Crop water stress diagnostic method based on multi-feature fusion of SIF diurnal curve

Abstract

The present application relates to the technical field of agricultural drought monitoring, and discloses a crop water stress diagnosis method based on SIF daily variation curve multi-feature fusion, which obtains crop canopy SIF and environmental observation data, obtains apparent fluorescence yield through pretreatment and illumination normalization; constructs a waterless stress SIFy daily variation physiological baseline based on historical data in the normal water supply period, compares the measured value to determine the SIFy physiological deviation amount; determines the turning point of the transition from light limitation to water limitation state according to the fluorescence light response sensitivity and the SIFy physiological deviation amount; extracts multiple daily variation stress features based on the SIFy physiological deviation amount and the turning point and carries out feature fusion to obtain a comprehensive water stress index, and divides the crop water stress grade according to the index. The present application can effectively exclude non-water stress factor interference, and realize early, accurate and comprehensive diagnosis of water stress.

Description

Technical Field

[0001] This invention relates to the field of agricultural drought monitoring technology, and in particular to a method for diagnosing crop water stress based on the fusion of multiple features of SIF diurnal variation curves. Background Technology

[0002] Sunlight-induced chlorophyll fluorescence (SIF) is a weak electromagnetic radiation signal emitted by chlorophyll during photosynthesis in plants under sunlight. Its intensity is closely related to the photochemical efficiency of the photosystem. Due to the mechanistic coupling between SIF and vegetation photosynthesis, SIF is highly sensitive to dynamic changes in plant water content, reflecting the actual photosynthetic capacity of vegetation and its response mechanism to environmental changes. Compared to traditional vegetation indices based on greenness, such as the normalized difference vegetation index (NDI), which mainly reflect structural changes such as leaf loss, chlorophyll reduction, or leaf curling, but are not sensitive enough to rapid changes in photosynthetic function, SIF is considered one of the most promising remote sensing signals for early stress detection. Therefore, SIF has been widely used in various fields such as photosynthetic productivity estimation, crop yield prediction, drought monitoring, evapotranspiration estimation, and vegetation phenology tracking.

[0003] However, existing SIF-based crop water stress diagnosis technologies still have many shortcomings. First, the canopy SIF signal is jointly controlled by multiple factors, including leaf physiology, canopy structure, radiation intensity, and solar-observation geometry. Besides water stress, changes in these non-stress factors can affect SIF observations, and effectively eliminating these interfering factors and accurately assessing drought severity remains a challenge. Second, existing methods mostly rely on single-phase SIF observations for stress discrimination, but single-phase observations cannot comprehensively reflect the physiological dynamics of crops within a diurnal cycle, resulting in insufficient reliability and stability of the diagnosis. Third, although some studies have used the ratio of SIF to photosynthetically active radiation (PAR) (i.e., apparent fluorescence yield SIFy) to eliminate the influence of incident radiation fluctuations, changes in canopy structure and biochemical components caused by crop growth have not been effectively eliminated. Furthermore, existing diagnostic methods based on the soil moisture deficit index rely on the installation of physical probes in the field, which are severely affected by soil spatial heterogeneity and cannot directly capture early signals of photosynthetic inhibition at the plant physiological level, resulting in high cost, significant disruption, insufficient representativeness, and delayed early warning response. Therefore, there is an urgent need for a crop water stress diagnosis method based on the fusion of multiple features of SIF diurnal variation curves to achieve accurate diagnosis of crop water stress. Summary of the Invention

[0004] The purpose of this invention is to provide a crop water stress diagnosis method based on the fusion of multiple features of SIF diurnal variation curves, aiming to solve one or more of the above-mentioned problems.

[0005] This invention provides a crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves, comprising: Acquire observation time series data of the target crop within a daily cycle. The observation time series data includes canopy sunlight-induced chlorophyll fluorescence value, canopy reflectance, incident photosynthetically active radiation, solar zenith angle, observation zenith angle, air temperature, air humidity, leaf area index, soil moisture, and canopy temperature. The observed time series data were preprocessed and normalized by illumination to obtain the apparent fluorescence yield. Historical observation data of the target crop during normal water supply period are obtained. Based on the historical observation data, a physiological baseline of the diurnal variation of SIFy under water stress is constructed. The apparent fluorescence yield is compared with the physiological baseline of the diurnal variation of SIFy under water stress to determine the physiological deviation of SIFy. The fluorescence response sensitivity is determined based on the apparent fluorescence yield, and the transition moment of the target crop from light-limited state to water-limited state is determined based on the fluorescence response sensitivity and the SIFy physiological deviation. Diurnal stress features are extracted based on the physiological deviation of SIFy and the turning point, and the diurnal stress features are fused to obtain a comprehensive water stress index. The water stress level of the target crop is determined based on the comprehensive water stress index.

[0006] Preferably, before performing preprocessing, the target crop is assessed for heat stress based on the canopy temperature. If no heat stress is present, preprocessing is performed. The preprocessing includes removing outliers and cloud shadow interference data, performing time resampling, and radiometric correction.

