nutrient deficiency, and pathogen infection stresses, and output a pixel-level crop composite stress mechanism diagnostic map.

[0009] The method includes the following steps:

[0010] Step S1, Multi-source remote sensing data processing and 3D physical parameter inversion: Acquire multi-frequency and multi-type remote sensing observation data covering the target area during crop growth stages, perform spatiotemporal coordinate alignment and inversion reconstruction on the data, and obtain a pixel-level 3D physical state vector sequence. This vector contains three core parameters: canopy equivalent temperature, canopy equivalent water thickness, and sunlight-induced chlorophyll fluorescence, which respectively reflect the crop's temperature changes, water metabolism, and carbon assimilation processes.

[0011] Step S2, Temporal Feature Extraction and Jacobi Dynamics Matrix Construction: In a continuous time coordinate system, calculate the first-order rate of change of the above three-dimensional physical state vector with time, and construct the Jacobi dynamics evolution matrix. This matrix contains the rate of change of temperature, moisture, and carbon assimilation, thereby extracting the speed characteristics of the physiological stress response of crops under stress.

[0012] Step S3, Spatial Feature Extraction and Laplace Operator Matrix Solving: In the two-dimensional spatial coordinate system of the target area, calculate the spatial second-order partial derivative of the chlorophyll fluorescence parameter to generate the Laplace operator matrix, thereby extracting spatial features that can reflect the spread of disease and sudden changes in crop growth status in local areas.

[0013] Step S4: Construction of Partial Differential Equations and Threshold Determination of Stress Type: Construct a multidimensional partial differential equation system for stress decoupling conditions, using the temporal characteristics of the Jacobian matrix and the spatial characteristics of the Laplace operator as input parameters; based on the differences in the physiological and physical change rates and spatial distribution characteristics corresponding to drought, nutrient deficiency, and disease, determine the type of stress suffered by the crop through partial differential thresholds.

[0014] Step S5: Decoupling of Complex Stresses and Optimization of Diagnostic Atlas Output: Based on the judgment results of the partial differential equation system, drought, nutrient deficiency, and pathogen infection stresses that result in similar spectra are accurately separated at the mathematical level. The results are then spatially smoothed and optimized to finally output a pixel-level diagnostic atlas of crop multi-factor complex stress mechanisms.

[0015] Step S1 includes the following steps:

[0016] Multimodal remote sensing data extraction and preprocessing: acquiring continuous time-series thermal infrared remote sensing data, shortwave infrared and microwave remote sensing data, and sunlight-induced chlorophyll fluorescence (SIF) remote sensing data of the target area; performing geometric correction on all data and completing pixel-level spatial coordinate alignment to ensure the spatiotemporal consistency of the data;

[0017] Pixel-level 3D physical state vector inversion: For each spatial pixel, a unified time step is reconstructed. The three-dimensional physical parameters define the pixel in Physical state vector at time t :

[0018]

[0019] in, The temperature dimension characterizes the canopy equivalent temperature due to heat dissipation from crop stomata closure. The water dimension characterizes the canopy equivalent water thickness, representing the water loss from crop mesophyll tissue. The carbon assimilation dimension characterizes the absolute value of chlorophyll fluorescence radiation when photosynthetic electron transport is hindered in crops. This represents the transpose of a matrix.

[0020] Preferably, step S2 includes the following steps:

[0021] Continuous time-space phase mapping maps the three-dimensional physical state vector of each pixel. This data is mapped onto a continuously changing three-dimensional multidimensional manifold phase space, transforming discrete remote sensing observation data into a continuously changing trajectory of crop physiological changes over time.

[0022] The Jacobian dynamics evolution matrix is ​​solved by calculating the three-dimensional physical state vector as a function of calendar time in the aforementioned three-dimensional phase space. First-order partial derivatives are used to construct the Jacobian matrix. This reflects the rate of evolution of crop physiological stress:

[0023]

[0024] in, Represents the rate of temperature change. Represents the rate of change in moisture content. This represents the rate of carbon assimilation.

