Environment-triggered multi-temporal sif cooperative observation method and system based on unmanned aerial vehicle and crop water stress identification method and system

Abstract

The present application relates to the technical field of SIF remote sensing detection, and discloses a method and system for unmanned aerial vehicle multi-temporal SIF cooperative observation and crop water stress identification based on environmental triggering, which first collects farmland environment and crop physiological parameters, combines crop categories and growth stage division to divide multi-class time sequence observation windows, and builds environmental triggering rules; after the rules take effect, the unmanned aerial vehicle is controlled to complete multi-period hyperspectral data collection with unified parameters; the original spectral data is corrected and processed, canopy SIF and apparent fluorescence yield are retrieved, a data set is constructed through atmospheric correction and space-time registration; interference signals are removed, different drought types are distinguished, crop water stress characteristics are analyzed by fusing multiple parameters, stress-related quantitative data are output, and variable irrigation equipment is linked. The present application can accurately distinguish soil drought and atmospheric drought, significantly improve the accuracy and reliability of crop water stress identification, accurately guide variable irrigation operation, and realize fine management and control of farmland water saving and efficiency improvement.

Description

Technical Field

[0001] This invention relates to the field of SIF remote sensing technology, and more specifically, to a method and system for multi-temporal SIF collaborative observation and crop water stress identification based on environment-triggered UAVs. Background Technology

[0002] Sunlight-induced chlorophyll fluorescence (SIF) can reflect the photosynthetic physiological state of crops and has been applied to remote sensing monitoring of water stress in farmland crops using unmanned aerial vehicle platforms. It has the advantages of high spatial resolution and early response to stress and has been widely studied and applied in the field of precision agriculture.

[0003] Current UAV-based SIF crop water stress identification mostly employs a single-temporal aerial survey acquisition mode triggered at fixed times. However, traditional single-temporal SIF monitoring methods rely solely on single observation data for stress discrimination, making it difficult to distinguish fluorescence signal changes caused by high light suppression, atmospheric drought, and soil drought, resulting in a high false alarm rate for stress identification. Furthermore, the lack of environmental-crop physiological constraints at observation times leads to highly subjective selection of observation windows, inconsistent flight conditions during multi-temporal acquisition, and poor comparability of spectral data from different time phases, directly interfering with the reliability of water stress diagnosis results. It is also difficult to eliminate fluorescence interference signals caused by non-water stress, making it impossible to effectively distinguish between soil drought and atmospheric drought. Additionally, the lack of a triggering mechanism linking environmental conditions and crop growth status during observation, coupled with arbitrary observation time selection and inconsistent parameters across multiple flights, results in inconsistent benchmarks for multi-temporal SIF data.

[0004] Therefore, it is necessary to design a method and system for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by UAVs, in order to solve the problems of insufficient identification accuracy caused by the difficulty in distinguishing high light suppression, atmospheric drought and soil drought stress signals in traditional UAV SIF monitoring, as well as the lack of environmental-physiological collaborative triggering mechanism and poor data comparability caused by inconsistent multi-temporal observation conditions. Summary of the Invention

[0005] In view of this, the present invention proposes a method and system for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by UAVs. It aims to solve the problems of insufficient identification accuracy caused by the difficulty in distinguishing high light suppression, atmospheric drought and soil drought stress signals in traditional UAV SIF monitoring, as well as the lack of environmental-physiological collaborative triggering mechanism and poor data comparability caused by inconsistent multi-temporal observation conditions.

[0006] In one aspect, this invention proposes a method for environmentally triggered UAV multi-temporal SIF collaborative observation and crop water stress identification, comprising: Collect environmental parameters of the target farmland and physiological parameters of crop stomatal conductance and leaf water potential. Based on the physiological parameters and crop type and growth stage parameters, determine the sensitive period of the crop, select the corresponding observation window set, and based on the selected observation window set, divide multiple time-series observation windows into multiple categories and construct environmental trigger judgment rules. When the environmental triggering judgment rule takes effect, the UAV carrying the hyperspectral calibration payload is controlled to collect multi-temporal spectral data of the target area in the corresponding time-series observation window with unified flight trajectory and observation parameters; After correcting and processing the raw spectral data collected by the UAV, the canopy SIF and apparent fluorescence yield are extracted based on the characteristic bands and oxygen absorption valleys. Atmospheric radiation correction is performed in combination with the UAV's low-altitude atmospheric parameters. Simultaneously, the spatiotemporal scale registration of multi-source data is completed to construct an observation dataset. Based on the observation dataset, various non-water stress interference signals are removed, the temporal characteristics of soil drought and atmospheric drought are distinguished, and crop physiological and soil and water meteorological parameters are integrated to obtain water stress characterization data. The stress level, spatial distribution, diagnostic confidence and irrigation priority data are output, and data linkage with variable irrigation equipment is carried out.

[0007] Furthermore, the process of dividing the time series observation windows into multiple categories includes: The environmental parameters, including photosynthetically active radiation, air temperature, relative humidity, water vapor pressure deficit, solar zenith angle, wind speed, cloud cover, and radiation stability parameters, are obtained. Based on the crop's stomatal conductance and leaf water potential physiological parameters, combined with the crop type and growth stage parameters, the water-sensitive and non-sensitive periods of the crop are determined. Based on the determined water-sensitive or non-sensitive periods, select the corresponding set of observation windows; Based on the selected set of observation windows, multiple types of time-series observation windows are divided in combination with environmental parameter thresholds; During the water-sensitive period, the time-series observation window is divided into a low-stress baseline observation window, a water-stress-sensitive observation window, and an afternoon recovery observation window. During the water-insensitive period, the time-series observation window is divided into a morning observation window and an afternoon observation window.

[0008] Furthermore, when dividing the time-series observation window into a low-stress baseline observation window, the following is included: Pre-set the minimum available threshold for photosynthetically active radiation, the low threshold for water vapor pressure deficit, and the maximum threshold for solar zenith angle; When the photosynthetically active radiation is greater than or equal to the minimum available threshold of photosynthetically active radiation, and the water vapor pressure deficit is less than the low threshold of water vapor pressure deficit, and the solar zenith angle is less than the maximum threshold of solar zenith angle, the current time period is determined to be the low stress benchmark observation window. When the photosynthetically active radiation is less than the minimum available threshold for photosynthetically active radiation, or the water vapor pressure deficit is greater than or equal to the low threshold for water vapor pressure deficit, or the solar zenith angle is greater than or equal to the maximum threshold for solar zenith angle, it is determined that the current time period does not meet the low stress benchmark observation window.

[0009] Furthermore, when dividing the time-series observation window into a water stress-sensitive observation window, the following is included: Pre-set the photosynthetically active radiation saturation threshold, water vapor pressure deficit stress threshold, cloud cover threshold, and radiation stability threshold; When the photosynthetically active radiation is greater than or equal to the photosynthetically active radiation saturation threshold, the water vapor pressure deficit is greater than or equal to the water vapor pressure deficit stress threshold, the cloud cover is less than the cloud cover threshold, and the radiation stability parameter is greater than the radiation stability threshold, the current time period is determined to be the water stress sensitive observation window. When the photosynthetically active radiation is less than the photosynthetically active radiation saturation threshold, or the water vapor pressure deficit is less than the water vapor pressure deficit stress threshold, or the cloud cover is greater than or equal to the cloud cover threshold, or the radiation stability parameter is less than or equal to the radiation stability threshold, it is determined that the current time period does not meet the water stress sensitive observation window.

[0010] Furthermore, when constructing environment-triggered judgment rules, the following are included: Obtain the first time period corresponding to the low-stress baseline observation window and the second time period corresponding to the water stress sensitive observation window; the first time period and the second time period are time periods within the same natural day, the first time period is earlier than the second time period, and the first time period and the second time period respectively meet the preset minimum effective duration; When both the first time period and the second time period are determined to be valid, an environment trigger command containing the flight mission of the first time period and the flight mission of the second time period is generated. When the time-series observation window includes an afternoon recovery observation window, the third time period corresponding to the afternoon recovery observation window is obtained, and the flight mission of the third time period is added to the environmental trigger command. When the first time period or the second time period is determined to be invalid, the environment trigger determination rule is determined to be ineffective, and the generation of the environment trigger instruction is terminated.

