Method and system for diagnosing crop water stress based on sunlight-induced chlorophyll fluorescence

Abstract

The present application relates to the technical field of agricultural drought monitoring and precision irrigation, and particularly relates to a crop water stress diagnosis method and system based on sunlight-induced chlorophyll fluorescence, which collects crop canopy oxygen absorption band spectral data through a hyperspectral sensor, and calculates sunlight-induced chlorophyll fluorescence intensity SIF values by using a fluorescence extraction algorithm. The SIF value of a non-water-stressed crop canopy is taken as a reference canopy SIF value, the difference between the SIF value of a to-be-measured crop canopy and the reference canopy SIF value is determined, then a relative SIF water deficit index SIFWDI is generated by combining a preset formula, a regression analysis model is used to establish a mapping relationship between the SIFWDI and a soil water deficit index to output a diagnosis report, a variable irrigation decision and early warning are triggered according to a stress level, and model parameter adaptive optimization is realized through a feedback mechanism. The system realizes non-destructive, real-time and high-precision water stress monitoring and irrigation management.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural drought monitoring and precision irrigation, and in particular to a crop water stress diagnosis method and system based on sunlight-induced chlorophyll fluorescence. BACKGROUND

[0002] Sunlight-induced chlorophyll fluorescence (SIF) is a weak electromagnetic radiation signal emitted by chlorophyll when plants perform photosynthesis under sunlight, and its intensity is closely related to the photochemical efficiency of the light system. When plants are in environmental stress or physiological regulation state, part of the absorbed light energy in the photosynthetic mechanism will be re-emitted in the form of fluorescence. By capturing such signals through remote sensing technology, the actual photosynthetic capacity of vegetation and its response mechanism to environmental changes can be inverted, so SIF can be used to directly diagnose the physiological drought of crops and diagnose their potential water stress conditions.

[0003] The existing crop drought diagnosis technology based on soil water deficit index has the following technical pain points in actual application, specifically: relying on soil probes laid in the field to measure water, the installation process not only damages the crop root layer, but also requires dense point distribution to accurately reflect the overall situation, making it difficult to achieve precise monitoring on a regional scale; at the same time, soil water conditions cannot monitor crop drought conditions from a physiological perspective, and cannot capture early physiological signals of plant drought. It can be seen that the existing crop water deficit diagnosis based on soil moisture faces high cost, limited monitoring range and monitoring lag when applied to intelligent irrigation decision-making. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a crop water stress diagnosis method and system based on sunlight-induced chlorophyll fluorescence, which solves the technical problems that the existing crop drought grade diagnosis technology based on soil water deficit index (SWDI) relies on field physical probe installation, is seriously disturbed by soil spatial heterogeneity, and cannot directly capture early signals of photosynthesis inhibition from the physiological level of plants, resulting in high cost, high destructive, insufficient representation, and lagging early warning response in the diagnosis process, which cannot meet the needs of modern agriculture for low-cost, non-destructive, large-area, and high-precision drought monitoring.

[0005] To solve the above technical problems, the technical content of the present application is as follows:

[0006] In a first aspect, the present application provides a crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, comprising:

[0007] Step S1, acquiring original spectral data of a water-free crop canopy and a to-be-measured crop canopy under the same environment by a hyperspectral sensor mounted on a remote sensing platform, the original spectral data including downwelling irradiance spectrum and upwelling radiance spectrum at oxygen molecule absorption band;

[0008] Step S2: Based on the downlink irradiance spectrum and uplink irradiance spectrum data, and the fluorescence extraction algorithm based on the principle of spectral filling, calculate the apparent reflectance in the oxygen molecule A absorption band, and solve the SIF value of the sunlight-induced chlorophyll fluorescence intensity in the ~760nm band of the crop canopy.

[0009] Step S3: Use the calculated SIF value of the crop canopy without water stress as the reference canopy SIF value, and then substitute the calculated SIF value of the crop canopy to be tested into the preset formula to generate the relative SIF water deficit index SIFWDI.

[0010] Step S4: Construct a regression analysis model between the relative SIF water deficit index SIFWDI and the soil water deficit index SWDI generated based on measured soil moisture data. By establishing the mapping relationship between the relative SIF water deficit index SIFWDI and the soil water deficit index SWDI, output a diagnostic report, which includes the crop water stress level.

[0011] Step S5: Based on the diagnostic report output by S4, if the crop water stress level is determined to exceed the preset threshold, a variable irrigation decision instruction is generated and an early warning is triggered. The relative SIF water deficit index, soil water deficit index SWDI, and crop water stress level corresponding to this determination are used as new data samples and automatically fed back into the regression analysis model to achieve adaptive optimization of the regression analysis model parameters.

[0012] Furthermore, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence described in this invention, step 1 includes:

[0013] Step S101: Install a spectral sensing device at a fixed height above the crop canopy and adjust the probe to point vertically towards the center area of ​​the canopy;

[0014] Step S102: Using a pre-installed spectral sensing device, simultaneously collect downlink irradiance and uplink irradiance data during a specified period of clear weather.

[0015] Step S103: Record and store the collected downlink irradiance and uplink irradiance data in the local storage unit. The data format is a continuous spectral curve covering the oxygen absorption band.

[0016] Furthermore, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence described in this invention, step 2 includes:

[0017] Step S201: Filter and baseline correct the downlink irradiance spectral data and the uplink irradiance spectral data to obtain corrected spectral data;

[0018] Step S202: The corrected spectral data obtained in step 201 is processed using a fluorescence extraction algorithm based on the Fraunhofer line depth principle to calculate the apparent fluorescence emission intensity of feature points inside and outside the dark line in the oxygen absorption band.

[0019] Step S203: Based on the preset parameter adjustment rules, using the apparent fluorescence emission intensity as input, dynamically adjust the band window width and weight parameters in the fluorescence extraction algorithm to extract the final sunlight-induced chlorophyll fluorescence value.