[0007] Preferably, the apparent fluorescence yield is determined according to the following formula: SIFy = SIF / PAR; Alternatively, SIFy = SIF / APAR; Wherein, SIFy represents the apparent fluorescence yield, SIF represents the canopy sunlight-induced chlorophyll fluorescence value, PAR represents the incident photosynthetically active radiation, and APAR represents the absorbed photosynthetically active radiation. The absorbed photosynthetically active radiation is the product of the incident photosynthetically active radiation and the photosynthetically active radiation absorption ratio, which is determined by an empirical model of canopy reflectivity and photosynthetically active radiation absorption ratio.

[0008] Preferably, historical observation data of the target crop during normal water supply periods are obtained, and a physiological baseline for the diurnal variation of SIFy under water stress is constructed based on the historical observation data, including: Obtain observation time series data of the target crop, compare the observation time series data with historical observation data, and if the data within the historical observation data collection period are the same as the growth stage, canopy structure and environmental conditions in the observation time series data, then set the historical observation data within that collection period as model construction data; Based on the model data, a physiological baseline for the diurnal variation of SIFy under water stress is constructed. The expression for the physiological baseline for the diurnal variation of SIFy under water stress is as follows: ; in, The saturation point represents the physiological baseline of the diurnal variation of SIFy without water stress, PAR represents incident photosynthetically active radiation, VPD represents saturated vapor pressure difference, θs represents solar zenith angle, θv represents observed zenith angle, LAI represents leaf area index, and P represents growth stage. The saturated water vapor pressure difference is determined based on air temperature and air humidity.

[0009] Preferably, the physiological deviation of SIFy is the relative deviation, ratio, standardized residual, or probability deviation between the apparent fluorescence yield and the physiological baseline of the diurnal variation of SIFy under water stress.

[0010] Preferably, when determining the fluorescence response sensitivity based on the apparent fluorescence yield, the fluorescence response sensitivity is determined according to the following formula: ; Where S represents the fluorescence response sensitivity, SIFy represents the apparent fluorescence yield, and PAR represents the incident photosynthetically active radiation.

[0011] Preferably, determining the transition moment from light-limited to water-limited state in the target crop based on the fluorescence response sensitivity and SIFy physiological deviation includes: Pre-set the deviation threshold and the photoresponse sensitivity threshold; The fluorescence photoresponse sensitivity is compared with the photoresponse sensitivity threshold, and the SIFy physiological deviation is compared with the deviation threshold. If the fluorescence response sensitivity is less than the photoresponse sensitivity threshold and the SIFy physiological deviation is greater than the deviation threshold, and the duration exceeds a preset time, then the moment when the fluorescence response sensitivity is less than the photoresponse sensitivity threshold and the SIFy physiological deviation is greater than the deviation threshold is determined as the turning point when the target crop transitions from a light-limited state to a water-limited state.

[0012] Preferably, the diurnal variation stress characteristics include: transition advance, maximum physiological deviation, cumulative deviation area of ​​water limitation, afternoon recovery rate, apparent fluorescence yield - hysteresis loop area of ​​incident photosynthetically active radiation, and morning-noon-afternoon asymmetry index.

[0013] Preferably, the comprehensive water stress index is determined according to the following formula: ; Wherein, WSI represents the comprehensive water stress index, Z(Fi) represents the i-th standardized diurnal variation stress feature, and wi represents the weight of the i-th diurnal variation stress feature.

[0014] Preferably, determining the water stress level of the target crop based on the comprehensive water stress index includes: A first water stress index, a second water stress index, and a third water stress index are preset, and the first water stress index, the second water stress index, and the third water stress index increase sequentially; The water stress level of the target crop is determined based on the relationship between the comprehensive water stress index and the first water stress index, the second water stress index, and the third water stress index. If the comprehensive water stress index is less than the first water stress index, the target crop is set to the normal level; If the comprehensive water stress index is greater than or equal to the first water stress index, and the comprehensive water stress index is less than the second water stress index, then the target crop is set to a mild water stress level. If the comprehensive water stress index is greater than or equal to the second water stress index, and the comprehensive water stress index is less than the third water stress index, then the target crop is set to a moderate water stress level. If the comprehensive water stress index is greater than or equal to the third water stress index, then the target crop is set to a severe water stress level.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention first preprocesses and normalizes the observation time series data to obtain the apparent fluorescence yield, thereby eliminating the influence of the diurnal variation of incident radiation intensity on the SIF observation value; then, using historical data of the target crop during normal water supply, a physiological baseline of the diurnal variation of SIFy under water stress is constructed, which integrates multiple environmental physiological parameters; the actual observed apparent fluorescence yield is compared with the baseline to obtain the physiological deviation of SIFy, thereby eliminating interference from non-stress factors such as light and weather, and accurately extracting water stress-specific signals.

[0016] Based on the apparent fluorescence yield, the fluorescence response sensitivity is obtained. Combined with the physiological deviation of SIFy, the turning point when the crop changes from light-limited to water-limited state is determined: when the fluorescence response sensitivity is less than the preset threshold and the physiological deviation of SIFy is greater than the preset threshold and continues to meet the time requirement, the moment is determined to be the turning point. This can accurately capture the key time of crop limitation state change within the daily cycle and support early warning and time-series diagnosis of water stress.