[0025] Furthermore, step S3 includes the following steps:

[0026] Spatial Laplacian operator calculation, in the two-dimensional spatial coordinate system of a pixel. Internally, chlorophyll fluorescence parameters related to carbon assimilation. Calculate the second-order spatial partial derivatives to obtain the Laplace operator matrix. :

[0027]

[0028] Spatial high-frequency mutation feature extraction involves extracting spatial mutation features within the local pixel neighborhood from the Laplacian operator matrix. Spatial mutation characteristic The quantitative criteria are as follows:

[0029] when When this is determined to be a spatial consistency feature, it indicates that the stress is uniformly distributed over a large area within the target region, where... The preset spatial smoothing threshold has a value range of [value range missing]. ;

[0030] when This was determined to be a spatial mutation characteristic, representing the stress spreading outward in patches or concentric circles from the source of infection. The preset spatial mutation threshold, and Furthermore, step S4 includes the following steps:

[0031] Acute drought stress identification: when the rate of temperature change is satisfied Equivalent water thickness change rate And spatial mutation characteristic quantity At that time, it was determined to be acute drought stress; among them The threshold for temperature increase, The threshold for the rate of water loss;

[0032] Chronic nutrient deficiency stress identification: when the rate of temperature change is met chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be chronic nutritional deficiency stress; among them This represents the upper limit for environmental noise fluctuations. This represents the photosynthetic inhibition threshold.

[0033] Infectious disease stress identification: when the rate of temperature change is satisfied chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be an infectious disease stress;

[0034] Normal growth state determination: When the absolute value of the time change rate of each dimension parameter is less than the corresponding preset steady-state threshold, and At that time, the crop was judged to be in a normal growth state.

[0035] Preferably, step S5 includes the following steps:

[0036] The probability transformation of the partial differential judgment result transforms the partial differential threshold judgment result in step 4 into a normalized probability matrix of each pixel belonging to drought, nutrient deficiency, or disease stress.

[0037] The results optimization and diagnostic map output utilize Markov random fields to perform spatiotemporal smoothing optimization on the probability matrix, eliminating local outliers, and finally outputting a pixel-level crop composite stress mechanism diagnostic map that accurately distinguishes the distribution of drought, nutrient deficiency, and pathogen infection.

[0038] Compared with the prior art, the present invention, employing the above technical solution, has the following beneficial effects:

[0039] (1) This invention combines the three core physiological processes of crop temperature change, water metabolism and carbon assimilation, and analyzes them together in a continuous time physical coordinate system. By utilizing the characteristics of different physiological responses of crops caused by drought, lack of fertilizer and disease, different stress factors are distinguished, thus solving the problem of misjudgment of different stresses.

[0040] (2) By monitoring the subtle changes in crop temperature, moisture and carbon assimilation, this invention can identify abnormal physiological changes 3-5 days before macroscopic damage occurs in crops, thus gaining crucial intervention time for precision agriculture irrigation, fertilization and pesticide application.

[0041] (3) This method determines the stress type by analyzing the relative rate of change and spatial distribution characteristics of crop physiological parameters. It mathematically isolates the influence of different meteorological backgrounds and regional environments, and has good adaptability to different climate zones and different crop varieties. The stability and universality of the model are greatly improved. Attached Figure Description

[0042] Figure 1 This is a schematic flowchart of a crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space proposed in this invention.

[0043] Figure 2 This is a three-dimensional phase space trajectory diagram of rice stress, which is the crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space proposed in this invention.

[0044] Figure 3 This is a graph showing the carbon assimilation rate of rice, based on a crop stress decoupling discrimination method using hot water carbon multidimensional manifold phase space proposed in this invention.