[0020] Furthermore, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence described in this invention, step 3 includes:

[0021] Step S301: Based on step 2, obtain the SIF value of the crop canopy without water stress as the reference canopy SIF value;

[0022] Step S302: Based on the SIF value of the canopy to be tested obtained in step 2, substitute the result into the preset formula for calculation, wherein the preset formula is:

[0023]

[0024] SIF well-watered SIF value of crop canopy without water stress (μW·cm) -2 ·mm -1 ·sr -1 SIF target SIF value of crop canopy (μW·cm) -2 ·mm -1 ·sr -1 SIFWDI: Relative to SIF moisture deficit index;

[0025] Step S304: Output the calculated value as the relative SIF water deficit index (SIFWDI).

[0026] Furthermore, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence described in this invention, step 4 includes:

[0027] Step 401: Construct a regression analysis model by comparing the relative SIF water deficit index SIFWDI with the soil water deficit index SWDI calculated from measured soil moisture data.

[0028] Step 402: The regression analysis model establishes a mapping relationship between the relative soil water deficit index (SIFWDI) and the soil water deficit index (SWDI), and outputs a diagnostic report. The diagnostic report includes the crop water stress level. The crop water stress level is used to characterize the degree of crop drought caused by water deficit. The classification of crop drought level is determined according to the SIFWDI value: SIFWDI < 0 is no drought, 0 ≤ SIFWDI < 1 is mild drought, 1 ≤ SIFWDI < 2 is moderate drought, 2 ≤ SIFWDI < 4 is severe drought, and SIFWDI ≥ 4 is extreme drought.

[0029] Step 403: Determine the final crop drought level based on the crop water stress level in the diagnostic report.

[0030] Furthermore, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence described in this invention, step 5 includes:

[0031] Step S501: Based on the crop water stress level in the diagnostic report output in step S4, monitor the temporal changes of the crop water stress level; when the crop water stress level exceeds the preset threshold in multiple consecutive monitoring periods, activate the early warning mechanism and generate early warning information.

[0032] Step S502: The warning information generated in step S501, along with the geographical location and timestamp, is encapsulated into a data packet and transmitted to the cloud platform via an encrypted communication protocol;

[0033] Step S503: After receiving the data packet, the cloud platform highlights the threatened area on the visualization interface and pushes a warning notification to the bound terminal.

[0034] Step S504: Based on the crop water stress level in the diagnostic report output in step S4, generate differentiated irrigation instructions; the differentiated irrigation instructions include maintenance plans, water-saving irrigation or emergency irrigation plans.

[0035] Furthermore, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence described in this invention, step 5 further includes:

[0036] Step S505: After executing the irrigation command, re-acquire the original spectral data of the crop canopy, and calculate the new sunlight-induced chlorophyll fluorescence intensity value based on the re-acquired data;

[0037] Step S506: Combine the new sunlight-induced chlorophyll fluorescence intensity value with the field soil moisture record, and use the sliding window algorithm to update the parameters of the regression analysis model;

[0038] Step S507: After verifying the accuracy of the regression analysis model, store the updated data in the database.

[0039] Furthermore, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence described in this invention, step 5 further includes:

[0040] Step 508: Based on the time series data of the relative SIF water deficit index generated in step S3, input it into the time series analysis model and output early abnormal signals.

[0041] Step 509: Input the early abnormal signal into the mutation detection algorithm for time-series mutation point analysis to obtain the mutation point location result;

[0042] Step 510: Based on the mutation point location results, the probability of early drought risk is output by calculating the intensity and trend changes of the abnormality within the preset time window.

[0043] Secondly, the crop water stress diagnosis system based on sunlight-induced chlorophyll fluorescence provided by the present invention is applied to the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence as described above, including:

[0044] The data acquisition module, mounted on the remote sensing platform and equipped with a hyperspectral sensor, is used to collect raw spectral data of the crop canopy. The raw spectral data includes the downward irradiance spectrum and the upward irradiance spectrum in the oxygen A absorption band.

[0045] The fluorescence inversion module, which is connected to the data acquisition module, is configured to: calculate the apparent reflectance in the oxygen molecule A absorption band based on the downlink irradiance spectrum and uplink irradiance spectrum data, and the fluorescence extraction algorithm based on the spectral filling principle, and solve the SIF value of the sunlight-induced chlorophyll fluorescence intensity in the ~760nm band of the crop canopy.

[0046] The index generation module, which communicates with the fluorescence inversion module, is configured to: obtain the SIF value of the canopy of a crop without water stress under the current environment; then obtain the SIF value of the canopy of the crop to be tested; substitute the result into a preset formula to generate the relative SIF water deficit index SIFWDI.

[0047] The regression diagnosis module, which communicates with the index generation module, is configured to: construct a regression analysis model between the relative SIF water deficit index (SIFWDI) and the soil water deficit index (SWDI) calculated based on measured soil moisture data; and output a diagnosis report by establishing a mapping relationship between the relative SIF water deficit index (SIFWDI) and the soil water deficit index (SWDI). The diagnosis report includes the crop water stress level; wherein, the crop water stress level is used to characterize the degree of crop drought caused by water deficit, and the classification of crop drought level is determined according to the SIFWDI value: SIFWDI<0 is no drought, 0 ≤ SIFWDI<1 is mild drought, 1 ≤ SIFWDI<2 is moderate drought, 2 ≤ SIFWDI<4 is severe drought, and SIFWDI ≥ 4 is extreme drought.

[0048] The decision optimization module, which communicates with the regression diagnosis module, is configured to: based on the diagnosis report, if the crop water stress level is determined to exceed a preset threshold, generate a variable irrigation decision instruction and trigger an early warning; use the relative SIF water deficit index (SIFWDI), soil water deficit index (SWDI), and crop water stress level corresponding to this determination as new data samples, and automatically feed them back into the regression analysis model to achieve adaptive optimization of the regression analysis model parameters.