[0017] Based on the physiological deviation and turning point of SIFy, this study extracts various diurnal stress characteristics, such as the advance of the turning point, the maximum physiological deviation, the cumulative deviation area of ​​water limitation, the afternoon recovery rate, the area of ​​the hysteresis loop between apparent fluorescence yield and incident photosynthetically active radiation, and the asymmetry between the morning, afternoon, and afternoon. The stress impact is characterized from six dimensions: the degree of advance of state transition, the peak intensity of stress, the cumulative effect of the diurnal cycle, the physiological recovery capacity, the degree of response lag, and the difference in response between the morning and afternoon. This overcomes the deficiency of insufficient diagnostic information from a single indicator and can comprehensively reflect the impact of water stress on the diurnal process of crop photosynthetic physiology.

[0018] After standardizing and weighting the extracted multidimensional diurnal variation stress features, a comprehensive water stress index is obtained. This index retains the sensitivity of each dimension to different stresses and improves the robustness of the diagnostic results. By comparing the index with preset grading thresholds, water stress is divided into four levels: normal, mild, moderate, and severe, thus achieving an objective quantitative grading diagnosis of the degree of water stress. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the crop water stress diagnosis method based on the fusion of multiple features from the SIF diurnal variation curve, as described in this invention. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0022] like Figure 1As shown, this invention provides a crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves, including: Acquire observation time series data of the target crop within a daily cycle. The observation time series data includes canopy sunlight-induced chlorophyll fluorescence value, canopy reflectance, incident photosynthetically active radiation, solar zenith angle, observation zenith angle, air temperature, air humidity, leaf area index, soil moisture, and canopy temperature. The observed time series data were preprocessed and normalized by illumination to obtain the apparent fluorescence yield. Historical observation data of the target crop during normal water supply period are obtained. Based on the historical observation data, a physiological baseline of the diurnal variation of SIFy under water stress is constructed. The apparent fluorescence yield is compared with the physiological baseline of the diurnal variation of SIFy under water stress to determine the physiological deviation of SIFy. The fluorescence response sensitivity is determined based on the apparent fluorescence yield, and the transition moment of the target crop from light-limited state to water-limited state is determined based on the fluorescence response sensitivity and the SIFy physiological deviation. Diurnal stress features are extracted based on the physiological deviation of SIFy and the turning point, and the diurnal stress features are fused to obtain a comprehensive water stress index. The water stress level of the target crop is determined based on the comprehensive water stress index.

[0023] This invention fully utilizes the crop physiological response information contained in the diurnal variation curve of SIFy. By integrating two types of information—the degree of physiological deviation and the characteristics of stress inflection—it overcomes the uncertainty limitation of traditional methods that rely solely on single-point observation of SIF values ​​to diagnose stress. This invention can more accurately and stably identify the occurrence and severity of crop water stress, providing reliable technical support for precision crop water management in the field.

[0024] In some embodiments of this application, before performing preprocessing operations, the canopy temperature is used to determine whether the target crop is under heat stress. If no heat stress is present, preprocessing operations are performed. The preprocessing operations include: removing outliers and cloud shadow interference data, performing time resampling and radiation correction.

[0025] Understandably, by performing quality screening and standardization on the original observation time series data, interference from factors such as instrument measurement errors, sudden environmental disturbances (such as cloud cover), and observational geometric inconsistencies on the original SIF signal can be eliminated. This ensures that the SIF data used in subsequent modeling and feature extraction can truly reflect the physiological state of the crop canopy, thereby guaranteeing the reliability and accuracy of water stress diagnosis results.

[0026] Specifically, outliers refer to observational data that significantly deviate from the normal fluctuation range due to sensor malfunction, signal transmission interruption, or sudden external interference (such as bird strikes or mechanical vibration). Outlier removal employs a statistical discrimination method: for each observation moment in the continuous time series, the local mean and standard deviation of the SIF observation value and its associated parameters (such as incident photosynthetically active radiation, temperature, etc.) at that moment are calculated. If an observation value exceeds the mean ± k times the standard deviation (k can be set according to the actual data quality, usually 3), the data at that moment is marked as an outlier and removed. Furthermore, values ​​that significantly exceed the physically reasonable range (such as negative values ​​or values ​​exceeding the sensor's measurement range) are also removed.

[0027] Cloud shadow interference rapidly reduces the incident photosynthetically active radiation reaching the canopy, causing a sharp drop in SIF (Solar Zenith Injection) values. Retaining such data would severely distort the crop's physiological response signals. Cloud shadow interference is identified based on the following principle: when incident photosynthetically active radiation exhibits drastic, abnormal, and rapid fluctuations within adjacent time windows (e.g., a drop in PAR value exceeding 40% within minutes followed by a rapid recovery), accompanied by a simultaneous decrease in SIF value, cloud shadow interference is determined to exist during that period, and all observational data within that period are discarded. Preferably, this can be aided by combining sky imager data or utilizing the theoretical relationship between solar zenith angle and PAR.

[0028] Because the data acquisition frequency of the original observation equipment may be inconsistent or the sampling intervals may be non-equidistant (for example, some instruments record every 5 minutes, while others record every 15 minutes), it is not convenient for subsequent unified analysis and feature extraction of diurnal variation curves. The purpose of time resampling is to uniformly resample the non-equidistant observation data after removing outliers and cloud shadow interference into a time series with equal time intervals through interpolation methods. Specifically, linear interpolation or cubic spline interpolation methods are used to resample the data to a preset fixed time step (preferably 5 minutes or 10 minutes) to ensure that the observation times within each diurnal cycle correspond one-to-one. For data segments with consecutive missing durations exceeding 30 minutes, no interpolation imputation is performed to avoid introducing excessive errors.