[0011] Furthermore, when controlling the UAV carrying the hyperspectral calibration payload to unify the flight trajectory and observation parameters, and collecting multi-temporal spectral data of the target area within the corresponding time-series observation window, the process includes: The flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy, and sensor observation angle of the UAV are preset. When the environmental trigger command is received, the UAV is controlled to perform the first spectral data acquisition in the first time period with the flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle. The UAV is controlled to perform a second spectral data acquisition in the second time period with the same flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle; When the environmental trigger command includes the third time period flight mission, the UAV is controlled to perform a third spectral data acquisition in the third time period with the same flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle.

[0012] Furthermore, after correcting and processing the raw spectral data collected by the UAV, the extraction of canopy SIF and apparent fluorescence yield based on characteristic bands and oxygen absorption valleys includes: The original spectral data is subjected to dark current correction, non-uniformity correction and radiometric calibration to obtain the first corrected spectral data; The first corrected spectral data is subjected to atmospheric scattering correction and atmospheric absorption correction to obtain the second corrected spectral data, wherein the atmospheric absorption correction is based on the oxygen absorption band corresponding to the oxygen absorption valley. Based on the second corrected spectral data, the first spectral reflectance corresponding to the characteristic band and the second spectral reflectance corresponding to the oxygen absorption valley are obtained; The canopy SIF is calculated based on the first spectral reflectance and the second spectral reflectance; The apparent fluorescence yield is calculated based on the ratio of the canopy SIF to the reference canopy SIF. Wherein, the reference canopy SIF is the canopy SIF corresponding to the low-stress baseline observation window, the canopy SIF and the reference canopy SIF have the same dimensions, and the apparent fluorescence yield is dimensionless.

[0013] Furthermore, when simultaneously completing the spatiotemporal scale registration of multi-source data and constructing the observation dataset, the following steps are included: Acquire real-time differential positioning data and ground control point coordinates during the flight of the UAV; Based on the real-time differential positioning data and the coordinates of the ground control points, the first image corresponding to the first spectral data acquisition and the second image corresponding to the second spectral data acquisition are registered at the pixel level. If the environmental triggering command includes a third time period flight mission, the third image corresponding to the third spectral data acquisition is also registered at the pixel level. This ensures that the same pixel in each image participating in the registration corresponds to the same canopy spatial unit; A multi-temporal observation dataset is constructed based on the registered images, with the canopy spatial units as the index. The multi-temporal observation dataset includes the canopy SIF and apparent fluorescence yield of each canopy spatial unit in the first time period and the second time period, respectively. When a third image is available, the dataset also includes the canopy SIF and apparent fluorescence yield for each canopy spatial unit at the third time period.

[0014] Furthermore, when obtaining water stress characterization data by removing various non-water stress interference signals from the observation dataset, distinguishing the temporal characteristics of soil drought and atmospheric drought, and fusing crop physiological and soil and water meteorological parameters, the process includes: Obtain the diurnal variation deviation of SIF for each canopy spatial unit in the observation dataset between the first time period and the second time period. If the observation dataset includes data for the third time period, obtain the SIF recovery of the canopy spatial unit between the second time period and the third time period. Based on the daily variation deviation of SIF, it is determined whether there is a significant decrease in SIF in each canopy spatial unit; When the daily variation deviation of SIF is greater than the preset deviation threshold, it is determined that there is a significant decrease in SIF, and the canopy temperature, air temperature, soil moisture content and water vapor pressure deficit of the corresponding canopy space unit are specifically obtained. When the difference between the canopy temperature and the air temperature is greater than a preset canopy temperature difference threshold, and the water vapor pressure deficit is greater than a preset water vapor stress threshold, and the soil moisture content is less than a preset soil moisture threshold, the current drought type is determined to be soil drought. When the difference between the canopy temperature and the air temperature is less than or equal to the preset canopy temperature difference threshold, and the water vapor pressure deficit is greater than the preset water vapor stress threshold, and the soil moisture content is greater than or equal to the preset soil moisture threshold, the current drought type is determined to be atmospheric drought. When the crown-temperature difference is greater than the preset crown-temperature difference threshold, and the vapor pressure deficit is greater than the preset vapor stress threshold, and the soil moisture content is greater than or equal to the preset soil moisture threshold, or when the crown-temperature difference is greater than the preset crown-temperature difference threshold, the vapor pressure deficit is less than or equal to the preset vapor stress threshold, and the soil moisture content is less than the preset soil moisture threshold, the current drought type is determined to be soil-atmosphere combined drought. When there is SIF recovery and the SIF recovery is greater than a preset recovery threshold, it is determined that the significant decrease in SIF is due to highlight suppression rather than moisture stress. When the SIF recovery amount is less than or equal to the preset recovery threshold, and the daily SIF deviation is greater than the preset deviation threshold, the significant decrease in SIF is determined to be due to water stress. Based on the drought type identification results, the stress level, diagnostic confidence level, and irrigation priority are calculated by combining the SIF decrease rate and soil moisture deficit degree. The stress level, spatial distribution, diagnostic confidence level, and irrigation priority corresponding to each canopy spatial unit are output.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: Based on parameters of farmland environment, crop physiology, and growth stage, time-series observation windows are divided and environmental triggering judgment rules are established. After the rules take effect, the UAV is controlled to complete multi-temporal spectral acquisition using a unified flight trajectory and observation parameters, ensuring the consistency of observation data benchmarks across different time phases. This overcomes the problems of existing technologies lacking an environment-physiology co-triggered mechanism and inconsistent multi-temporal observation conditions leading to poor data comparability. Canopy SIF and apparent fluorescence yield are obtained through correction and inversion. By relying on the multi-temporal observation dataset to eliminate interference signals from non-water stress, the temporal characteristics of soil drought and atmospheric drought are distinguished, effectively identifying false stress signals caused by high light suppression, reducing the false judgment rate of stress identification, and solving the problem of insufficient identification accuracy caused by the difficulty of distinguishing high light suppression, atmospheric drought, and soil drought stress signals in traditional schemes.

[0016] On the other hand, this application also provides an environment-triggered UAV multi-temporal SIF collaborative observation and crop water stress identification system, including: The environmental perception and window segmentation module is configured to collect environmental parameters of the target farmland and physiological parameters of crop stomatal conductance and leaf water potential, and to divide multiple time-series observation windows by combining crop type and growth stage parameters, and to construct environmental trigger judgment rules. The UAV aerial survey control module is configured to control the UAV carrying the hyperspectral calibration payload to uniformly fly the flight trajectory and observe the parameters when the environmental trigger judgment rule takes effect, and to collect multi-temporal spectral data of the target area in the corresponding time-series observation window. The spectral processing and SIF inversion module is configured to correct and process the raw spectral data collected by the UAV, extract the canopy SIF and apparent fluorescence yield based on the characteristic bands and oxygen absorption valleys, perform atmospheric radiation correction in combination with the UAV's low-altitude atmospheric parameters, and simultaneously complete the spatiotemporal scale registration of multi-source data to construct the observation dataset. The stress identification and prescription generation module is configured to remove various non-water stress interference signals based on the observation dataset, distinguish the temporal characteristics of soil drought and atmospheric drought, integrate crop physiological and soil and water meteorological parameters to obtain water stress characterization data, output stress level, spatial distribution, diagnostic confidence and irrigation priority data, and perform data linkage with variable irrigation equipment.