[0045] Figure 4 This is a Laplace space feature map of rice, which is the crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space proposed in this invention. Detailed Implementation

[0046] The technical solutions of 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 some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] Example 1:

[0048] This invention provides a crop stress decoupling discrimination method based on the multidimensional manifold phase space of hot water carbon, comprising the following steps:

[0049] Step S1, Multi-source remote sensing data processing and 3D physical parameter inversion: Acquire multi-frequency and multi-type remote sensing observation data covering the target area during crop growth stages, perform spatiotemporal coordinate alignment and inversion reconstruction on the data, and obtain a pixel-level 3D physical state vector sequence. This vector contains three core parameters: canopy equivalent temperature, canopy equivalent water thickness, and sunlight-induced chlorophyll fluorescence, which respectively reflect the crop's temperature changes, water metabolism, and carbon assimilation processes.

[0050] Step S1 includes the following steps:

[0051] Multimodal remote sensing data extraction and preprocessing: acquiring continuous time-series thermal infrared remote sensing data, shortwave infrared and microwave remote sensing data, and sunlight-induced chlorophyll fluorescence (SIF) remote sensing data of the target area; performing geometric correction on all data and completing pixel-level spatial coordinate alignment to ensure the spatiotemporal consistency of the data;

[0052] Pixel-level 3D physical state vector inversion: For each spatial pixel, a unified time step is reconstructed. The three-dimensional physical parameters define the pixel in Physical state vector at time t :

[0053]

[0054] in, The temperature dimension characterizes the equivalent canopy temperature (T) of crop stomatal closure and heat dissipation. The water dimension characterizes the canopy equivalent water thickness, representing the water loss from crop mesophyll tissue. The carbon assimilation dimension characterizes the absolute value of chlorophyll fluorescence radiation when photosynthetic electron transport is hindered in crops. This represents the transpose of a matrix.

[0055] Step S2, Temporal Feature Extraction and Jacobi Dynamics Matrix Construction: In a continuous time coordinate system, calculate the first-order rate of change of the above three-dimensional physical state vector with time, and construct the Jacobi dynamics evolution matrix. This matrix contains the rate of change of temperature, moisture, and carbon assimilation, thereby extracting the speed characteristics of the physiological stress response of crops under stress.

[0056] Step S2 includes the following steps:

[0057] Continuous time-space phase mapping maps the three-dimensional physical state vector of each pixel. This data is mapped onto a continuously changing three-dimensional multidimensional manifold phase space, transforming discrete remote sensing observation data into a continuously changing trajectory of crop physiological changes over time.

[0058] The Jacobian dynamics evolution matrix is ​​solved by calculating the three-dimensional physical state vector as a function of calendar time in the aforementioned three-dimensional phase space. First-order partial derivatives are used to construct the Jacobian matrix. This reflects the rate of evolution of crop physiological stress:

[0059]

[0060] in, Represents the rate of temperature change. Represents the rate of change in moisture content. This represents the rate of carbon assimilation.

[0061] Step S3, Spatial Feature Extraction and Laplace Operator Matrix Solving: In the two-dimensional spatial coordinate system of the target area, calculate the spatial second-order partial derivative of the chlorophyll fluorescence parameter to generate the Laplace operator matrix, thereby extracting spatial features that can reflect the spread of disease and sudden changes in crop growth status in local areas.

[0062] Step S3 includes the following steps:

[0063] Spatial Laplacian operator calculation, in the two-dimensional spatial coordinate system of a pixel. Internally, chlorophyll fluorescence parameters related to carbon assimilation. Calculate the second-order spatial partial derivatives to obtain the Laplace operator matrix. :

[0064]

[0065] Spatial high-frequency mutation feature extraction involves extracting spatial mutation features within the local pixel neighborhood from the Laplacian operator matrix. Spatial mutation characteristic The quantitative criteria are as follows:

[0066] when When this is determined to be a spatial consistency feature, it indicates that the stress is uniformly distributed over a large area within the target region, where... The preset spatial smoothing threshold has a value range of [value range missing]. ;

[0067] when This was determined to be a spatial mutation characteristic, representing the stress spreading outward in patches or concentric circles from the source of infection. The preset spatial mutation threshold, and .