[0029] The purpose of radiometric correction is to convert the digital signals recorded by the instrument into physically meaningful radiance or irradiance values, thereby eliminating the influence of factors such as sensor response differences, dark current drift, and optical system attenuation on the measured values. The specific operation of radiometric correction is as follows: using the radiometric calibration coefficients specified by the observation equipment at the factory, the original DN values ​​are converted into canopy SIF radiance and incident photosynthetically effective radiative flux density; simultaneously, considering changes in the solar zenith angle, the SIF radiance is corrected to the equivalent value under standard observation geometry conditions to eliminate geometric differences introduced by changes in observation angle. Furthermore, for long-term series data, calibration coefficient drift caused by instrument attenuation must be considered, and radiometric calibration checks and corrections should be performed periodically using a standard light source.

[0030] Specifically, after removing outliers and cloud shadow interference data, to ensure the continuity and integrity of the SIF diurnal variation time series, this application preferably performs corresponding repair or removal processing on missing data. Specifically, for short-term data gaps with a continuous missing duration not exceeding a preset threshold (preferably 30 minutes), cubic spline interpolation is used for filling. In order to ensure the physical consistency of the interpolated data, related parameters such as SIFy, PAR, and temperature are co-interpolated within the same missing period to ensure the rationality of the time difference ratio for subsequent calculation of fluorescence response sensitivity. For long-term data gaps with a continuous missing duration exceeding the preset threshold (preferably 30 minutes) or a total missing data volume exceeding 20% ​​to 30% of the total observed data volume on that day, it is determined that the data on that day can no longer effectively restore the true diurnal variation pattern. No further interpolation is performed, and all observed data on that day are directly marked as invalid and removed, and do not enter the subsequent physiological baseline construction and stress feature extraction process. By employing a hierarchical processing strategy based on a missing duration threshold, the integrity of data for effective observation days is ensured to meet the requirements of continuous time series for diurnal variation feature extraction, while avoiding the problem of distorted diagnostic results caused by forced interpolation when information is severely insufficient.

[0031] In some embodiments of this application, the apparent fluorescence yield is determined according to the following formula: SIFy = SIF / PAR; or, SIFy = SIF / APAR; Wherein, SIFy represents the apparent fluorescence yield, SIF represents the canopy sunlight-induced chlorophyll fluorescence value, PAR represents the incident photosynthetically active radiation, and APAR represents the absorbed photosynthetically active radiation. The absorbed photosynthetically active radiation is the product of the incident photosynthetically active radiation and the photosynthetically active radiation absorption ratio, which is determined by an empirical model of canopy reflectivity and photosynthetically active radiation absorption ratio.

[0032] Specifically, the canopy sunlight-induced chlorophyll fluorescence value can be obtained by using far-red SIF around 760nm, or by using red SIF around 687nm and far-red SIF around 760nm simultaneously.

[0033] Understandably, by normalizing the original SIF (Self-Effective Fluorescence) values ​​of the canopy, the direct impact of fluctuations in incident photosynthetically active radiation intensity within the diurnal cycle on the SIF values ​​can be eliminated. This allows the normalized apparent fluorescence yield to more accurately reflect the fluorescence emission efficiency of the crop canopy under unit light conditions, thereby removing the interference of light intensity, a non-physiological factor, on the SIF signal. This provides a standardized indicator that is decoupled from light conditions and reflects the intrinsic physiological state of the crop for subsequent water stress diagnosis based on diurnal SIF variation characteristics.

[0034] Specifically, since there is a high positive correlation between the intensity of sunlight-induced chlorophyll fluorescence signal and incident photosynthetically active radiation (SIF), even if the crop's physiological state remains completely unchanged, the natural fluctuations in SIF caused by diurnal variations in incident radiation can still mask or interfere with the true physiological signals induced by water stress. Therefore, this application converts the raw SIF observations into apparent fluorescence yields decoupled from incident radiation by performing light normalization.

[0035] The normalization of incident photosynthetically active radiation (SEPA) is given by SIFy = SIF / PAR; where SIFy represents the apparent fluorescence yield, which is physically the SIF intensity emitted by the crop canopy per unit of incident SEPA received; SIF represents the pre-treated canopy light-induced chlorophyll fluorescence value; and PAR represents the incident SEPA. This scheme is simple to implement, directly using synchronously observed incident SEPA to normalize the ratio of SIF, and is suitable for application scenarios with relatively homogeneous canopy structure and small variations in leaf area index.

[0036] The normalized value of absorbed photosynthetically active radiation is SIFy = SIF / APAR. This scheme further considers the actual absorption capacity of the canopy for incident photosynthetically active radiation, and can more accurately reflect the fluorescence emission efficiency of crop photosynthetic apparatus based on effective radiation absorption. Here, APAR represents the absorbed photosynthetically active radiation, which physically refers to the portion of photosynthetically active radiation actually absorbed by the crop canopy and available for photosynthesis.