[0017] It is understandable that the above-mentioned method and system for multi-temporal SIF collaborative observation of crops based on environment triggering by UAVs has the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0018] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart of a method for environmentally triggered UAV multi-temporal SIF collaborative observation and crop water stress identification provided in an embodiment of the present invention; Figure 2 This is a functional block diagram of an environment-triggered UAV multi-temporal SIF collaborative observation and crop water stress identification system provided in an embodiment of the present invention. Detailed Implementation

[0019] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] Reference Figure 1 In some embodiments of this application, the present invention proposes a method for multi-temporal SIF collaborative observation and crop water stress identification based on environment-triggered UAVs, comprising the following steps: Step S100: Collect environmental parameters of the target farmland and physiological parameters of crop stomatal conductance and leaf water potential. Based on the physiological parameters, combine crop type and growth stage parameters to determine the sensitive period of the crop, select the corresponding observation window set, and based on the selected observation window set, combine environmental parameter thresholds to divide multiple types of time-series observation windows and construct environmental trigger judgment rules. Step S200: When the environmental trigger determination rule takes effect, control the UAV carrying the hyperspectral calibration payload to collect multi-temporal spectral data of the target area in the corresponding time-series observation window with unified flight trajectory and observation parameters; Step S300: After correcting and processing the raw spectral data collected by the UAV, extract the canopy SIF and apparent fluorescence yield based on the characteristic bands and oxygen absorption valleys, perform atmospheric radiation correction in combination with the UAV's low-altitude atmospheric parameters, and simultaneously complete the spatiotemporal scale registration of multi-source data to construct an observation dataset. Step S400: Based on the observation dataset, remove various non-water stress interference signals, distinguish the temporal characteristics of soil drought and atmospheric drought, integrate crop physiological and soil and water meteorological parameters to obtain water stress characterization data, output stress level, spatial distribution, diagnostic confidence and irrigation priority data, and link with variable irrigation equipment.

[0021] Specifically, a hyperspectral calibration payload is an unmanned aerial vehicle (UAV) payload with spectral imaging acquisition and radiometric calibration capabilities, including at least a hyperspectral imager, an uplink irradiance sensor, and a radiometric reference device. The radiometric reference device is a calibration plate or an airborne built-in radiometric reference source. Among them, the hyperspectral imager is used to acquire hyperspectral images of farmland canopy, the uplink irradiance sensor is used to simultaneously acquire the incident solar irradiance above the canopy, and the radiometric reference device is used for radiometric calibration of spectral data.

[0022] Specifically, using the canopy spatial unit as the smallest output unit, water stress characterization data is mapped to multi-level stress levels, generating a corresponding raster-formatted stress spatial distribution map. The diagnostic confidence of each unit is calculated by combining pixel spectral quality, observation window validity, and multi-parameter threshold deviation. Irrigation priorities for each spatial unit are quantified based on stress level, soil moisture, and crop water requirements. A prescription data package containing spatial location, stress level, diagnostic confidence, irrigation water demand, and irrigation priority is sent to variable irrigation equipment. The variable irrigation equipment performs differentiated irrigation based on the spatial information within the data package, while simultaneously receiving operational status data from the variable irrigation equipment, completing the observation-diagnosis-irrigation closed loop.

[0023] The above embodiments divide the time-series observation windows based on farmland environment, crop physiology, and growth stage parameters, and establish environmental triggering judgment rules. After the rules take effect, the UAV is controlled to complete multi-temporal spectral acquisition using a unified flight trajectory and observation parameters, ensuring the consistency of observation data benchmarks across different time phases. This overcomes the problems of existing technologies lacking an environment-physiology co-triggered mechanism and inconsistent multi-temporal observation conditions leading to poor data comparability. By correcting and inverting, the canopy SIF and apparent fluorescence yield are obtained. Based on the multi-temporal observation dataset, non-water stress interference signals are eliminated, and the temporal characteristics of soil drought and atmospheric drought are distinguished. This effectively identifies false stress signals caused by high light suppression, reduces the false judgment rate of stress identification, and solves the problem of insufficient identification accuracy caused by traditional schemes that are difficult to distinguish between high light suppression, atmospheric drought, and soil drought stress signals.

[0024] Specifically, when dividing time series observation windows into multiple categories, the following are included: Acquire environmental parameters including photosynthetically active radiation, air temperature, relative humidity, water vapor pressure deficit, solar zenith angle, wind speed, cloud cover, and radiation stability parameters; Based on the physiological parameters of crop stomatal conductance and leaf water potential, combined with crop type and growth stage parameters, determine the water-sensitive and non-sensitive periods of the crop. Based on the determined water-sensitive or non-sensitive periods, select the corresponding set of observation windows; Based on the selected set of observation windows, and combined with environmental parameter thresholds, multiple types of time-series observation windows are classified. During the water-sensitive period, the time-series observation window is divided into a low-stress baseline observation window, a water-stress-sensitive observation window, and an afternoon recovery observation window. During the water-insensitive period, the time-series observation window is divided into a morning observation window and an afternoon observation window.

[0025] Specifically, environmental parameters are obtained from multi-element meteorological sensing terminals deployed in farmland. Sampling is completed at a set time step. The sampling content includes photosynthetically active radiation, air temperature, relative humidity, water vapor pressure deficit, solar zenith angle, wind speed, cloud cover, and radiation stability parameters. Among them, photosynthetically active radiation refers to solar radiation in the 400-700nm band that can be used by crops for photosynthesis; water vapor pressure deficit is the difference between saturated water vapor pressure and actual air water vapor pressure at the same temperature; solar zenith angle represents the angle between the direction of solar incidence and the local zenith direction; radiation stability parameter is the short-term fluctuation coefficient of photosynthetically active radiation within a specified time window; cloud cover is the proportion of cloud cover within the field of view; and wind speed is the measured wind speed at a fixed height above the farmland canopy. The crop type parameter records the variety of crops planted in the target farmland. The growth stage parameter is taken from the output of the crop growth model or field phenological observation records. Based on the crop type, growth stage parameter, and physiological parameters, the growth stage of the target crop is determined, and water-sensitive and non-sensitive periods are identified. The water-sensitive period is the growth stage in which the crop responds significantly to water deficit, while the non-sensitive period is the growth stage in which the crop is less affected by short-term water deficit. Based on the determined water-sensitive or non-sensitive period, a corresponding set of observation windows is selected. Based on the selected set of observation windows and combined with environmental parameter thresholds, multiple types of time-series observation windows are divided. For crops in the water-sensitive period, three independent time intervals are defined according to meteorological element threshold conditions, serving as the low-stress baseline observation window, the water-stress sensitive observation window, and the afternoon recovery observation window, respectively. For crops in the water-non-sensitive period, two time intervals are divided according to the local solar cycle, serving as the morning observation window and the afternoon observation window, respectively. Each observation window outputs two sets of time stamps: the start time and the end time.

[0026] Understandably, by collecting real-time field environmental parameters such as farmland light, temperature and humidity, atmospheric dryness and wetness, sky radiation and meteorological stability, and combining them with the inherent growth characteristics of different crops and the phenological characteristics of real-time growth stages, it is possible to distinguish between growth periods when crops are highly sensitive to water deficiency and non-sensitive growth periods when they are more tolerant. Based on the differences in water response at different growth stages of crops, observation periods are differentiated. During the water-sensitive period of crops, multiple time-series observation windows are constructed to represent the basic physiological state, stress response state and stress recovery state of crops, respectively. During the non-sensitive period, the observation windows are simplified to morning and afternoon. In this way, based on environmental conditions and crop growth patterns, standardized observation time intervals suitable for different growth stages are matched, providing accurate time determination basis for subsequent environmentally triggered observations.

[0027] Specifically, when dividing the time-series observation window into a low-stress baseline observation window, the following is included: Pre-set the minimum available threshold for photosynthetically active radiation, the low threshold for water vapor pressure deficit, and the maximum threshold for solar zenith angle; When the photosynthetically active radiation is greater than or equal to the minimum available threshold of photosynthetically active radiation, and the water vapor pressure deficit is less than the low threshold of water vapor pressure deficit, and the solar zenith angle is less than the maximum threshold of solar zenith angle, the current time period is determined to be the low stress benchmark observation window. When the photosynthetically active radiation is less than the minimum available threshold for photosynthetically active radiation, or the water vapor pressure deficit is greater than or equal to the low threshold for water vapor pressure deficit, or the solar zenith angle is greater than or equal to the maximum threshold for solar zenith angle, it is determined that the current time period does not meet the low stress benchmark observation window.