[0049] Beneficial effects of this invention:

[0050] The beneficial effects of this invention are reflected in the following aspects: A remote sensing monitoring method based on sunlight-induced chlorophyll fluorescence (SIF) is used. Hyperspectral sensors are employed to non-destructively acquire crop canopy spectral data. A fluorescence extraction algorithm based on the spectral filling principle is used to calculate the SIF value, generating a relative SIF water deficit index (SIFWDI). This replaces traditional soil probe measurements, avoiding damage to crop roots during installation and overcoming the high costs associated with dense deployment. By establishing a mapping relationship between SIFWDI and the SIFWDI, a diagnostic report is output, enabling the direct capture of early signals of photosynthetic inhibition at the plant physiological level, solving the problem of soil moisture changes lagging behind physiological responses. Regression analysis models and adaptive optimization mechanisms improve diagnostic accuracy and real-time performance. Finally, through variable irrigation decisions and early warning triggers, precise irrigation management is supported, water use efficiency is improved, and the needs of modern agriculture for low-cost, non-destructive, large-area, and high-precision drought monitoring are met. Attached Figure Description

[0051] To more clearly illustrate the present invention, the accompanying drawings used in the embodiments will be briefly described below. Obviously, those skilled in the art can obtain other drawings based on the accompanying drawings without any creative effort.

[0052] Figure 1This is a flowchart illustrating the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence provided in an embodiment of the present invention.

[0053] Figure 2 The correlation diagram between the Soil Water Deficiency Index (SWDI) and the Relative SIF Water Stress Index (SIFWDI) provided in the embodiments of the present invention. Detailed Implementation

[0054] To make the objectives, advantages, and features of this invention clearer, the invention will be described clearly and completely below with reference to specific embodiments and accompanying drawings. Obviously, the described embodiments are only one module of the invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention. The invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the objectives of this invention, it will be described in further detail below.

[0055] In a first aspect, the present invention provides a method for diagnosing crop water stress based on sunlight-induced chlorophyll fluorescence, comprising:

[0056] Step S1: Using a hyperspectral sensor mounted on a remote sensing platform, collect raw spectral data of the canopy of the crop under test in the same environment without water stress. The raw spectral data includes the downward irradiance spectrum and the upward irradiance spectrum in the oxygen molecule absorption band (~760 nm).

[0057] Step S2: Based on the downlink irradiance spectrum and uplink irradiance spectrum data, and the fluorescence extraction algorithm based on the principle of spectral filling, calculate the apparent reflectance in the oxygen molecule A absorption band, and solve the SIF value of the sunlight-induced chlorophyll fluorescence intensity in the ~760nm band of the crop canopy.

[0058] Step S3: Use the calculated SIF value of the crop canopy without water stress as the reference canopy SIF value, and then substitute the calculated SIF value of the crop canopy to be tested into the preset formula to generate the relative SIF water deficit index SIFWDI.

[0059] Step S4: Construct a regression analysis model between the relative SIF water deficit index SIFWDI and the soil water deficit index SWDI generated based on measured soil moisture data. By establishing the mapping relationship between the relative SIF water deficit index SIFWDI and the soil water deficit index SWDI, output a diagnostic report, which includes the crop water stress level.

[0060] Step S5: Based on the diagnostic report output by S4, if the crop water stress level is determined to exceed the preset threshold, a variable irrigation decision instruction is generated and an early warning is triggered. The relative SIF water deficit index corresponding to this judgment and the crop water stress level of this judgment are used as new data samples and automatically fed back into the regression analysis model to achieve adaptive optimization of the regression analysis model parameters.

[0061] The crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence first achieves data acquisition in step S1. Specifically, a hyperspectral sensor mounted on a remote sensing platform is used to collect raw spectral data of the crop canopy, covering the downward irradiance spectrum and upward irradiance spectrum of the oxygen A absorption band. The hyperspectral sensor is installed at a fixed height above the crop canopy, with the probe vertically pointed towards the center of the canopy. During a specified period in clear weather, downward irradiance and upward irradiance data are collected simultaneously and recorded as a continuous spectral curve covering the oxygen absorption band. This step ensures the acquisition of high-quality spectral data, providing a foundation for subsequent processing.

[0062] Step S2 processes the acquired spectral data to extract the sunlight-induced chlorophyll fluorescence intensity (SIF) values. Downward and upward irradiance spectral data are filtered and baseline-corrected to eliminate noise and environmental interference. The corrected spectral data are then processed using a fluorescence extraction algorithm based on the Fraunhofer line depth principle to calculate the apparent fluorescence emission intensity of characteristic points inside and outside the dark line in the oxygen absorption band, and finally, the sunlight-induced chlorophyll fluorescence intensity (SIF) values ​​in the ~760 nm band of the crop canopy are calculated. This step ensures accurate extraction of the fluorescence signal through the spectral filling principle, providing crucial input for water stress diagnosis.

[0063] Step S3 generates the Relative SIF Water Deficiency Index (SIFWDI) to quantify the degree of water stress. The SIF value of the canopy of a crop without water stress is used as a reference, while the SIF value of the canopy of the crop to be tested is also obtained. The reference SIF value and the SIF value of the crop to be tested are substituted into a preset formula for calculation. The preset formula is based on the difference in fluorescence intensity between the no-water-stress state and the tested state, outputting the Relative SIF Water Deficiency Index (SIFWDI). This index directly reflects the degree of crop water deficit, replacing traditional soil probe measurements.

[0064] Step S4 uses a regression analysis model to output a diagnostic report. A regression analysis model is constructed by comparing the relative soil moisture deficit index (SIFWDI) with the soil moisture deficit index (SWDI) calculated from measured soil moisture data. By establishing a mapping relationship between SIFWDI and SWDI, a diagnostic report is generated. The diagnostic report includes the crop water stress level, which is categorized based on SIFWDI values ​​as no drought, mild drought, moderate drought, severe drought, and extreme drought, characterizing the degree of crop drought caused by water deficit.

[0065] Step S5 executes decision-making and feedback optimization based on the diagnostic report. If the crop water stress level exceeds a preset threshold and the dominant factor is soil moisture deficit, a variable irrigation decision instruction is generated and an early warning is triggered. The instruction includes a maintenance plan, water-saving irrigation, or emergency irrigation scheme. Simultaneously, the relative soil moisture deficit index (SIFWDI) and crop water stress level corresponding to this judgment are used as new data samples and automatically fed back into the regression analysis model to achieve adaptive optimization of model parameters. This closed-loop feedback mechanism continuously improves system accuracy by re-collecting data, calculating new fluorescence values, and combining them with field soil moisture records, using a sliding window algorithm to ensure the real-time nature and accuracy of the diagnosis. The entire method integrates optical sensing and data processing to achieve non-destructive and efficient crop water stress monitoring and irrigation decision support.