[0037] APAR is determined by the following formula: APAR = PAR × FAPAR; where FAPAR represents the proportion of photosynthetically active radiation absorbed, that is, the proportion of photosynthetically active radiation absorbed by the canopy to the incident photosynthetically active radiation, and its value ranges from 0 to 1. FAPAR is determined through an empirical model of canopy reflectivity and FAPAR.

[0038] Specifically, the Normalized Difference Vegetation Index (NDVI) is first calculated based on the preprocessed red and near-infrared reflectance, expressed as: NDVI = (NIR - Red) / (NIR + Red); where NIR represents the reflectance in the near-infrared band and Red represents the reflectance in the red band. Then, the NDVI is converted to FAPAR using a pre-established NDVI-FAPAR empirical model. This empirical model preferably takes the following form: FAPAR = a × NDVI + b or FAPAR = a × NDVI 2 +b×NDVI+c; where a, b, and c are empirical fitting coefficients, obtained by regression analysis of paired NDVI data collected at multiple growth stages for the target crop type and measured FAPAR data. Since the NDVI-FAPAR relationship may differ between different crop types and growth stages, preferably, the above empirical coefficients are dynamically selected and adjusted according to the type of target crop and the current growth stage.

[0039] In some embodiments of this application, historical observation data of the target crop during normal water supply periods are obtained, and a physiological baseline for the diurnal variation of SIFy under water stress is constructed based on the historical observation data. This includes: obtaining observation time series data of the target crop, comparing the observation time series data with historical observation data, and if the data within the historical observation data collection period are the same as the growth stage, canopy structure and environmental conditions in the observation time series data, then the historical observation data within that collection period is set as model construction data. Based on the model data, a physiological baseline for the diurnal variation of SIFy under water stress is constructed. The expression for the physiological baseline for the diurnal variation of SIFy under water stress is as follows: ; in, The saturation point represents the physiological baseline of the diurnal variation of SIFy without water stress, PAR represents incident photosynthetically active radiation, VPD represents saturated vapor pressure difference, θs represents solar zenith angle, θv represents observed zenith angle, LAI represents leaf area index, and P represents growth stage. The saturated water vapor pressure difference is determined based on air temperature and air humidity.

[0040] Specifically, the baseline is obtained through any one or more combinations of the following methods: historical observation data of the same plot on normal water supply dates, synchronous observation data of the normal water supply control area, simulation data based on the radiative transfer model, regression model or machine learning model trained on normal samples, and self-calibration of the baseline based on observation data of the low stress period in the early morning of the day.

[0041] The baseline is self-calibrated based on the observation data during the low-stress period in the early morning of the day, including: using the measured SIFy data within a preset time window after sunrise as the no-stress reference value for the day, and shifting and / or scaling the initial baseline obtained from historical data, control area observations or model simulations to eliminate baseline differences between different plots and fertility periods.

[0042] Understandably, historical observation data of the target crop under normal water supply (i.e., no water stress) conditions are used to construct a reference baseline that reflects the diurnal variation of SIFy under healthy physiological conditions. This baseline is not a simple diurnal mean curve, but rather establishes a functional relationship between apparent fluorescence yield and multiple factors such as photosynthetically active radiation, saturated vapor pressure difference, solar zenith angle, observed zenith angle, leaf area index, and growth stage. Through this baseline, the actual observed apparent fluorescence yield at any given time can be compared with the "no-stress baseline value that should be achieved under that time, environmental conditions, and canopy structure," thereby eliminating the superimposed influence of non-stress factors such as light, weather, canopy structure, and growth process on the diurnal variation of SIFy, allowing the subsequently extracted physiological deviations to be specifically attributed to water stress.

[0043] Specifically, the normal water supply period is a period in which the soil moisture supply is sufficient, allowing the crop to be in a state of complete water stress-free condition, provided that the crop is healthy and disease-free, has sufficient nutrients, and the environmental conditions such as temperature are suitable and there are no other abiotic stresses.

[0044] In some embodiments of this application, the physiological deviation of SIFy is the relative deviation, ratio, standardized residual, or probability deviation of the apparent fluorescence yield from the physiological baseline of the diurnal variation of SIFy under water stress.

[0045] Understandably, quantitatively comparing the measured apparent fluorescence yield with the physiological baseline of diurnal variation of SIFy without water stress, and constructing different types of deviation indicators, transforms the difference between the two into quantifiable values, thereby achieving a quantitative characterization of the presence and relative degree of water stress. Furthermore, by providing multiple deviation calculation schemes, this method can adapt to the differentiated needs of various application scenarios regarding data accuracy, computational complexity, and statistical interpretability, enhancing the method's flexibility and universality.

[0046] In some embodiments of this application, when determining the fluorescence response sensitivity based on the apparent fluorescence yield, the fluorescence response sensitivity is determined according to the following formula: ; Where S represents the fluorescence response sensitivity, SIFy represents the apparent fluorescence yield, and PAR represents the incident photosynthetically active radiation.