[0028] Specifically, the minimum available threshold for photosynthetically active radiation (PANR), the low threshold for water vapor pressure deficit, and the maximum threshold for solar zenith angle are pre-stored. The thresholds are calibrated according to the target crop species. The minimum available threshold for PANR is set to 800 μmol·m⁻²·s⁻¹, the low threshold for water vapor pressure deficit is set to 1.2 kPa, and the maximum threshold for solar zenith angle is set to 60°. PANR is the shortwave solar radiation flux reaching the farmland canopy in the 400-700 nm band. Water vapor pressure deficit refers to the difference between the saturated water vapor pressure and the actual water vapor pressure under the current temperature conditions. Solar zenith angle is the angle between the incident solar rays and the vertical line at the observation location. The environmental parameters collected at each time step are read and compared with the preset thresholds. When the real-time photosynthetically active radiation is greater than or equal to 800 μmol·m⁻²·s⁻¹, the real-time water vapor pressure deficit is less than 1.2 kPa, and the real-time solar zenith angle is less than 60°, the continuous time segment is marked as a low-stress baseline observation window, and the start and end times of the window are output. When the real-time photosynthetically active radiation is less than 800 μmol·m⁻²·s⁻¹, or the real-time water vapor pressure deficit is greater than or equal to 1.2 kPa, or the real-time solar zenith angle is greater than or equal to 60°, the time segment is not marked as a low-stress baseline observation window.

[0029] Understandably, the pre-calibrated minimum usable threshold for photosynthetically active radiation, low threshold for water vapor pressure deficit, and maximum threshold for solar zenith angle are retrieved, and the corresponding environmental parameters collected in real time in the field are continuously read. Each parameter is then compared with its corresponding threshold. Only when all three environmental parameters—photosynthetically active radiation, water vapor pressure deficit, and solar zenith angle—simultaneously meet the constraints are the continuous period identified as a low-stress baseline observation window. If any one of the environmental parameters fails to meet the threshold constraint, the period is excluded. This process filters out time intervals with sufficient sunlight, low atmospheric water vapor deficit, and suitable solar incidence angles to obtain baseline spectral observation data for crops under conditions of no significant water stress.

[0030] Specifically, when dividing the time-series observation window into a water stress-sensitive observation window, the following is included: Pre-set the photosynthetically active radiation saturation threshold, water vapor pressure deficit stress threshold, cloud cover threshold, and radiation stability threshold; When the photosynthetically active radiation is greater than or equal to the photosynthetically active radiation saturation threshold, the water vapor pressure deficit is greater than or equal to the water vapor pressure deficit stress threshold, the cloud cover is less than the cloud cover threshold, and the radiation stability parameter is greater than the radiation stability threshold, the current time period is determined to be the water stress sensitive observation window. When the photosynthetically active radiation is less than the photosynthetically active radiation saturation threshold, or the water vapor pressure deficit is less than the water vapor pressure deficit stress threshold, or the cloud cover is greater than or equal to the cloud cover threshold, or the radiation stability parameter is less than or equal to the radiation stability threshold, it is determined that the current time period does not meet the water stress sensitive observation window.

[0031] Specifically, the window segmentation process relies on four types of fixed calibration thresholds to screen time periods. Fixed parameter thresholds adapted to field observation scenarios are pre-stored in the system: the photosynthetically active radiation saturation threshold is set at 1200 μmol·m⁻²·s⁻¹, the water vapor pressure deficit stress threshold is set at 1.8 kPa, the cloud cover threshold is set at 20%, and the radiation stability threshold is set at 0.95. Among these, the photosynthetically active radiation saturation threshold is the minimum light intensity required for crop photosynthetic response to reach saturation; the water vapor pressure deficit stress threshold is the critical value at which atmospheric aridity induces a water stress response in crops; the cloud cover threshold is the upper limit of the proportion of cloud cover in the sky, used to avoid spectral data distortion caused by cloud obstruction; and the radiation stability parameter is the fluctuation stability coefficient of photosynthetically active radiation per unit time period. The closer the value is to 1, the smaller the light fluctuation and the more stable the light environment. Its calculation formula is:

[0032] in, For radiation stability parameters; The standard deviation of photosynthetically active radiation per unit time period; This represents the average photosynthetically active radiation (PADR) within the same time period. The system collects real-time environmental data of farmland every second and simultaneously conducts parallel verification under multiple conditions. Only when the real-time PADR is not less than 1200 μmol·m⁻²·s⁻¹, the real-time vapor pressure deficit is not less than 1.8 kPa, the real-time cloud cover percentage is less than 20%, and the real-time radiation stability coefficient is greater than 0.95, is the current continuous period determined to be an effective water stress sensitive observation window. If any of the above four parameters fails to meet the corresponding threshold standard, the current period is directly determined not to meet the observation window requirements and is not included in the stress observation period.

[0033] Understandably, by calling pre-set thresholds for photosynthetically active radiation saturation, water vapor pressure deficit stress, cloud cover, and radiation stability, four corresponding environmental parameters are continuously acquired in real time in the field. Threshold comparison calculations are performed on the four parameters simultaneously. Only when the light reaches the crop photosynthetic saturation level, the atmospheric water vapor deficit reaches a level that can induce crop stress response, the cloud interference is small, and the light radiation state remains stable, is the continuous period identified as a water stress sensitive observation window. If any environmental parameter does not meet the threshold constraint, the period is removed. This is used to screen time intervals that can stimulate the physiological response of crops to water stress, which are then used to capture the spectral fluorescence signal of crops after being subjected to water stress.

[0034] Specifically, when constructing environment-triggered judgment rules, the following are included: Obtain the first time period corresponding to the low-stress baseline observation window and the second time period corresponding to the water stress sensitive observation window; the first time period and the second time period are time periods within the same natural day, the first time period is earlier than the second time period, and the first time period and the second time period respectively meet the preset minimum effective duration; When both the first time period and the second time period are determined to be valid, an environment trigger command containing the flight mission of the first time period and the flight mission of the second time period is generated. When the time-series observation window includes an afternoon recovery observation window, the third time period corresponding to the afternoon recovery observation window is obtained, and the flight mission of the third time period is added to the environmental trigger command. When the first time period or the second time period is determined to be invalid, the environment trigger determination rule is determined to be ineffective, and the generation of the environment trigger instruction is terminated.

[0035] Specifically, the system first reads the low-stress baseline observation window and the water-stress sensitive observation window, which have been filtered by environmental parameter thresholds for the day. These are defined as the first observation period and the second observation period, respectively. The criterion for determining the validity of a single period is that the continuous valid duration is not less than 40 minutes. The system executes a joint validity verification logic for the two periods. Only if both the first and second periods are determined to be valid and usable periods, a standardized UAV aerial survey environment trigger command is generated. The command is embedded with the fixed-point, fixed-orbit, and fixed-parameter flight mission sequence corresponding to the first and second periods. If the current crop is in a water-sensitive period, the system checks whether a valid afternoon recovery observation window has been generated. If a valid third period exists, the independent flight mission corresponding to that period is synchronously superimposed and entered into the environmental trigger command, forming a three-period linked aerial survey mission queue. If either the first or second period does not meet the valid duration standard or the parameter threshold verification fails, the system directly determines that the environmental trigger conditions for the day are not met, locks the mission generation program, terminates the editing and issuance process of all UAV aerial survey commands, and does not retain the aerial survey mission queue for the day.

[0036] Understandably, the system retrieves the window information of the first and second time periods after threshold filtering, and performs validity identification on the two time periods separately. Only when the first and second time periods are both valid will an environmental trigger command carrying the corresponding flight mission be generated. If there is an available afternoon recovery observation window, the flight mission of the third time period will be included in the command. Once either the first or second time period is determined to be invalid, the environmental trigger judgment rule will be directly invalidated, and the generation process of the environmental trigger command will be stopped. This completes the multi-window joint verification and determines whether to start the UAV multi-temporal acquisition mission.

[0037] Specifically, when controlling a UAV carrying a hyperspectral calibration payload to unify its flight trajectory and observation parameters, and collecting multi-temporal spectral data of the target area within the corresponding time-series observation window, the process includes: Pre-set the drone's flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy, and sensor observation angle; When an environmental trigger command is received, the UAV is controlled to perform the first spectral data acquisition in the first time period with the flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle. The UAV was controlled to perform a second spectral data acquisition in the second time period with the same flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle; When the environmental trigger command includes a third-period flight mission, the UAV is controlled to perform a third spectral data acquisition in the third period with the same flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy, and sensor observation angle.