[0068] Step S4: Construction of Partial Differential Equations and Threshold Determination of Stress Type: Construct a multidimensional partial differential equation system for stress decoupling conditions, using the temporal characteristics of the Jacobian matrix and the spatial characteristics of the Laplace operator as input parameters; based on the differences in the physiological and physical change rates and spatial distribution characteristics corresponding to drought, nutrient deficiency, and disease, determine the type of stress suffered by the crop through partial differential thresholds.

[0069] Step S4 includes the following steps:

[0070] Acute drought stress identification: when the rate of temperature change is satisfied Equivalent water thickness change rate And spatial mutation characteristic quantity At that time, it was determined to be acute drought stress; among them The threshold for temperature increase, The threshold for the rate of water loss;

[0071] Chronic nutrient deficiency stress identification: when the rate of temperature change is met chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be chronic nutritional deficiency stress; among them This represents the upper limit for environmental noise fluctuations. This represents the photosynthetic inhibition threshold.

[0072] Infectious disease stress identification: when the rate of temperature change is satisfied chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be an infectious disease stress;

[0073] Normal growth state determination: When the absolute value of the time change rate of each dimension parameter is less than the corresponding preset steady-state threshold, and When the crop is in a normal growth state, it is determined that the crop is growing normally. Step S5, Decoupling of Complex Stress and Optimization of Diagnostic Atlas Output: Based on the judgment results of the partial differential equation system, drought, nutrient deficiency and pathogen infection stresses that lead to similar spectra are accurately separated at the mathematical level. Then, the results are spatially smoothed and optimized, and finally a pixel-level diagnostic atlas of crop multi-factor complex stress mechanism is output.

[0074] Step S5 includes the following steps:

[0075] The probability transformation of the partial differential judgment result transforms the partial differential threshold judgment result in step 4 into a normalized probability matrix of each pixel belonging to drought, nutrient deficiency, or disease stress.

[0076] The results optimization and diagnostic map output utilize Markov random fields to perform spatiotemporal smoothing optimization on the probability matrix, eliminating local outliers, and finally outputting a pixel-level crop composite stress mechanism diagnostic map that accurately distinguishes the distribution of drought, nutrient deficiency, and pathogen infection.

[0077] Example 2 shows the specific implementation effect at a rice base:

[0078] Step S1, Multi-source remote sensing data processing and 3D physical parameter inversion: Acquire multi-frequency and multi-type remote sensing observation data covering the target area during crop growth stages, perform spatiotemporal coordinate alignment and inversion reconstruction on the data, and obtain a pixel-level 3D physical state vector sequence. This vector contains three core parameters: canopy equivalent temperature, canopy equivalent water thickness, and sunlight-induced chlorophyll fluorescence, which respectively reflect the crop's temperature changes, water metabolism, and carbon assimilation processes.

[0079] Step S1 includes the following steps:

[0080] Multimodal remote sensing data extraction and preprocessing: Continuous time series data from April 2024 were acquired, including thermal infrared data from the Landsat-9 satellite for canopy equivalent temperature inversion, microwave data from the Sentinel-1 satellite for canopy equivalent water thickness inversion, and solar-induced chlorophyll fluorescence (SIF) data from the OCO-3 satellite.

[0081] Pixel-level 3D physical state vector inversion: Using a geometric correction algorithm, pixels of different resolutions are uniformly resampled to a spatial resolution of 30 meters, and linear interpolation is performed with a time step of 1 day to construct a pixel-level 3D physical state vector sequence. :

[0082]

[0083] in, The temperature dimension characterizes the equivalent canopy temperature (T) of crop stomatal closure and heat dissipation. The water dimension characterizes the canopy equivalent water thickness, representing the water loss from crop mesophyll tissue. The carbon assimilation dimension characterizes the absolute value of chlorophyll fluorescence radiation when photosynthetic electron transport is hindered in crops. This represents the transpose of a matrix.