[0047] Understandably, by calculating the first derivative of apparent fluorescence yield with respect to incident photosynthetically active radiation, the instantaneous response rate of crop canopy fluorescence emission efficiency to changes in light intensity can be quantitatively characterized, i.e., the physiological sensitivity of crop photosynthetic apparatus to a unit increase in light intensity. This fluorescence response sensitivity can effectively distinguish between two different physiological limitation states. When the crop is in a light-limited state, SIFy increases significantly with increasing PAR, and the sensitivity is a relatively high positive value. When the crop is subjected to water stress and transitions to a water-limited state, the photosynthetic apparatus is damaged, and SIFy tends to saturate or even decrease with increasing PAR, with the sensitivity decreasing or even approaching zero or becoming negative. Through this index, the critical moment of the transition from a light-limited state to a water-limited state can be accurately captured during diurnal variations, providing an objective and quantitative basis for determining the subsequent turning point.

[0048] In some embodiments of this application, determining the transition moment of a target crop from a light-limited state to a water-limited state based on the fluorescence photoresponse sensitivity and the SIFy physiological deviation includes: pre-setting a deviation threshold and a photoresponse sensitivity threshold; comparing the fluorescence photoresponse sensitivity with the photoresponse sensitivity threshold, and comparing the SIFy physiological deviation with the deviation threshold; if the fluorescence photoresponse sensitivity is less than the photoresponse sensitivity threshold and the SIFy physiological deviation is greater than the deviation threshold, and the duration exceeds a preset time, then the moment corresponding to the fluorescence photoresponse sensitivity being less than the photoresponse sensitivity threshold and the SIFy physiological deviation being greater than the deviation threshold is determined as the transition moment of the target crop from a light-limited state to a water-limited state.

[0049] Understandably, by utilizing two physically complementary indicators—fluorescence photoresponse sensitivity and SIFy physiological deviation—a dual-threshold discrimination mechanism can accurately identify the critical moment when crops transition from light-limited to water-limited states in diurnal time series. Specifically, in the early morning when light intensity is low, crop photosynthesis is primarily limited by insufficient light. At this time, fluorescence photoresponse sensitivity is high (SIFy increases rapidly with PAR), and SIFy physiological deviation is low (measured values ​​are close to the baseline without stress). As daytime light intensity increases and the effects of water stress accumulate, crops gradually transition to a water-limited state. At this point, fluorescence photoresponse sensitivity drops below the threshold (SIFy becomes insensitive to PAR), and SIFy physiological deviation increases above the threshold. By simultaneously determining the achievement of both indicators and combining this with duration constraints, the precise time point of this state transition can be reliably captured. This provides crucial temporal segmentation for the subsequent extraction of diurnal stress characteristics and also offers a physiologically meaningful temporal reference indicator for early warning of water stress.

[0050] In some embodiments of this application, the diurnal variation stress characteristics include: transition advance, maximum physiological deviation, cumulative deviation area of ​​water limitation, afternoon recovery rate, apparent fluorescence yield-incident photosynthetically active radiation hysteresis loop area, and morning-noon-afternoon asymmetry index.

[0051] Specifically, the turning point advance Δt is calculated as follows: Δt = tc - tc,0, where tc is the turning point time, and tc,0 is the turning point time under normal conditions without moisture stress. The smaller Δt is, the earlier the stress occurs.

[0052] The maximum physiological deviation amplitude Dmax is calculated as follows: Dmax = max|D(t)|, t∈[tc,tend], where tend is the end time of the daily cycle observation, and Dmax represents the maximum deviation of the measured SIFy from the baseline after entering the water limitation stage.

[0053] The cumulative deviation area A of water limitation is calculated as follows: A = ∫[B(t) - SIFy(t)] dt, t∈[tc,tend], where B(t) is the physiological baseline of the diurnal variation of SIFy without water stress, SIFy(t) is the measured apparent fluorescence yield, and A represents the cumulative area between the deviation curve and the baseline after entering the water limitation stage.

[0054] The afternoon recovery rate Rrec is calculated as follows: Rrec = d[SIFy(t)] / dt, t∈[tnoon,tend]; where tnoon is the noon time, and tend is the end time of the daily observation cycle. Rrec represents the rate of change of SIFy over time in the afternoon period, which is used to distinguish between temporary high light suppression and persistent moisture stress.

[0055] The SIFy-PAR hysteresis loop area H is calculated as follows: H = ∮[SIFy(PAR)]d(PAR); that is, the SIFy-PAR response curves in the morning and afternoon are combined into a closed or quasi-closed curve, and the hysteresis loop area is calculated to quantify the "memory effect" and stress level of crop light energy utilization.

[0056] The morning-noon-afternoon asymmetry indicators include: comparing the difference between the rate of increase of SIFy in the morning and the rate of decrease of SIFy in the noon, as well as the rate of recovery of SIFy in the afternoon, to distinguish between temporary high light suppression and persistent moisture stress.

[0057] In some embodiments of this application, the comprehensive water stress index is determined according to the following formula: ; Wherein, WSI represents the comprehensive water stress index, Z(Fi) represents the i-th standardized diurnal variation stress feature, and wi represents the weight of the i-th diurnal variation stress feature.