[0038] Specifically, the system pre-configures and stores all aerial survey parameters for the UAV, unifies the acquisition hardware and flight reference parameters for all-time data collection, and avoids dynamic random adjustments to geometric flight parameters. Specifically, the flight altitude is fixed at 120m (relative to ground elevation); the flight trajectory uses a full-coverage parallel strip-shaped route, with pre-entered target farmland vector boundaries generating a fixed waypoint sequence; the flight heading is uniformly set to due north; the heading overlap is fixed at 85%, representing the longitudinal overlap ratio of adjacent preceding and following images; the lateral overlap is fixed at 80%, representing the lateral overlap ratio of adjacent parallel route images; the exposure control strategy is set to a unified fixed-point automatic exposure mode, with the payload integration time adaptively adjusted according to lighting conditions based on this exposure control strategy; the sensor observation angle maintains a vertical, 0° orthogonal observation attitude with no tilted viewpoint.

[0039] Specifically, the system monitors environmental trigger commands in real time. Upon successful receipt of a command, it uses all preset fixed aerial survey parameters to complete the first round of full-domain hyperspectral data acquisition during the first time period. In the second time period, it strictly reuses identical flight paths, flight attitudes, and imaging parameters to complete the second round of spectral data acquisition. If the issued environmental trigger command includes a third-time resumption observation flight mission, the system continues to use the same unchanged aerial survey configuration to complete the third round of spectral data acquisition during the corresponding afternoon resumption observation period, ensuring that the acquisition hardware parameters and flight geometry parameters remain completely consistent across all time phases.

[0040] Understandably, the flight geometry parameters and hyperspectral payload imaging parameters of the UAV are pre-configured. After receiving the environmental trigger command, the same set of preset flight and imaging parameters are used to complete two spectral acquisitions in the first and second time periods, respectively. When the command carries the flight mission for the third time period, the same set of parameters is used to carry out the third spectral acquisition in the third time period. This keeps the flight path, flight attitude, and sensor imaging conditions consistent in different time-series observation processes, eliminates external variables introduced by differences in aerial survey conditions, and obtains multiple sets of multi-temporal original spectral images with unified observation conditions.

[0041] Specifically, after correcting and processing the raw spectral data collected by the UAV, the extraction of canopy SIF and apparent fluorescence yield based on characteristic bands and oxygen absorption valleys includes: The original spectral data is subjected to dark current correction, non-uniformity correction and radiometric calibration to obtain the first corrected spectral data; The first corrected spectral data is subjected to atmospheric scattering correction and atmospheric absorption correction to obtain the second corrected spectral data, wherein the atmospheric absorption correction is based on the oxygen absorption band corresponding to the oxygen absorption valley. Based on the second corrected spectral data, the first spectral reflectance corresponding to the characteristic band and the second spectral reflectance corresponding to the oxygen absorption valley are obtained; The canopy SIF is calculated based on the first spectral reflectance and the second spectral reflectance; The apparent fluorescence yield is calculated based on the ratio of the canopy SIF to the reference canopy SIF. Wherein, the reference canopy SIF is the canopy SIF corresponding to the low-stress baseline observation window, the canopy SIF and the reference canopy SIF have the same dimensions, and the apparent fluorescence yield is dimensionless.

[0042] Specifically, the first step is to perform basic hardware error correction and radiometric calibration on the raw hyperspectral data acquired by the UAV aerial photography. The raw spectral data consists of raw digitally quantized grayscale values ​​directly collected by the UAV's hyperspectral payload. Dark current correction is used to eliminate the inherent dark current noise of the sensor under dark conditions. By acquiring the dark current mean matrix under the darkened state of the device, the dark current offset of the raw data is deducted pixel by pixel. Non-uniformity correction relies on the response non-uniformity coefficient matrix calibrated by the sensor at the factory to correct the differences in photosensitivity response of different pixels. Radiometric calibration uses the standard radiometric calibration coefficients provided with the device to convert the spectral grayscale values ​​into absolute radiance values. After the calibration is completed, the first calibrated spectral data is output.

[0043] Specifically, low-altitude atmospheric interference correction processing is performed on the first corrected spectral data, including two types of correction procedures: atmospheric scattering correction and atmospheric absorption correction. Atmospheric scattering correction is used to eliminate spectral radiation deviations caused by scattering from low-altitude atmospheric aerosols and air molecules in farmland. Atmospheric absorption correction is carried out based on the 760nm characteristic band of the oxygen absorption valley. This band is a characteristic band of strong atmospheric oxygen absorption and is the core reference band for low-altitude atmospheric radiation correction. By using the atmospheric absorption attenuation coefficient of this band, the atmospheric absorption loss of the spectral data is corrected, and finally, the second corrected spectral data with atmospheric interference eliminated is output.

[0044] Specifically, the oxygen absorption valley is the 760nm absorption band of oxygen A (O2-A); the area inside the oxygen A absorption band is the absorption valley band, and 750nm and 770nm on both sides are selected as out-of-band characteristic reference bands; the canopy SIF can be extracted using Fraunhofer line discrimination (FLD), three-band Fraunhofer line discrimination (3FLD), or spectral fitting (SFM); for UAV low-altitude observation scenarios, the improved 3FLD or improved SFM algorithm with low-altitude atmospheric parameter compensation is preferred to adapt to the radiative transfer characteristics of low-altitude aerosols and oxygen absorption, reducing the inversion error introduced by low-altitude observation. The canopy SIF calculation uses the oxygen absorption valley filling method, and the calculation formula is: Apparent fluorescence yield The average reference reflectance of the characteristic bands outside the oxygen absorption valley on both sides; The measured reflectance within the oxygen absorption valley; is the incident spectral radiance of the canopy; SIFref is the reference canopy SIF, i.e., the canopy SIF corresponding to the low-stress benchmark observation window.

[0045] Specifically, two types of core spectral parameters are extracted from the second corrected spectral data. The 750nm and 770nm characteristic bands that are sensitive to crop photosynthetic fluorescence response are selected, and the spectral reflectance of the corresponding bands is extracted as the first spectral reflectance. The 760nm oxygen absorption valley center band is accurately extracted, and the spectral reflectance at this position is extracted as the second spectral reflectance. The extraction accuracy of all bands is controlled within ±0.5nm.

[0046] Specifically, the canopy SIF is calculated using the classical oxygen absorption valley filling method based on two sets of spectral reflectance parameters. The core calculation formula is as follows: In the formula The intensity of sunlight-induced chlorophyll fluorescence in the canopy at 760 nm is given in units of [missing information]. ; The average reference reflectance of the characteristic bands on both sides of the oxygen absorption valley; Measured reflectance of the oxygen absorption valley; The radiance of the incident spectrum of the canopy.

[0047] Specifically, based on the calculated canopy SIF value and the reference canopy SIF parameter, the apparent fluorescence yield is calculated. The core calculation formula is as follows: The apparent fluorescence yield is dimensionless. The fluorescence intensity of the canopy at 760 nm. The reference canopy SIF is the canopy SIF corresponding to the low-stress baseline observation window; SIF and SIFref are both canopy radiance with the same dimension, and the resulting apparent fluorescence yield is dimensionless.

[0048] It should be noted that the "apparent fluorescence yield" mentioned in this embodiment is a relative fluorescence index obtained based on the SIF of the low-stress baseline observation window on the same day. It is a custom index in this embodiment and is different from the fluorescence quantum yield, which is the ratio of fluorescence output to absorbed excitation light in the traditional physical definition. It is used to characterize the degree of relative change of the canopy SIF relative to the unstressed baseline state.

[0049] Understandably, the raw spectral data from the UAV is first subjected to dark current correction, non-uniformity correction, and radiometric calibration to eliminate sensor hardware acquisition errors and complete radiometric quantization conversion, resulting in preliminary calibrated spectral data. Then, atmospheric scattering and oxygen absorption valley band-specific atmospheric absorption correction are performed on this data to eliminate spectral interference from the low-altitude atmospheric environment, obtaining high-precision surface canopy spectral data. Subsequently, the spectral reflectance of the characteristic band and the oxygen absorption valley band are extracted separately. Based on the two types of reflectance parameters, the canopy SIF value is obtained by inversion using the oxygen absorption valley filling algorithm. Finally, combined with the reference canopy SIF parameter, the dimensionless crop canopy apparent fluorescence yield is obtained by ratio calculation, achieving accurate and standardized extraction of multi-temporal fluorescence characteristic parameters.