[0066] Specifically, the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence of the present invention includes step 1 as follows:

[0067] Step S101: Install a spectral sensing device at a fixed height above the crop canopy and adjust the probe to point vertically towards the center area of ​​the canopy;

[0068] Step S102: Using a pre-installed spectral sensing device, simultaneously collect downlink irradiance and uplink irradiance data during a specified period of clear weather.

[0069] Step S103: Record and store the collected downlink irradiance and uplink irradiance data in the local storage unit. The data format is a continuous spectral curve covering the oxygen absorption band.

[0070] The crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence involves a data acquisition process in step 1, specifically comprising three sub-steps. In step S101, the spectral sensing device is installed at a fixed height above the crop canopy. The installation height is determined based on crop growth characteristics and remote sensing platform requirements, typically maintained within 1.5 to 2.5 meters above the canopy to minimize environmental interference. The probe is adjusted to point vertically towards the center of the canopy, ensuring the optical path is perpendicular to the canopy surface, thereby accurately capturing downward and upward irradiance data and avoiding spectral distortion caused by angular deviations. In step S102, using the pre-installed spectral sensing device, downward and upward irradiance data are simultaneously collected daily from 10:00 to 14:00 under clear, cloudless weather conditions. This time period is chosen based on maximizing the solar altitude angle to reduce atmospheric scattering and shadow effects, improving data quality. During the acquisition process, the device is kept stable, avoiding movement or vibration to ensure spectral continuity. In step S103, the collected downlink and uplink irradiance data are recorded and stored in real time to a local storage unit. The storage unit uses a solid-state drive or embedded memory. The data format is a continuous spectral curve covering the oxygen absorption band, including the band near 760nm of the oxygen molecule A absorption band, with a spectral resolution better than 1nm to ensure the accuracy of subsequent fluorescence extraction. The stored data includes a timestamp and geographic location information for easy traceability and integrated processing. These steps are logically coherent: S101 establishes the hardware foundation, S102 optimizes the acquisition conditions to obtain raw data, and S103 ensures complete data storage, providing reliable input for subsequent spectral processing and fluorescence inversion. The entire data acquisition process aims to minimize external interference, ensuring that the spectral data represents the true state of the crop canopy and supports the accuracy of water stress diagnosis.

[0071] Specifically, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence of the present invention, step 2 includes:

[0072] Step S201: Filter and baseline correct the downlink irradiance spectral data and the uplink irradiance spectral data to obtain corrected spectral data;

[0073] Step S202: The corrected spectral data obtained in step 201 is processed using a fluorescence extraction algorithm based on the Fraunhofer line depth principle to calculate the apparent fluorescence emission intensity of feature points inside and outside the dark line in the oxygen absorption band.

[0074] Step S203: Based on the preset parameter adjustment rules, using the apparent fluorescence emission intensity as input, dynamically adjust the band window width and weight parameters in the fluorescence extraction algorithm to extract the final sunlight-induced chlorophyll fluorescence value.

[0075] In the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, step 2 involves spectral data processing and fluorescence extraction, specifically including three sub-steps. Step S201 filters and performs baseline correction on the downlink irradiance spectral data and the uplink irradiance spectral data to eliminate environmental noise and instrument errors. Filtering uses digital filtering algorithms such as moving average or Gaussian filtering to smooth random fluctuations. Baseline correction restores the true spectral characteristics by fitting the spectral baseline and subtracting the offset, outputting the corrected spectral data to provide clean input for subsequent analysis. Step S202 uses a fluorescence extraction algorithm based on the Fraunhofer line depth principle to process the corrected spectral data. The algorithm focuses on the oxygen absorption band, such as around 760 nm, and calculates the irradiance difference between feature points inside and outside the dark line. The area inside the dark line corresponds to the absorption trough point, and the area outside the dark line corresponds to the continuous spectral point. By comparing the apparent fluorescence emission intensity of these points, the sunlight-induced chlorophyll fluorescence signal is preliminarily estimated. Step S203, based on preset parameter adjustment rules, uses apparent fluorescence emission intensity as input to dynamically adjust the band window width and weight parameters in the fluorescence extraction algorithm. The band window width defines the spectral calculation range, and the weight parameters optimize the contribution ratio of different bands. Through iterative optimization, the final sunlight-induced chlorophyll fluorescence value is extracted, which is used for subsequent water stress index calculation. The logic of the entire step 2 is coherent, from data preprocessing to fluorescence extraction, with progressively refined processing to ensure the accuracy and reliability of the fluorescence signal and support the accuracy of water stress diagnosis.

[0076] Specifically, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence of the present invention, step 3 includes:

[0077] Step S301: Based on step 2, obtain the SIF value of the crop canopy without water stress as the reference canopy SIF value;

[0078] Step S302: Based on the SIF value of the canopy to be tested obtained in step 2, substitute the result into the preset formula for calculation, wherein the preset formula is:

[0079]

[0080] SIF well-watered SIF value of crop canopy without water stress (μW·cm) -2 ·mm -1 ·sr -1 SIF target SIF value of crop canopy (μW·cm) -2 ·mm -1 ·sr -1 SIFWDI: Relative to SIF moisture deficit index;

[0081] Step S304: Output the calculated value as the relative SIF water deficit index (SIFWDI).

[0082] In the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, step 3 involves generating the relative SIF water deficit index (SIFWDI), which includes three sub-steps. Step S301 obtains the SIF value of the canopy of a crop without water stress as a reference canopy SIF value. The reference SIF value represents the typical sunlight-induced chlorophyll fluorescence intensity of the canopy of a healthy crop, providing a benchmark for subsequent calculations. Step S302, based on the test canopy SIF value obtained in step 2, substitutes the reference SIF value and the test SIF value into a preset formula for calculation. The preset formula quantifies the degree of water deficit by comparing the fluorescence intensity difference between the no-water-stress state and the test state. The calculation process focuses on the relative change of fluorescence signal, avoiding dependence on absolute values, and outputs a preliminary index result. Step S304 outputs the calculated value as the relative SIF water deficit index (SIFWDI), which directly reflects the crop water stress level and is used for subsequent diagnostic report generation. The logic of the entire step 3 is coherent, from data retrieval to index calculation, ensuring that the index is based on actual physiological signals and supports accurate assessment of water stress.