[0058] Understandably, the aforementioned extracted diurnal variation stress features, including the turning point advance, maximum physiological deviation, cumulative deviation area of ​​water limitation, afternoon recovery rate, SIFy-PAR hysteresis loop area, and morning-noon-afternoon asymmetry index, are integrated into a single comprehensive water stress index through standardization and weighted fusion. This fusion process has the following three benefits: First, each stress feature characterizes the impact of water stress on the diurnal variation curve of SIFy from different dimensions. A single feature may be sensitive to a certain type or degree of stress but insensitive to others. Weighted fusion of multiple features can leverage their respective advantages, compensate for the limitations of a single feature, and improve the overall robustness of diagnosis. Second, standardization eliminates the incomparability caused by differences in units and numerical ranges between features, making different features comparable and additive on a unified scale. Third, weighted allocation allows for differentiated processing of the importance of different features, flexibly adjusting the contribution of each feature in stress discrimination according to crop type, growth stage, or application scenario. The resulting comprehensive water stress index, as a dimensionless quantitative indicator, can intuitively reflect the overall water stress level of crops, providing a unified and comparable numerical basis for the subsequent objective classification of water stress levels.

[0059] Specifically, in this embodiment, the corresponding feature weights are usually reasonably obtained through the following technical path: Sample calibration and statistical fitting: Collect a large number of field measured datasets covering different crop types, different growth stages, and different soil moisture gradients. Use the calibration value of the actual water stress degree of the crop as the baseline dependent variable and the corresponding standardized diurnal variation stress features as independent variables. Calculate the regression coefficients corresponding to each feature through statistical fitting methods such as multiple linear regression and partial least squares regression. After normalization, the weight values ​​that fit the sample distribution can be obtained. This method can intuitively reflect the contribution of each diurnal variation stress feature to the comprehensive water stress index.

[0060] Machine learning training and calibration acquisition: A multi-temporal crop stress remote sensing / observation dataset with labeled water stress levels is used. Standardized diurnal variation stress features are used as input features, and the ground truth value of a pre-defined comprehensive water stress index is used as the label. The model is trained using supervised models with feature importance assessment capabilities, such as Random Forest and XGBoost. The feature importance results output by the model are normalized to obtain the corresponding weights. This method can automatically mine the differences in the nonlinear contributions of different stress characteristics to the overall stress situation.

[0061] Domain experience combined with weighted analysis of analytic hierarchy process: Professionals in the field of crop water stress research combine their experience with the sensitivity of crops to water deficit response to different diurnal stress characteristics (such as diurnal temperature range, diurnal fluctuation range of vegetation index, etc.) to pairwise comparisons of the stress indication ability of different characteristics, construct judgment matrices, and calculate the ranking weight of each characteristic after consistency verification. This method can ensure the agronomic rationality of the weights in scenarios with insufficient sample data.

[0062] Combined optimization weighting acquisition: The above statistical objective weighting results are coupled with the domain experience subjective weighting results. By minimizing the optimization objective of the sum of squared residuals between the comprehensive water stress index and the measured stress true value, the final weights that take into account both data-driven and agronomic professionalism are obtained, further improving the stress characterization accuracy of the comprehensive index.

[0063] In some embodiments of this application, determining the water stress level of a target crop based on the comprehensive water stress index includes: pre-setting a first water stress index, a second water stress index, and a third water stress index, wherein the first water stress index, the second water stress index, and the third water stress index increase sequentially; determining the water stress level of the target crop based on the relationship between the comprehensive water stress index and the first water stress index, the second water stress index, and the third water stress index; if the comprehensive water stress index is less than the first water stress index, the target crop is set to a normal level; if the comprehensive water stress index is greater than or equal to the first water stress index and the comprehensive water stress index is less than the second water stress index, the target crop is set to a mild water stress level; if the comprehensive water stress index is greater than or equal to the second water stress index and the comprehensive water stress index is less than the third water stress index, the target crop is set to a moderate water stress level; if the comprehensive water stress index is greater than or equal to the third water stress index, the target crop is set to a severe water stress level.

[0064] Understandably, transforming the continuous Comprehensive Water Stress Index (WSI) obtained through multi-feature weighted fusion into a discrete water stress level with clear agricultural semantics allows users (such as farmers, agricultural managers, or irrigation decision-making systems) to intuitively understand the severity of current crop water stress and make corresponding irrigation management decisions. By setting three progressively increasing thresholds (first, second, and third water stress indices), the continuous WSI value range is divided into four corresponding level ranges: normal, mild water stress, moderate water stress, and severe water stress, thus constructing a clear and well-defined diagnostic standard for crop water stress. This grading system balances the precision of diagnosis with the operability of practical application. It avoids the limitation of relying solely on a single value to directly guide production, and provides richer intermediate state information than simple binary classification (stressed / non-stressed) through multi-level division, providing a quantitative basis for the formulation of differentiated regulation strategies in precision irrigation.

[0065] Specifically, it also includes outputting diagnostic results, which include crop water stress level, stress onset time, stress duration, water stress spatial distribution map, diagnostic confidence level, and optional irrigation trigger signal.