[0050] Specifically, when simultaneously completing the spatiotemporal scale registration of multi-source data and constructing the observation dataset, the following steps are included: Acquire real-time differential positioning data and ground control point coordinates during UAV flight; Based on the real-time differential positioning data and the coordinates of the ground control points, the first image corresponding to the first spectral data acquisition and the second image corresponding to the second spectral data acquisition are registered at the pixel level. If the environmental triggering command includes a third time period flight mission, the third image corresponding to the third spectral data acquisition is also registered at the pixel level. This ensures that the same pixel in each image participating in the registration corresponds to the same canopy spatial unit; Based on the registered images, a multi-temporal observation dataset is constructed with the canopy spatial units as the index. The multi-temporal observation dataset contains the canopy SIF and apparent fluorescence yield of each canopy spatial unit in the first time period and the second time period, respectively. When a third image is available, the dataset also includes the canopy SIF and apparent fluorescence yield for each canopy spatial unit at the third time period.

[0051] Specifically, throughout each UAV aerial survey operation, onboard RTK real-time differential positioning data is continuously collected at a frequency of 10Hz. This data includes the original attitude information of the UAV at the time of exposure, including longitude, latitude, flight altitude, pitch angle, roll angle, and yaw angle. At least six ground control points are pre-deployed in the target farmland. These control points are white square targets, evenly distributed along the edges and interior of the farmland. The absolute planar and elevation coordinates of each control point are measured using high-precision GNSS equipment, with an accuracy of ±1cm, serving as the reference coordinate parameters for image spatial correction.

[0052] Specifically, a polynomial correction model is used to complete the pixel-level spatial registration of multi-temporal images. The correction model formula is as follows: Where x and y are the original pixel coordinates of the image, and X and Y are the geographic coordinates of the ground control points. These are polynomial correction coefficients. Hyperspectral images acquired at least in the first and second time periods are substituted into the model to complete geometric correction and resampling. When a third image acquired in the third time period exists, it is also included in the correction. Bilinear interpolation is used for resampling, and the spatial resolution of the output images is uniformly 0.1m. After registration, the pixel position deviation of images from multiple periods is controlled within 0.2 pixels, ensuring that the same pixel in images from different time periods corresponds to a unique crop canopy spatial unit in the field.

[0053] Specifically, a canopy spatial unit is defined as the smallest geographic raster unit of a single registered image, using a combination of raster row and column numbers and geographic coordinates as a unique index identifier. After registration, inversion parameters corresponding to each time period are extracted for each unit and uniformly collected to construct a structured multi-temporal observation dataset. Each data record in the dataset contains a unique canopy spatial unit code, unit geographic coordinates, canopy SIF value for the first time period, apparent fluorescence yield for the first time period, canopy SIF value for the second time period, apparent fluorescence yield for the third time period, and the corresponding observation timestamp. If a third image exists, the canopy SIF value and apparent fluorescence yield for the third time period are further recorded, achieving one-to-one binding and storage of fluorescence parameters and spatial units across multiple time periods.

[0054] Understandably, by collecting real-time differential positioning data and high-precision ground control point coordinates throughout the entire UAV aerial survey process, spatial positioning benchmarks for multi-phase spectral images are obtained. Based on these benchmarks, pixel-level spatial registration is uniformly performed on multi-source images corresponding to different observation periods to eliminate geometric positional deviations in multi-phase images. This ensures that each phase image pixel corresponds precisely to a field crop canopy spatial unit. Then, using the standardized canopy spatial unit as a unique index, the canopy SIF and apparent fluorescence yield data corresponding to each time period that have been collected are organized and summarized. This completes the spatiotemporal fusion of multi-source heterogeneous data, ultimately constructing a standardized multi-temporal observation dataset that is temporally continuous and spatially aligned.

[0055] Specifically, when obtaining water stress characterization data by removing various non-water stress interference signals from the observation dataset, distinguishing the temporal characteristics of soil drought and atmospheric drought, and fusing crop physiological and soil and water meteorological parameters, the following are included: Obtain the diurnal variation deviation of SIF for each canopy spatial unit in the observation dataset between the first and second time periods. If the observation dataset includes data for the third time period, obtain the SIF recovery of that canopy spatial unit between the second and third time periods. Based on the daily deviation of SIF, determine whether there is a significant decrease in SIF in each canopy spatial unit; When the daily variation deviation of SIF exceeds the preset deviation threshold, it is determined that there is a significant decrease in SIF, and the canopy temperature, air temperature, soil moisture content and water vapor pressure deficit of the corresponding canopy spatial unit are obtained. When the difference between the canopy temperature and the air temperature is greater than the preset canopy temperature difference threshold, the water vapor pressure deficit is greater than the preset water vapor stress threshold, and the soil moisture content is less than the preset soil moisture threshold, the current drought type is determined to be soil drought. When the difference between the canopy temperature and the air temperature is less than or equal to the preset canopy temperature difference threshold, the water vapor pressure deficit is greater than the preset water vapor stress threshold, and the soil moisture content is greater than or equal to the preset soil moisture threshold, the current drought type is determined to be atmospheric drought. When the crown-temperature difference is greater than the preset crown-temperature difference threshold, and the vapor pressure deficit is greater than the preset vapor stress threshold, and the soil moisture content is greater than or equal to the preset soil moisture threshold, or when the crown-temperature difference is greater than the preset crown-temperature difference threshold, the vapor pressure deficit is less than or equal to the preset vapor stress threshold, and the soil moisture content is less than the preset soil moisture threshold, the current drought type is determined to be soil-atmosphere combined drought. When there is SIF recovery and the SIF recovery is greater than the preset recovery threshold, the significant decrease in SIF is determined to be due to highlight suppression rather than moisture stress. When the SIF recovery amount is less than or equal to the preset recovery threshold and the daily SIF deviation is greater than the preset deviation threshold, the significant decrease in SIF is determined to be water stress. Based on the drought type identification results, the stress level, diagnostic confidence level, and irrigation priority are calculated by combining the SIF decrease rate and soil moisture deficit degree. The stress level, spatial distribution, diagnostic confidence level, and irrigation priority corresponding to each canopy spatial unit are output.

[0056] Specifically, each canopy spatial unit is taken as an independent analysis object, and parameter difference calculation is performed based on multi-temporal observation datasets. The daily variation deviation of SIF is the numerical difference between the canopy SIF in the first time period and the second time period, and the preset deviation threshold is fixed at 0.03mW・m-2・sr-1・nm-1. When the observation dataset includes observation data in the third time period, the SIF recovery is the numerical difference between the canopy SIF in the second time period and the third time period, and the preset recovery threshold is fixed at 0.02mW・m-2・sr-1・nm-1.

[0057] Specifically, the diurnal variation deviation data of SIF for all canopy spatial units are traversed, and the deviation values ​​of each unit are compared with preset deviation thresholds to quantify the SIF decline characteristics. When the diurnal variation deviation of SIF for a canopy spatial unit is less than or equal to the preset deviation threshold of 0.03 mW·m⁻²·sr⁻¹·nm⁻¹, the canopy spatial unit is determined to have no significant SIF decline and is therefore considered to be without water stress, requiring no further drought type identification. For target units with a diurnal variation deviation of SIF greater than 0.03, multi-source measured parameters matched to that unit are simultaneously retrieved, including four basic parameters: canopy temperature, near-surface 2m air temperature, soil moisture content in the 0-20cm soil layer, and atmospheric water vapor pressure deficit.

[0058] Specifically, the system has built-in fixed discrimination thresholds, with the canopy temperature difference threshold set at 3℃, the water vapor stress threshold set at 1.8kPa, and the soil moisture threshold set at 60% of field capacity. When the difference between the unit canopy temperature and the air temperature is greater than 3℃, the water vapor pressure deficit is greater than 1.8kPa, and the soil moisture content is less than 60% of field capacity, the unit is marked as drought type, which is mainly caused by insufficient soil water supply leading to crop water stress.

[0059] Specifically, when the difference between the unit canopy temperature and the air temperature is less than or equal to 3℃, the water vapor pressure deficit is greater than 1.8kPa, and the soil moisture content is greater than or equal to 60% of the field capacity, the drought type of the unit is marked as atmospheric drought. This type is caused only by atmospheric drought and there is no soil water shortage.