[0083] Specifically, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence of the present invention, step 4 includes:

[0084] Step 401: Construct a regression analysis model by comparing the relative SIF water deficit index SIFWDI with the soil water deficit index SWDI calculated from measured soil moisture data.

[0085] Step 402: The regression analysis model establishes a mapping relationship between the relative soil water deficit index (SIFWDI) and the soil water deficit index (SWDI), and outputs a diagnostic report. The diagnostic report includes the crop water stress level. The crop water stress level is used to characterize the degree of crop drought caused by water deficit. The classification of crop drought level is determined according to the SIFWDI value: SIFWDI < 0 is no drought, 0 ≤ SIFWDI < 1 is mild drought, 1 ≤ SIFWDI < 2 is moderate drought, 2 ≤ SIFWDI < 4 is severe drought, and SIFWDI ≥ 4 is extreme drought.

[0086] Step 403: Determine the final crop drought level based on the crop water stress level in the diagnostic report.

[0087] In the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, step 4 involves the generation of a diagnostic report and the determination of drought level, specifically including three sub-steps. Step 401 constructs a regression analysis model by combining the relative soil water deficit index (SIFWDI) and the soil water deficit index (SWDI) calculated from measured soil moisture data. The regression analysis model uses a linear or nonlinear regression algorithm and is trained using a historical dataset to obtain a mapping function. The historical data includes SIFWDI values ​​and corresponding SWDI values ​​for different crop varieties across multiple growing seasons. The model's input layer receives SIFWDI and soil moisture data, and the output layer predicts the SWDI value, thus realizing the conversion from optical signals to soil moisture status. In step 402, the regression analysis model establishes a mapping relationship between the relative soil water deficit index (SIFWDI) and the soil water deficit index (SWDI), and outputs a diagnostic report. The diagnostic report includes the crop water stress level. The mapping relationship is based on least squares or machine learning to optimize parameters so that the SIFWDI value and the SWDI value are negatively correlated, that is, an increase in the SIFWDI value indicates an increase in water stress. The crop water stress level is used to characterize the degree of crop drought caused by water deficit. The level is determined based on the SIFWDI value: SIFWDI < 0 is no drought, 0 ≤ SIFWDI < 1 is mild drought, 1 ≤ SIFWDI < 2 is moderate drought, 2 ≤ SIFWDI < 4 is severe drought, and SIFWDI ≥ 4 is extreme drought. This classification is based on statistical data from field trials and reflects the quantitative relationship between crop physiological response and water deficit. Step 403 determines the final drought level of the crop based on the crop water stress level in the diagnostic report. The diagnostic report is output in numerical or categorical form, and the system automatically matches the level description. For example, when the SIFWDI value is 1.5, it corresponds to the moderate drought level, triggering the corresponding early warning and irrigation decision. The logic of the entire step 4 is coherent, from data input and model processing to result output, ensuring that the diagnosis is based on multi-source data fusion, improving the accuracy and reliability of water stress assessment, and supporting agricultural management decisions.

[0088] Specifically, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence of the present invention, step 5 includes:

[0089] Step S501: Based on the crop water stress level in the diagnostic report output in step S4, monitor the temporal changes of the crop water stress level; when the crop water stress level exceeds the preset threshold in multiple consecutive monitoring periods, activate the early warning mechanism and generate early warning information.

[0090] Step S502: The warning information generated in step S501, along with the geographical location and timestamp, is encapsulated into a data packet and transmitted to the cloud platform via an encrypted communication protocol;

[0091] Step S503: After receiving the data packet, the cloud platform highlights the threatened area on the visualization interface and pushes a warning notification to the bound terminal.

[0092] Step S504: Based on the crop water stress level in the diagnostic report output in step S4, generate differentiated irrigation instructions; the differentiated irrigation instructions include maintenance plans, water-saving irrigation or emergency irrigation plans.

[0093] In the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, step 5 involves early warning generation and irrigation decision-making, specifically including four sub-steps. Step S501, based on the crop water stress level in the diagnostic report output in step S4, monitors the temporal changes in the crop water stress level, using a time series analysis algorithm to track level fluctuations. When the crop water stress level exceeds a preset threshold for multiple consecutive monitoring periods, an early warning mechanism is activated and early warning information is generated. The early warning information includes the stress level value, change trend, and duration, indicating the severity of water deficit. Step S502 encapsulates the early warning information generated in step S501 with geographic coordinates and a timestamp into a structured data packet. The data packet is organized in JSON or XML format and transmitted to the cloud platform via encrypted communication protocols such as TLS or SSL to ensure data transmission security and integrity. In step S503, after receiving the data packet, the cloud platform parses the early warning information and highlights the stressed area on the visualization interface. The visualization interface integrates a geographic information system map, rendering the stress intensity with color gradients. Simultaneously, it pushes early warning notifications to bound terminals such as mobile apps or monitoring centers, including the stress location, level, and recommended measures. Step S504, based on the crop water stress level in the diagnostic report output in step S4, generates differentiated irrigation instructions. The instruction generation module selects the appropriate strategy according to the level range: outputting a maintenance plan instruction when there is no drought or mild drought; outputting a water-saving irrigation instruction, such as reducing irrigation volume, when there is moderate drought; and outputting an emergency irrigation instruction, such as increasing irrigation frequency, when there is severe or extreme drought. The instructions are then distributed to the irrigation execution equipment via the cloud platform. The entire logic of step 5 is coherent, from monitoring and early warning to instruction generation, achieving automated response, ensuring timely intervention in water stress, and supporting healthy crop growth.

[0094] Specifically, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence of the present invention, step 5 further includes:

[0095] Step S505: After executing the irrigation command, re-acquire the original spectral data of the crop canopy, and calculate the new sunlight-induced chlorophyll fluorescence intensity value based on the re-acquired data;

[0096] Step S506: Combine the new sunlight-induced chlorophyll fluorescence intensity value with the field soil moisture record, and use the sliding window algorithm to update the parameters of the regression analysis model;

[0097] Step S507: After verifying the accuracy of the regression analysis model, store the updated data in the database.