[0084] Step S2, Temporal Feature Extraction and Jacobi Dynamics Matrix Construction: In a continuous time coordinate system, calculate the first-order rate of change of the above three-dimensional physical state vector with time, and construct the Jacobi dynamics evolution matrix. This matrix contains the rate of change of temperature, moisture, and carbon assimilation, thereby extracting the speed characteristics of the physiological stress response of crops under stress.

[0085] Step S2 includes the following steps:

[0086] Continuous-time phase-space mapping: mapping the three-dimensional physical state vector of each pixel. This data is mapped onto a continuously changing three-dimensional multidimensional manifold phase space, transforming discrete remote sensing observation data into a continuously changing trajectory of crop physiological changes over time.

[0087] The Jacobian dynamics evolution matrix is ​​solved by calculating the three-dimensional physical state vector as a function of calendar time in the aforementioned three-dimensional phase space. First-order partial derivatives are used to construct the Jacobian matrix. This reflects the rate of evolution of crop physiological stress:

[0088]

[0089] in, Represents the rate of temperature change. Represents the rate of change in moisture content. This represents the rate of carbon assimilation.

[0090] Figure 2 This is a three-dimensional phase space trajectory diagram of rice stress, taking healthy rice, drought, and pest / disease conditions as examples. Healthy rice exhibits high [stress / stress] in phase space. ,Low ,high A stable minimum region; in drought-stricken rice, due to the closure of stomata and the cessation of transpiration and heat dissipation, its trajectory will follow... The axis rose sharply, and at the same time A significant drop in the axis; pests and diseases are the first to show their effects in rice. The decline, and and Initial changes were relatively slow.

[0091] Figure 3 The graph shows the carbon assimilation rate of rice. Taking healthy, drought-stricken, and pest-infested rice as examples, healthy rice shows a small fluctuation around 0; drought-stricken rice shows a smaller fluctuation in carbon assimilation rate than pest-infested rice.

[0092] Step S3, Spatial Feature Extraction and Laplace Operator Matrix Solving: In the two-dimensional spatial coordinate system of the target area, calculate the spatial second-order partial derivative of the chlorophyll fluorescence parameter to generate the Laplace operator matrix, thereby extracting spatial features that can reflect the spread of disease and sudden changes in crop growth status in local areas.

[0093] Step S3 includes the following steps:

[0094] Spatial Laplacian operator calculation, in the two-dimensional spatial coordinate system of a pixel. Internally, chlorophyll fluorescence parameters related to carbon assimilation. Calculate the second-order spatial partial derivatives to obtain the Laplace operator matrix. :

[0095]

[0096] Spatial high-frequency mutation feature extraction involves extracting spatial mutation features within the local pixel neighborhood from the Laplacian operator matrix. Spatial mutation characteristic The quantitative criteria are as follows:

[0097] when When this is determined to be a spatial consistency feature, it indicates that the stress is uniformly distributed over a large area within the target region, where... The preset spatial smoothing threshold has a value range of [value range missing]. ;

[0098] when This was determined to be a spatial mutation characteristic, representing the stress spreading outward in patches or concentric circles from the source of infection. The preset spatial mutation threshold, and .

[0099] Figure 4 This is a Laplace spatial feature map of rice. Take 0.08, Take 1. Arid regions The response value is greater than 0.08 and relatively uniform, as shown in the boxed area of ​​the graph; extremely high values ​​will appear at the site of pest and disease occurrence. The response value is greater than 1 and shows a significant abrupt change, which is represented by the circled area in the figure.

[0100] Step S4: Construction of Partial Differential Equations and Threshold Determination of Stress Type: Construct a multidimensional partial differential equation system for stress decoupling conditions, using the temporal characteristics of the Jacobian matrix and the spatial characteristics of the Laplace operator as input parameters; based on the differences in the physiological and physical change rates and spatial distribution characteristics corresponding to drought, nutrient deficiency, and disease, determine the type of stress suffered by the crop through partial differential thresholds.