[0060] When the canopy-temperature difference is greater than 3℃, the vapor pressure deficit is greater than 1.8kPa, and the soil moisture content is greater than or equal to 60% of the field capacity, or when the canopy-temperature difference is greater than 3℃, the vapor pressure deficit is less than or equal to 1.8kPa, and the soil moisture content is less than 60% of the field capacity, it is determined to be a soil-atmosphere combined drought. This branch covers two types of boundary conditions that traditional algorithms cannot identify: one is that the soil moisture is sufficient, but the stress response is induced by the combined effects of the atmosphere and the canopy; the other is a mixed stress scenario where the soil moisture is deficient, but the degree of atmospheric vapor pressure deficit is not high.

[0061] Specifically, given the availability of observation data from a third time period, the SIF recovery value of the selected units with significant SIF decline is compared with a preset recovery threshold. If the SIF recovery value is greater than 0.02, the SIF decline during that time period is determined to be a photosynthetic light inhibition phenomenon caused by high light intensity at midday, and is classified as a non-water stress interference signal. If the SIF recovery value is less than or equal to 0.02, and the diurnal variation deviation of SIF is greater than 0.03, light inhibition interference is excluded, and the significant SIF decline characteristic is determined to be caused by actual crop water stress.

[0062] Specifically, based on the drought type identification results of soil drought, atmospheric drought, and soil-atmosphere combined drought, combined with the SIF decline and soil water deficit, the stress level, diagnostic confidence and irrigation priority of each canopy spatial unit are calculated; the stress level, spatial distribution, diagnostic confidence and irrigation priority of each canopy spatial unit are output; and finally, standardized crop water stress characterization data are generated by combining crop physiological and soil and water meteorological parameters.

[0063] Understandably, the SIF data of each canopy spatial unit within the multi-temporal observation dataset is read to calculate the diurnal variation deviation of SIF between the first and second time periods. When observation data for the third time period is available, the SIF recovery between the second and third time periods is further calculated. The diurnal variation deviation of SIF is compared with a preset deviation threshold to identify spatial units with significantly decreased SIF. For the identified units, canopy temperature, air temperature, soil moisture content, and vapor pressure deficit parameters are retrieved. Based on multiple threshold combinations of canopy temperature difference, vapor pressure deficit, and soil moisture content, soil drought, atmospheric drought, and soil-atmosphere combined drought are distinguished, compensating for the shortcomings of traditional binary classification algorithms in identifying mixed stress boundary conditions. Then, the threshold of SIF recovery is used to distinguish between pseudo-stress signals caused by high light suppression and real water stress signals, completing the removal of non-water stress interference signals. Based on drought type, SIF decrease magnitude, and soil moisture deficit degree, the stress level, diagnostic confidence, and irrigation priority are quantitatively output. Combined with crop physiological and soil and water meteorological parameters, water stress characterization data are generated.

[0064] In another preferred embodiment based on the above embodiments, see [reference] Figure 2 As shown, this embodiment provides an environment-triggered UAV multi-temporal SIF collaborative observation and crop water stress identification system, including: an environment perception and window division module, a UAV aerial survey control module, a spectral processing and SIF inversion module, and a stress identification and prescription generation module.

[0065] The environmental perception and window segmentation module is configured to collect environmental parameters of the target farmland and physiological parameters of crop stomatal conductance and leaf water potential, and to divide multiple time-series observation windows by combining crop type and growth stage parameters, and to construct environmental trigger judgment rules. The UAV aerial survey control module is configured to control the UAV carrying the hyperspectral calibration payload to uniformly fly the flight trajectory and observe the parameters when the environmental trigger judgment rule takes effect, and to collect multi-temporal spectral data of the target area in the corresponding time-series observation window. The spectral processing and SIF inversion module is configured to correct and process the raw spectral data collected by the UAV, extract the canopy SIF and apparent fluorescence yield based on the characteristic bands and oxygen absorption valleys, perform atmospheric radiation correction in combination with the UAV's low-altitude atmospheric parameters, and simultaneously complete the spatiotemporal scale registration of multi-source data to construct the observation dataset. The stress identification and prescription generation module is configured to remove various non-water stress interference signals based on the observation dataset, distinguish the temporal characteristics of soil drought and atmospheric drought, integrate crop physiological and soil and water meteorological parameters to obtain water stress characterization data, output stress level, spatial distribution, diagnostic confidence and irrigation priority data, and perform data linkage with variable irrigation equipment.

[0066] It is understandable that the above-mentioned method and system for multi-temporal SIF collaborative observation of crops based on environment triggering by UAVs has the same beneficial effects, and will not be elaborated further here.

[0067] 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.

[0068] 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.

[0069] 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.

[0070] 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.

[0071] 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 method for multi-temporal SIF collaborative observation and crop water stress identification based on environment-triggered UAVs, characterized in that, include: Collect environmental parameters of the target farmland and physiological parameters of crop stomatal conductance and leaf water potential. Based on the physiological parameters and crop type and growth stage parameters, determine the sensitive period of the crop, select the corresponding observation window set, and based on the selected observation window set, divide multiple types of time-series observation windows into environmental parameter thresholds to construct environmental trigger judgment rules. When the environmental triggering judgment rule takes effect, the UAV carrying the hyperspectral calibration payload is controlled to collect multi-temporal spectral data of the target area in the corresponding time-series observation window with unified flight trajectory and observation parameters; After correcting and processing the raw spectral data collected by the UAV, the canopy SIF and apparent fluorescence yield are extracted based on the characteristic bands and oxygen absorption valleys. Atmospheric radiation correction is performed in combination with the UAV's low-altitude atmospheric parameters. Simultaneously, the spatiotemporal scale registration of multi-source data is completed to construct an observation dataset. Based on the observation dataset, various non-water stress interference signals are removed, the temporal characteristics of soil drought and atmospheric drought are distinguished, and crop physiological and soil and water meteorological parameters are integrated to obtain water stress characterization data. The stress level, spatial distribution, diagnostic confidence and irrigation priority data are output, and data linkage with variable irrigation equipment is carried out.

2. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment-triggered UAVs according to claim 1, characterized in that, The process of dividing time series observation windows into multiple categories includes: The environmental parameters, including photosynthetically active radiation, air temperature, relative humidity, water vapor pressure deficit, solar zenith angle, wind speed, cloud cover, and radiation stability parameters, are obtained. Based on the crop's stomatal conductance, leaf water potential physiological parameters, and the crop type and growth stage parameters, the water-sensitive and non-sensitive periods of the crop are determined. Based on the determined water-sensitive or non-sensitive periods, select the corresponding set of observation windows; Based on the selected set of observation windows, multiple types of time-series observation windows are divided in combination with environmental parameter thresholds; During the water-sensitive period, the time-series observation window is divided into a low-stress baseline observation window, a water-stress-sensitive observation window, and an afternoon recovery observation window. During the water-insensitive period, the time-series observation window is divided into a morning observation window and an afternoon observation window.

3. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by unmanned aerial vehicles according to claim 2, characterized in that, When dividing the time-series observation window into a low-stress baseline observation window, the following is included: Pre-set the minimum available threshold for photosynthetically active radiation, the low threshold for water vapor pressure deficit, and the maximum threshold for solar zenith angle; When the photosynthetically active radiation is greater than or equal to the minimum available threshold of photosynthetically active radiation, and the water vapor pressure deficit is less than the low threshold of water vapor pressure deficit, and the solar zenith angle is less than the maximum threshold of solar zenith angle, the current time period is determined to be the low stress benchmark observation window. When the photosynthetically active radiation is less than the minimum available threshold for photosynthetically active radiation, or the water vapor pressure deficit is greater than or equal to the low threshold for water vapor pressure deficit, or the solar zenith angle is greater than or equal to the maximum threshold for solar zenith angle, it is determined that the current time period does not meet the low stress benchmark observation window.

4. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by unmanned aerial vehicles according to claim 2, characterized in that, When dividing the time-series observation window into a water stress-sensitive observation window, the following is included: Pre-set the photosynthetically active radiation saturation threshold, water vapor pressure deficit stress threshold, cloud cover threshold, and radiation stability threshold; When the photosynthetically active radiation is greater than or equal to the photosynthetically active radiation saturation threshold, the water vapor pressure deficit is greater than or equal to the water vapor pressure deficit stress threshold, the cloud cover is less than the cloud cover threshold, and the radiation stability parameter is greater than the radiation stability threshold, the current time period is determined to be the water stress sensitive observation window. When the photosynthetically active radiation is less than the photosynthetically active radiation saturation threshold, or the water vapor pressure deficit is less than the water vapor pressure deficit stress threshold, or the cloud cover is greater than or equal to the cloud cover threshold, or the radiation stability parameter is less than or equal to the radiation stability threshold, it is determined that the current time period does not meet the water stress sensitive observation window.

5. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by unmanned aerial vehicles according to claim 4, characterized in that, When constructing environment trigger determination rules, the following are included: Obtain the first time period corresponding to the low-stress baseline observation window and the second time period corresponding to the water stress sensitive observation window; the first time period and the second time period are time periods within the same natural day, the first time period is earlier than the second time period, and the first time period and the second time period respectively meet the preset minimum effective duration; When both the first time period and the second time period are determined to be valid, an environment trigger command containing the flight mission of the first time period and the flight mission of the second time period is generated. When the time-series observation window includes an afternoon recovery observation window, the third time period corresponding to the afternoon recovery observation window is obtained, and the flight mission of the third time period is added to the environmental trigger command. When the first time period or the second time period is determined to be invalid, the environment trigger determination rule is determined to be ineffective, and the generation of the environment trigger instruction is terminated.

6. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by unmanned aerial vehicles according to claim 5, characterized in that, When controlling a UAV equipped with a hyperspectral calibration payload to unify flight trajectory and observation parameters, and collecting multi-temporal spectral data of the target area within the corresponding time-series observation window, the following are included: The flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy, and sensor observation angle of the UAV are preset. When the environmental trigger command is received, the UAV is controlled to perform the first spectral data acquisition in the first time period with the flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle. The UAV is controlled to perform a second spectral data acquisition in the second time period with the same flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle; When the environmental trigger command includes the third time period flight mission, the UAV is controlled to perform a third spectral data acquisition in the third time period with the same flight altitude, flight heading, flight trajectory, heading overlap, lateral overlap, exposure control strategy and sensor observation angle.

7. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by unmanned aerial vehicles according to claim 6, characterized in that, After correcting and processing the raw spectral data collected by the UAV, the extraction of canopy SIF and apparent fluorescence yield based on characteristic bands and oxygen absorption valleys includes: The original spectral data is subjected to dark current correction, non-uniformity correction and radiometric calibration to obtain the first corrected spectral data; The first corrected spectral data is subjected to atmospheric scattering correction and atmospheric absorption correction to obtain the second corrected spectral data, wherein the atmospheric absorption correction is based on the oxygen absorption band corresponding to the oxygen absorption valley. Based on the second corrected spectral data, the first spectral reflectance corresponding to the characteristic band and the second spectral reflectance corresponding to the oxygen absorption valley are obtained; The canopy SIF is calculated based on the first spectral reflectance and the second spectral reflectance; The apparent fluorescence yield is calculated based on the ratio of the canopy SIF to the reference canopy SIF. Wherein, the reference canopy SIF is the canopy SIF corresponding to the low-stress baseline observation window, the canopy SIF and the reference canopy SIF have the same dimensions, and the apparent fluorescence yield is dimensionless.

8. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by unmanned aerial vehicles according to claim 7, characterized in that, When simultaneously completing the spatiotemporal scale registration of multi-source data and constructing the observation dataset, the following steps are included: Acquire real-time differential positioning data and ground control point coordinates during the flight of the UAV; Based on the real-time differential positioning data and the coordinates of the ground control points, the first image corresponding to the first spectral data acquisition and the second image corresponding to the second spectral data acquisition are registered at the pixel level. If the environmental triggering command includes a third time period flight mission, the third image corresponding to the third spectral data acquisition is also registered at the pixel level. This ensures that the same pixel in each image participating in the registration corresponds to the same canopy spatial unit; A multi-temporal observation dataset is constructed based on the registered images, with the canopy spatial units as the index. The multi-temporal observation dataset includes the canopy SIF and apparent fluorescence yield of each canopy spatial unit in the first time period and the second time period, respectively. When a third image is available, the dataset also includes the canopy SIF and apparent fluorescence yield for each canopy spatial unit at the third time period.

9. The method for multi-temporal SIF collaborative observation and crop water stress identification based on environment triggering by unmanned aerial vehicles according to claim 8, characterized in that, When obtaining water stress characterization data by removing various non-water stress interference signals from the observation dataset, distinguishing the temporal characteristics of soil drought and atmospheric drought, and fusing crop physiological and soil and water meteorological parameters, the following are included: Obtain the diurnal variation deviation of SIF for each canopy spatial unit in the observation dataset between the first time period and the second time period. If the observation dataset includes data for the third time period, obtain the SIF recovery of the canopy spatial unit between the second time period and the third time period. Based on the daily variation deviation of SIF, it is determined whether there is a significant decrease in SIF in each canopy spatial unit; When the daily variation deviation of SIF is greater than the preset deviation threshold, it is determined that there is a significant decrease in SIF, and the canopy temperature, air temperature, soil moisture content and water vapor pressure deficit of the corresponding canopy space unit are specifically obtained. When the difference between the canopy temperature and the air temperature is greater than a preset canopy temperature difference threshold, and the water vapor pressure deficit is greater than a preset water vapor stress threshold, and the soil moisture content is less than a preset soil moisture threshold, the current drought type is determined to be soil drought. When the difference between the canopy temperature and the air temperature is less than or equal to the preset canopy temperature difference threshold, and the water vapor pressure deficit is greater than the preset water vapor stress threshold, and the soil moisture content is greater than or equal to the preset soil moisture threshold, the current drought type is determined to be atmospheric drought. When the crown-temperature difference is greater than the preset crown-temperature difference threshold, and the vapor pressure deficit is greater than the preset vapor stress threshold, and the soil moisture content is greater than or equal to the preset soil moisture threshold, or when the crown-temperature difference is greater than the preset crown-temperature difference threshold, the vapor pressure deficit is less than or equal to the preset vapor stress threshold, and the soil moisture content is less than the preset soil moisture threshold, the current drought type is determined to be soil-atmosphere combined drought. When there is SIF recovery and the SIF recovery is greater than a preset recovery threshold, it is determined that the significant decrease in SIF is due to highlight suppression rather than moisture stress. When the SIF recovery amount is less than or equal to the preset recovery threshold, and the daily SIF deviation is greater than the preset deviation threshold, the significant decrease in SIF is determined to be due to water stress. Based on the drought type identification results, the stress level, diagnostic confidence level, and irrigation priority are calculated by combining the SIF decrease rate and soil moisture deficit degree. The stress level, spatial distribution, diagnostic confidence level, and irrigation priority corresponding to each canopy spatial unit are output.

10. A UAV multi-temporal SIF collaborative observation and crop water stress identification system based on environment triggering, applied to the UAV multi-temporal SIF collaborative observation and crop water stress identification method based on environment triggering as described in any one of claims 1-9, characterized in that, include: The environmental perception and window segmentation module is configured to collect environmental parameters of the target farmland and physiological parameters of crop stomatal conductance and leaf water potential, and to divide multiple time-series observation windows by combining crop type and growth stage parameters, and to construct environmental trigger judgment rules. The UAV aerial survey control module is configured to control the UAV carrying the hyperspectral calibration payload to uniformly fly the flight trajectory and observe the parameters when the environmental trigger judgment rule takes effect, and to collect multi-temporal spectral data of the target area in the corresponding time-series observation window. The spectral processing and SIF inversion module is configured to correct and process the raw spectral data collected by the UAV, extract the canopy SIF and apparent fluorescence yield based on the characteristic bands and oxygen absorption valleys, perform atmospheric radiation correction in combination with the UAV's low-altitude atmospheric parameters, and simultaneously complete the spatiotemporal scale registration of multi-source data to construct the observation dataset. The stress identification and prescription generation module is configured to remove various non-water stress interference signals based on the observation dataset, distinguish the temporal characteristics of soil drought and atmospheric drought, integrate crop physiological and soil and water meteorological parameters to obtain water stress characterization data, output stress level, spatial distribution, diagnostic confidence and irrigation priority data, and perform data linkage with variable irrigation equipment.

Images