[0098] Detailed technical solution

[0099] In the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, step 5 further includes three sub-steps for post-irrigation data updates and model optimization. Step S505, after executing the irrigation command, re-acquires the original spectral data of the crop canopy. Specifically, it uses a hyperspectral sensor mounted on a remote sensing platform to acquire the downward irradiance spectrum and upward irradiance spectrum of the oxygen molecule A absorption band under the same conditions. Based on the re-acquired spectral data, a fluorescence extraction algorithm based on the spectral filling principle is used to calculate the apparent reflectance of the oxygen absorption band, and the sunlight-induced chlorophyll fluorescence intensity (SIF) value of the crop canopy in the ~760nm band is calculated. This process ensures that the latest fluorescence signal after irrigation is obtained, reflecting changes in crop water status.

[0100] Step S506 combines the new sunlight-induced chlorophyll fluorescence intensity value with the field soil moisture record; the sliding window algorithm is used to process the time series data. The size of the sliding window is set according to the data acquisition frequency and crop growth cycle. The algorithm extracts the correlation characteristics between fluorescence intensity and soil moisture data within the window, and dynamically updates the coefficients and mapping relationship of the regression analysis model to adapt the model to the new data distribution and improve the diagnostic accuracy.

[0101] Step S507 verifies the accuracy of the regression analysis model. Verification methods include calculating prediction error, correlation coefficient, or cross-validation index. After the accuracy meets the standard, the updated baseline data and model parameters are stored in the database. The database adopts a relational or time-series database structure, supporting fast querying and subsequent analysis, completing data closed-loop management. The entire additional process in step 5 is logically coherent: S505 acquires new data, S506 processes and updates the model, and S507 verifies and stores the data, ensuring continuous system optimization and enhancing the reliability and real-time performance of water stress diagnosis.

[0102] Specifically, in the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence of the present invention, step 5 further includes:

[0103] Step 508: Based on the time series data of the relative SIF water deficit index generated in step S3, input it into the time series analysis model and output early abnormal signals.

[0104] Step 509: Input the early abnormal signal into the mutation detection algorithm for time-series mutation point analysis to obtain the mutation point location result;

[0105] Step 510: Based on the mutation point location results, the probability of early drought risk is output by calculating the intensity and trend changes of the abnormality within the preset time window.

[0106] In the crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, step 5 further includes three sub-steps for early abnormal signal detection and drought risk probability assessment. Step 508, based on the time-series data of the relative SIF water deficit index generated in step S3, inputs it into the time-series analysis model and outputs early abnormal signals. The time-series analysis model uses a time series decomposition algorithm to separate trend components, seasonal components, and residual components. By monitoring the fluctuation amplitude and frequency of the residual components, it identifies abnormal points that deviate from the normal pattern. Early abnormal signals are manifested as a sudden increase or decrease in the plant physiological water stress index at consecutive time points, indicating the potential occurrence of water stress.

[0107] Step 509 inputs the early abnormal signal into the mutation detection algorithm for time series mutation point analysis to obtain the mutation point location result. The mutation detection algorithm is based on the cumulative sum control chart or Bayesian change point detection method to calculate the statistical characteristic changes of time series data. When the data distribution parameters such as mean or variance change significantly, they are marked as mutation points. The mutation point location result provides the specific time location of the abnormal signal, which is used to accurately locate the start time of water stress.

[0108] Step 510, based on the mutation point location results, calculates the intensity and trend change of the abnormality within a preset time window to output the early drought risk probability. The preset time window is set according to the crop growth cycle and monitoring frequency. The intensity of the abnormality is obtained by integrating the amplitude of the abnormal signal over the time window. The trend change is quantified using the slope of linear regression analysis. Combining the intensity and trend direction, the probability value is calculated. The early drought risk probability is expressed as a value between 0 and 1, with a higher probability indicating a greater likelihood of drought. The entire additional process in Step 5 is logically coherent: S508 detects abnormal signals, S509 locates mutation points, and S510 assesses the risk probability, achieving a closed loop from data monitoring to risk quantification, supporting early warning and intervention decisions.

[0109] Secondly, the crop water stress diagnosis system based on sunlight-induced chlorophyll fluorescence provided by the present invention, applied to the aforementioned crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence, includes:

[0110] The data acquisition module, mounted on the remote sensing platform and equipped with a hyperspectral sensor, is used to collect raw spectral data of the crop canopy. The raw spectral data includes the downward irradiance spectrum and the upward irradiance spectrum in the oxygen A absorption band.

[0111] The fluorescence inversion module, which is connected to the data acquisition module, is configured to: calculate the apparent reflectance in the oxygen molecule A absorption band based on the downlink irradiance spectrum and uplink irradiance spectrum data, and the fluorescence extraction algorithm based on the spectral filling principle, and solve the SIF value of the sunlight-induced chlorophyll fluorescence intensity in the ~760nm band of the crop canopy.

[0112] The index generation module, communicating with the fluorescence inversion module, is configured to: acquire the SIF value of the crop canopy under the current environment without water stress; then acquire the SIF value of the crop canopy to be tested; substitute the results into a preset formula to generate the relative SIF water deficit index SIFWDI. ;

[0113] The regression diagnosis module, communicating with the index generation module, is configured to: input the relative SIF water deficit index (SIFWDI) and measured soil moisture data into a regression analysis model built based on historical data; establish a mapping relationship between the relative SIF water deficit index (SIFWDI) and the soil moisture deficit index (SWDI); and output a diagnosis report. The diagnosis report includes the crop water stress level; the crop water stress level is used to characterize the degree of crop drought caused by water deficit. The classification of crop drought level is determined according to the SIFWDI value: SIFWDI < 0 indicates no drought, 0 ≤ SIFWDI < 1 indicates mild drought, 1 ≤ SIFWDI < 2 indicates moderate drought, 2 ≤ SIFWDI < 4 indicates severe drought, and SIFWDI ≥ 4 indicates extreme drought.

[0114] The decision optimization module, which communicates with the regression diagnosis module, is configured to: based on the diagnosis report, if the crop water stress level is determined to exceed a preset threshold and the dominant factor is soil water deficit, generate a variable irrigation decision instruction and trigger an early warning; use the relative SIF water deficit index (SIFWDI) and the crop water stress level determined in this judgment as new data samples, and automatically feed them back to the regression analysis model to achieve adaptive optimization of the regression analysis model parameters.