[0066] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0067] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0068] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0069] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0070] 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 it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves, characterized in that, include: Acquire observation time series data of the target crop within a daily cycle. The observation time series data includes canopy sunlight-induced chlorophyll fluorescence value, canopy reflectance, incident photosynthetically active radiation, solar zenith angle, observation zenith angle, air temperature, air humidity, leaf area index, soil moisture, and canopy temperature. The observed time series data were preprocessed and normalized by illumination to obtain the apparent fluorescence yield. Historical observation data of the target crop during normal water supply period are obtained. Based on the historical observation data, a physiological baseline of the diurnal variation of SIFy under water stress is constructed. The apparent fluorescence yield is compared with the physiological baseline of the diurnal variation of SIFy under water stress to determine the physiological deviation of SIFy. The fluorescence response sensitivity is determined based on the apparent fluorescence yield, and the transition moment of the target crop from light-limited state to water-limited state is determined based on the fluorescence response sensitivity and the SIFy physiological deviation. Diurnal stress features are extracted based on the physiological deviation of SIFy and the turning point, and the diurnal stress features are fused to obtain a comprehensive water stress index. The water stress level of the target crop is determined based on the comprehensive water stress index.

2. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, Before performing the pretreatment operation, the canopy temperature is used to determine whether the target crop is under heat stress. If there is no heat stress, the pretreatment operation is performed. The preprocessing operations are as follows: removing outliers and cloud shadow interference data, performing time resampling and radiometric correction.

3. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, The apparent fluorescence yield is determined according to the following formula: SIFy = SIF / PAR; Alternatively, SIFy = SIF / APAR; Wherein, SIFy represents the apparent fluorescence yield, SIF represents the canopy sunlight-induced chlorophyll fluorescence value, PAR represents the incident photosynthetically active radiation, and APAR represents the absorbed photosynthetically active radiation. The absorbed photosynthetically active radiation is the product of the incident photosynthetically active radiation and the photosynthetically active radiation absorption ratio, which is determined by an empirical model of canopy reflectivity and photosynthetically active radiation absorption ratio.

4. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, Historical observation data of the target crop during normal water supply periods were obtained, and a physiological baseline for the diurnal variation of SIFy under water stress was constructed based on the historical observation data, including: Obtain observation time series data of the target crop, compare the observation time series data with historical observation data, and if the data within the historical observation data collection period are the same as the growth stage, canopy structure and environmental conditions in the observation time series data, then set the historical observation data within that collection period as model construction data; Based on the model data, a physiological baseline for the diurnal variation of SIFy under water stress is constructed. The expression for the physiological baseline for the diurnal variation of SIFy under water stress is as follows: ； in, The saturation point represents the physiological baseline of the diurnal variation of SIFy without water stress, PAR represents incident photosynthetically active radiation, VPD represents saturated vapor pressure difference, θs represents solar zenith angle, θv represents observed zenith angle, LAI represents leaf area index, and P represents growth stage. The saturated water vapor pressure difference is determined based on air temperature and air humidity.

5. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, The physiological deviation of SIFy is the relative deviation, ratio, standardized residual, or probability deviation between the apparent fluorescence yield and the physiological baseline of the diurnal variation of SIFy under water stress.

6. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, When determining the fluorescence response sensitivity based on the apparent fluorescence yield, the fluorescence response sensitivity is determined according to the following formula: ； Where S represents the fluorescence response sensitivity, SIFy represents the apparent fluorescence yield, and PAR represents the incident photosynthetically active radiation.

7. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, The transition moment from light-limited to water-limited state in the target crop is determined based on the fluorescence response sensitivity and SIFy physiological deviation, including: Pre-set the deviation threshold and the photoresponse sensitivity threshold; The fluorescence photoresponse sensitivity is compared with the photoresponse sensitivity threshold, and the SIFy physiological deviation is compared with the deviation threshold. If the fluorescence response sensitivity is less than the photoresponse sensitivity threshold and the SIFy physiological deviation is greater than the deviation threshold, and the duration exceeds a preset time, then the moment when the fluorescence response sensitivity is less than the photoresponse sensitivity threshold and the SIFy physiological deviation is greater than the deviation threshold is determined as the turning point when the target crop transitions from a light-limited state to a water-limited state.

8. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, The diurnal variation stress characteristics include: transition advance, maximum physiological deviation, cumulative deviation area of ​​water limitation, afternoon recovery rate, apparent fluorescence yield - hysteresis loop area of ​​incident photosynthetically active radiation, and morning-noon-afternoon asymmetry index.

9. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, The comprehensive water stress index is determined according to the following formula: ； Wherein, WSI represents the comprehensive water stress index, Z(Fi) represents the i-th standardized diurnal variation stress feature, and wi represents the weight of the i-th diurnal variation stress feature.

10. The crop water stress diagnosis method based on multi-feature fusion of SIF diurnal variation curves according to claim 1, characterized in that, The water stress level of the target crop is determined based on the comprehensive water stress index, including: A first water stress index, a second water stress index, and a third water stress index are preset, and the first water stress index, the second water stress index, and the third water stress index increase sequentially; The water stress level of the target crop is determined based on the relationship between the comprehensive water stress index and the first water stress index, the second water stress index, and the third water stress index. If the comprehensive water stress index is less than the first water stress index, the target crop is set to the normal level; If the comprehensive water stress index is greater than or equal to the first water stress index, and the comprehensive water stress index is less than the second water stress index, then the target crop is set to a mild water stress level. If the comprehensive water stress index is greater than or equal to the second water stress index, and the comprehensive water stress index is less than the third water stress index, then the target crop is set to a moderate water stress level. If the comprehensive water stress index is greater than or equal to the third water stress index, then the target crop is set to a severe water stress level.

Images