[0101] Step S4 includes the following steps:

[0102] Acute drought stress identification: when the rate of temperature change is satisfied Equivalent water thickness change rate And spatial mutation characteristic quantity At that time, it was determined to be acute drought stress; among them The threshold for temperature increase, The threshold for the rate of water loss;

[0103] Chronic nutrient deficiency stress identification: when the rate of temperature change is met chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be chronic nutritional deficiency stress; among them This represents the upper limit for environmental noise fluctuations. This represents the photosynthetic inhibition threshold.

[0104] Infectious disease stress identification: when the rate of temperature change is satisfied chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be an infectious disease stress;

[0105] Normal growth state determination: When the absolute value of the time change rate of each dimension parameter is less than the corresponding preset steady-state threshold, and At that time, the crop was judged to be in a normal growth state.

[0106] Table 1 Specific Threshold Recommendations

[0107]

[0108] Step S5: Decoupling of Complex Stresses and Optimization of Diagnostic Atlas Output: Based on the judgment results of the partial differential equation system, drought, nutrient deficiency, and pathogen infection stresses that result in similar spectra are accurately separated at the mathematical level. The results are then spatially smoothed and optimized to finally output a pixel-level diagnostic atlas of crop multi-factor complex stress mechanisms.

[0109] Step S5 includes the following steps:

[0110] The probability transformation of the partial differential judgment result transforms the partial differential threshold judgment result in step 4 into a normalized probability matrix of each pixel belonging to drought, nutrient deficiency, or disease stress.

[0111] The results optimization and diagnostic map output utilize Markov random fields to perform spatiotemporal smoothing optimization on the probability matrix, eliminating local outliers, and finally outputting a pixel-level crop composite stress mechanism diagnostic map that accurately distinguishes the distribution of drought, nutrient deficiency, and pathogen infection.

[0112] Finally, it should be noted that 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 the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space, characterized in that, Includes the following steps: Step S1, Multi-source remote sensing data processing and 3D physical parameter inversion: Obtain multi-frequency and multi-type remote sensing observation data of the target area covering the crop growth stage, perform spatiotemporal coordinate alignment and inversion reconstruction on the data, and obtain a pixel-level 3D physical state vector sequence. This vector contains three core parameters: canopy equivalent temperature, canopy equivalent water thickness, and sunlight-induced chlorophyll fluorescence, which respectively reflect the crop's temperature change, water metabolism, and carbon assimilation process. Step S2, Time Feature Extraction and Jacobi Dynamics Matrix Construction: In a continuous time coordinate system, calculate the first-order rate of change of the above three-dimensional physical state vector with time, and construct the Jacobi dynamics evolution matrix. This matrix contains the rate of change of temperature, moisture, and carbon assimilation, thereby extracting the speed characteristics of the physiological stress response of crops under stress. Step S3, Spatial Feature Extraction and Laplace Operator Matrix Solving: In the two-dimensional spatial coordinate system of the target area, the spatial second-order partial derivative of the chlorophyll fluorescence parameter is calculated to generate the Laplace operator matrix, thereby extracting spatial features that can reflect the spread of disease and sudden changes in crop growth status in local areas. Step S4, Construction of Partial Differential Equations and Threshold Judgment of Stress Type: Construct a multidimensional partial differential equation system for stress decoupling conditions, using the temporal characteristics of the Jacobian matrix and the spatial characteristics of the Laplace operator as input parameters; based on the differences in the physiological and physical change rates and spatial distribution characteristics of drought, nutrient deficiency, and disease, determine the type of stress suffered by the crop through partial differential thresholds; Step S5: Decoupling of Complex Stresses and Optimization of Diagnostic Atlas Output: Based on the judgment results of the partial differential equation system, drought, nutrient deficiency, and pathogen infection stresses that result in similar spectra are accurately separated at the mathematical level. The results are then spatially smoothed and optimized to finally output a pixel-level diagnostic atlas of crop multi-factor complex stress mechanisms.

2. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, The multimodal remote sensing data extraction and preprocessing described in step S1 specifically involves acquiring continuous time-series thermal infrared remote sensing data, shortwave infrared and microwave remote sensing data, and sunlight-induced chlorophyll fluorescence (SIF) remote sensing data of the target area, performing geometric correction on all data, and completing pixel-level spatial coordinate alignment.

3. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, The pixel-level three-dimensional physical state vector inversion described in step S1 specifically involves reconstructing a uniform time step for each spatial pixel. The three-dimensional physical parameters define the pixel in Physical state vector at time t : in, The temperature dimension characterizes the canopy equivalent temperature due to heat dissipation from crop stomata closure. The water dimension characterizes the canopy equivalent water thickness, representing the water loss from crop mesophyll tissue. The carbon assimilation dimension characterizes the absolute value of chlorophyll fluorescence radiation when photosynthetic electron transport is hindered in crops. This represents the transpose of a matrix.

4. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, The continuous time phase space mapping described in step S2 specifically involves mapping the three-dimensional physical state vector of each pixel. This data is mapped onto a continuously changing three-dimensional multidimensional manifold phase space, transforming discrete remote sensing observation data into a continuously changing trajectory of crop physiological changes over time.

5. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, The Jacobian dynamic evolution matrix solution in step S2 specifically involves calculating the three-dimensional physical state vector as a function of calendar time in the aforementioned three-dimensional phase space. First-order partial derivatives are used to construct the Jacobian matrix. This reflects the rate of evolution of crop physiological stress: in, Represents the rate of temperature change. Represents the rate of change in moisture content. This represents the rate of carbon assimilation.

6. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, The spatial Laplacian operator calculation described in step S3 is specifically performed in the two-dimensional spatial coordinate system of the pixel. Internally, chlorophyll fluorescence parameters related to carbon assimilation. Calculate the second-order spatial partial derivatives to obtain the Laplace operator matrix. : 。 7. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, Step S3, the extraction of spatial high-frequency mutation features, specifically involves extracting spatial mutation feature quantities within the local pixel neighborhood from the Laplacian operator matrix. The spatial mutation feature quantity The quantitative criteria are as follows: when When this is determined to be a spatial consistency feature, it indicates that the stress is uniformly distributed over a large area within the target region, where... The preset spatial smoothing threshold has a value range of [value range missing]. ; when This was determined to be a spatial mutation characteristic, representing the stress spreading outward in patches or concentric circles from the source of infection. The preset spatial mutation threshold, and .

8. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that: In step S4, the specific decision logic for the multidimensional stress decoupling condition partial differential equation system is as follows: Acute drought stress identification: when the rate of temperature change is satisfied Equivalent water thickness change rate And spatial mutation characteristic quantity At that time, it was determined to be acute drought stress; among them The threshold for temperature increase, The threshold for the rate of water loss; Chronic nutrient deficiency stress identification: when the rate of temperature change is met chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be chronic nutritional deficiency stress; among them This represents the upper limit for environmental noise fluctuations. This represents the photosynthetic inhibition threshold. Infectious disease stress identification: when the rate of temperature change is satisfied chlorophyll fluorescence change rate And spatial mutation characteristic quantity At that time, it was determined to be an infectious disease stress; Normal growth state determination: When the absolute value of the time change rate of each dimension parameter is less than the corresponding preset steady-state threshold, and At that time, the crop was judged to be in a normal growth state.

9. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, The probability transformation of the partial differential judgment result in step S5 specifically involves transforming the partial differential threshold judgment result in step 4 into a normalized probability matrix for each pixel belonging to drought, nutrient deficiency, or disease stress.

10. The crop stress decoupling discrimination method based on hot water carbon multidimensional manifold phase space according to claim 1, characterized in that, The result optimization and diagnostic map output described in step S5 specifically involves using Markov random fields to perform spatiotemporal smoothing optimization on the probability matrix, eliminating local outliers, and finally outputting a pixel-level crop composite stress mechanism diagnostic map that distinguishes the distribution of drought, nutrient deficiency, and pathogen infection.

Images