[0115] The embodiments of the present invention are as follows:

[0116] To verify the applicability of this invention, winter wheat from a certain region was selected as the test crop. Three water treatments were first established: full irrigation (lower limit of 75% field capacity, upper limit of 95% field capacity); slight deficit irrigation (60-70% of full irrigation volume); and moderate deficit irrigation (50-60% of full irrigation volume). Winter wheat was irrigated with 40 mm of water at sowing and after entering the greening stage, respectively, as emergence water and greening water. Drought treatment began at the jointing and heading stage.

[0117] Observation indicators: The downward and upward irradiance of the vegetation canopy were observed in real time using a solar-induced chlorophyll fluorescence spectrometer (SIF), and the solar-induced chlorophyll fluorescence index of the vegetation was calculated using the 3FLD algorithm. On a clear, cloudless day 10 days after the drought treatment, the average SIF index of different irrigation treatment plots was calculated between 10:00 AM and 2:00 PM.

[0118] The average soil moisture content for the above-mentioned period was calculated simultaneously using soil moisture probes.

[0119] The soil water deficit index (SWDI) calculated from soil moisture content showed a strong linear relationship with the soil water stress index (SIFWDI) calculated from the daylight-induced chlorophyll fluorescence index (SIFWDI = -0.4045SWDI, R² = 0.7427). Figure 2 Furthermore, based on the drought levels defined by the SWDI value, the SIFWDI value can also be used to define corresponding drought levels (Table 1), namely, the SIFWDI values ​​for extreme drought, severe drought, moderate drought, mild drought, and no drought are >4, 2~4, 1~2, 0~1, and <0, respectively. Drought early warning can be issued in the future based on the monitored SIF index.

[0120] Table 1. Aridity levels based on SIFWDI values ​​corresponding to SWDI values.

[0121]

[0122] This invention addresses the problems of existing crop drought diagnosis techniques based on soil moisture deficit index (SMI) that rely on field physical probe installation, are severely affected by soil spatial heterogeneity, and cannot directly capture early photosynthetic inhibition signals from a plant physiological perspective. First, a hyperspectral sensor mounted on a remote sensing platform collects raw spectral data of the crop canopy, including the downward and upward irradiance spectra of the oxygen A absorption band. This avoids damage to the crop root system caused by deploying physical probes in the field, achieving non-destructive data acquisition. Second, a fluorescence extraction algorithm based on the spectral filling principle calculates the SMI value, which directly reflects the photochemical efficiency of the crop's photosynthetic apparatus. This captures early photosynthetic inhibition signals caused by water stress from a plant physiological perspective, overcoming the limitation that soil moisture changes lag behind physiological responses. Then, a relative SMI water deficit index (SIFWDI) is generated. By comparing the fluorescence intensity difference between the no-water-stress state and the test state, the degree of water stress is quantified, replacing traditional soil probe measurements and reducing the high cost of densely deployed probes. Furthermore, SIFWDI and measured soil moisture data are input into a regression analysis model. By establishing a mapping relationship, a diagnostic report is output, including the crop water stress level. Combining multi-source data improves diagnostic accuracy. Finally, based on the diagnostic report, early warnings are triggered and irrigation decisions are generated. A feedback mechanism updates baseline data and model parameters, achieving adaptive optimization and supporting real-time early warning and precision irrigation. This meets the needs of modern agriculture for non-destructive, real-time, and high-precision drought monitoring. The entire solution integrates optical remote sensing and data processing, achieving a closed loop from direct monitoring of plant physiological signals to water stress diagnosis, effectively solving the problems of high cost, limited monitoring range, and monitoring lag in existing technologies.

Claims

1. A method for diagnosing crop water stress based on sunlight-induced chlorophyll fluorescence, characterized in that, include: Step S1: Using a hyperspectral sensor mounted on a remote sensing platform, collect raw spectral data of the canopy of the crop under test in the same environment without water stress. The raw spectral data includes the downward irradiance spectrum and the upward irradiance spectrum in the oxygen molecule absorption band. Step S2: Based on the downlink irradiance spectrum and uplink irradiance spectrum data, and the fluorescence extraction algorithm based on the principle of spectral filling, calculate the apparent reflectance in the oxygen molecule A absorption band, and solve the SIF value of the sunlight-induced chlorophyll fluorescence intensity in the ~760nm band of the crop canopy. Step S3: Use the calculated SIF value of the crop canopy without water stress as the reference canopy SIF value, and then substitute the calculated SIF value of the crop canopy to be tested into the preset formula to generate the relative SIF water deficit index SIFWDI. Step 3 includes: Step S301: Based on step 2, obtain the SIF value of the crop canopy without water stress as the reference canopy SIF value; Step S302: Based on the SIF value of the canopy to be tested obtained in step 2, substitute the result into the preset formula for calculation, wherein the preset formula is: SIF well-watered SIF value of crop canopy without water stress, μW·cm -2 ·mm -1 ·sr -1 SIF target SIF value of the crop canopy to be tested, μW·cm -2 ·mm -1 ·sr -1 SIFWDI: Relative to SIF moisture deficit index; Step S304: Output the calculated value as the relative SIF water deficit index (SIFWDI). Step S4: Construct a regression analysis model between the relative SIF water deficit index SIFWDI and the soil water deficit index SWDI generated based on measured soil moisture data. By establishing the mapping relationship between the relative SIF water deficit index SIFWDI and the soil water deficit index SWDI, output a diagnostic report, which includes the crop water stress level. Step S5: Based on the diagnostic report output by S4, if the crop water stress level is determined to exceed the preset threshold, a variable irrigation decision instruction is generated and an early warning is triggered. The relative SIF water deficit index, soil water deficit index SWDI, and crop water stress level corresponding to this determination are used as new data samples and automatically fed back into the regression analysis model to achieve adaptive optimization of the regression analysis model parameters.

2. The crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence according to claim 1, characterized in that, Step 1 includes: Step S101: Install a spectral sensing device at a fixed height above the crop canopy and adjust the probe to point vertically towards the center area of ​​the canopy; Step S102: Using a pre-installed spectral sensing device, simultaneously collect downlink irradiance and uplink irradiance data during a specified period of clear weather. Step S103: Record and store the collected downlink irradiance and uplink irradiance data in the local storage unit. The data format is a continuous spectral curve covering the oxygen absorption band.

3. The crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence according to claim 2, characterized in that, Step 2 includes: Step S201: Filter and baseline correct the downlink irradiance spectral data and the uplink irradiance spectral data to obtain corrected spectral data; Step S202: The corrected spectral data obtained in step 201 is processed using a fluorescence extraction algorithm based on the Fraunhofer line depth principle to calculate the apparent fluorescence emission intensity of feature points inside and outside the dark line in the oxygen absorption band. Step S203: Based on the preset parameter adjustment rules, using the apparent fluorescence emission intensity as input, dynamically adjust the band window width and weight parameters in the fluorescence extraction algorithm to extract the final sunlight-induced chlorophyll fluorescence value.

4. The crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence according to claim 3, characterized in that, Step 4 includes: Step 401: Construct a regression analysis model by comparing the relative SIF water deficit index SIFWDI with the soil water deficit index SWDI calculated from measured soil moisture data. Step 402: The regression analysis model establishes a mapping relationship between the relative soil water deficit index (SIFWDI) and the soil water deficit index (SWDI), and outputs a diagnostic report. The diagnostic report includes the crop water stress level. The crop water stress level is used to characterize the degree of crop drought caused by water deficit. The classification of crop drought level is determined according to the SIFWDI value: SIFWDI < 0 is no drought, 0 ≤ SIFWDI < 1 is mild drought, 1 ≤ SIFWDI < 2 is moderate drought, 2 ≤ SIFWDI < 4 is severe drought, and SIFWDI ≥ 4 is extreme drought. Step 403: Determine the final crop drought level based on the crop water stress level in the diagnostic report.

5. The crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence according to claim 4, characterized in that, Step 5 includes: Step S501: Based on the crop water stress level in the diagnostic report output in step S4, monitor the temporal changes of the crop water stress level; when the crop water stress level exceeds the preset threshold in multiple consecutive monitoring periods, activate the early warning mechanism and generate early warning information. Step S502: The warning information generated in step S501, along with the geographical location and timestamp, is encapsulated into a data packet and transmitted to the cloud platform via an encrypted communication protocol; Step S503: After receiving the data packet, the cloud platform highlights the threatened area on the visualization interface and pushes a warning notification to the bound terminal. Step S504: Based on the crop water stress level in the diagnostic report output in step S4, generate differentiated irrigation instructions; the differentiated irrigation instructions include maintenance plans, water-saving irrigation or emergency irrigation plans.

6. The crop water stress diagnosis method based on sunlight-induced chlorophyll fluorescence according to claim 5, characterized in that, Step 5 also includes: Step S505: After executing the irrigation command, re-acquire the original spectral data of the crop canopy, and calculate the new sunlight-induced chlorophyll fluorescence intensity value based on the re-acquired data; Step S506: Combine the new sunlight-induced chlorophyll fluorescence intensity value with the field soil moisture record, and use the sliding window algorithm to update the parameters of the regression analysis model; Step S507: After verifying the accuracy of the regression analysis model, store the updated data in the database.

7. The method for diagnosing crop water stress based on sunlight-induced chlorophyll fluorescence according to claim 6, characterized in that, Step 5 also includes: Step 508: Based on the time series data of the relative SIF water deficit index generated in step S3, input it into the time series analysis model and output early abnormal signals. Step 509: Input the early abnormal signal into the mutation detection algorithm for time-series mutation point analysis to obtain the mutation point location result; Step 510: Based on the mutation point location results, the probability of early drought risk is output by calculating the intensity and trend changes of the abnormality within the preset time window.

8. A crop water stress diagnostic system based on sunlight-induced chlorophyll fluorescence, applied to the crop water stress diagnostic method based on sunlight-induced chlorophyll fluorescence as described in any one of claims 1 to 7, characterized in that, include: The data acquisition module, mounted on the remote sensing platform, is equipped with a hyperspectral sensor to collect raw spectral data of the crop canopy; The raw spectral data include the descending irradiance spectrum and the ascending irradiance spectrum of the oxygen molecule A absorption band; The fluorescence inversion module, which is connected to the data acquisition module, is configured to: calculate the apparent reflectance in the oxygen molecule A absorption band based on the downlink irradiance spectrum and uplink irradiance spectrum data, and the fluorescence extraction algorithm based on the spectral filling principle, and solve the SIF value of the sunlight-induced chlorophyll fluorescence intensity in the ~760nm band of the crop canopy. The index generation module, communicating with the fluorescence inversion module, is configured to: acquire the SIF value of the crop canopy under the current environment without water stress; then acquire the SIF value of the crop canopy to be tested; substitute the results into a preset formula to generate the relative SIF water deficit index SIFWDI. ; The regression diagnosis module, which communicates with the index generation module, is configured to: construct a regression analysis model between the relative SIF water deficit index (SIFWDI) and the soil water deficit index (SWDI) calculated based on measured soil moisture data; and output a diagnosis report by establishing a mapping relationship between the relative SIF water deficit index (SIFWDI) and the soil water deficit index (SWDI). The diagnosis report includes the crop water stress level; whereby the crop water stress level is used to characterize the degree of crop drought caused by water deficit, and the classification of crop drought level is determined according to the SIFWDI value: SIFWDI < 0 is no drought, 0 ≤ SIFWDI < 1 is mild drought, 1 ≤ SIFWDI < 2 is moderate drought, 2 ≤ SIFWDI < 4 is severe drought, and SIFWDI ≥ 4 is extreme drought. The decision optimization module, which communicates with the regression diagnosis module, is configured to: based on the diagnosis report, if the crop water stress level is determined to exceed a preset threshold, generate a variable irrigation decision instruction and trigger an early warning; use the relative SIF water deficit index (SIFWDI), soil water deficit index (SWDI), and crop water stress level corresponding to this determination as new data samples, and automatically feed them back into the regression analysis model to achieve adaptive optimization of the regression analysis model parameters.

Images