An agricultural irrigation water full-cycle supervision and prediction system based on multi-temporal remote sensing data

Abstract

The application relates to the technical field of agricultural water supervision and management, and discloses a full-cycle supervision and prediction system for agricultural irrigation water based on multi-temporal remote sensing data, which comprises a data acquisition module, a feature extraction module, a logic operation module, a correlation analysis module, a state recognition module and a cycle prediction module.The data acquisition module is used for acquiring upstream gate flow scheduling sequences and remote sensing image data containing geographical spatial attributes; the feature extraction module is used for extracting reflectivity response pulse sequences of target image elements from the remote sensing image data; the logic operation module is used for determining expected water delivery time delays; the correlation analysis module is used for determining measured response time delays; the state recognition module is used for recognizing irrigation compliance states based on the time deviation of the measured response time delays and the expected water delivery time delays; and the cycle prediction module is used for outputting water demand prediction instructions under the compliance states.The application establishes a physical time delay constraint mechanism between scheduling instructions and surface reflectivity responses, realizes causal auditing for water distribution business compliance, and ensures the authenticity of resource scheduling data and the accuracy of supervision instructions.

Description

Technical Field

[0001] This invention belongs to the field of agricultural water supervision and management technology, and in particular relates to a full-cycle monitoring and prediction system for agricultural irrigation water based on multi-temporal remote sensing data. Background Technology

[0002] Currently, irrigation monitoring in large irrigation areas mainly relies on satellite remote sensing sequences to obtain farmland water distribution characteristics, which serve as the basis for decision-making in water allocation and scheduling. Existing methods mostly select reflectance images from specific time phases to invert soil moisture content, aiming to achieve spatial positioning of irrigation demand and thus improve the integration of field monitoring. While existing technologies attempt to improve the end-side sensing effect by optimizing the deployment of sensing devices through hardware structure improvements—for example, patent CN205667196U discloses a connection device between an agricultural irrigation device and a remote sensor, using an H-shaped fixing component— The hook and support mechanical components integrate remote sensors onto the housing of irrigation pumps and motors to improve the hardware's field survival and portability. However, this regulatory logic is still limited to the idea of ​​endpoint mounting and fails to address the deep-seated contradictions involved in the control logic. Due to the lack of physical modeling of the dynamic response relationship between scheduling commands and surface feedback, the sensing data and water distribution flow are disconnected, making it impossible to explain the physical time delay generated by hydraulic transmission in the canal system. This hardware-led improvement approach also makes it difficult to establish a reliable causal audit mechanism when faced with observation sequence breakpoints or natural precipitation interference.

[0003] However, in actual water resource management operations, there is a physical transmission delay between the issuance of water dispatch instructions and the physiological response of farmland crops. Furthermore, due to limitations such as atmospheric scattering and cloud cover, satellite payloads cannot continuously provide uninterrupted observation sequences. This situation leads to observation blind spots in water evolution models at key growth nodes, resulting in cumulative prediction bias. At the same time, simply increasing the frequency of image revisits cannot establish a causal relationship between dispatch instructions and water changes. Consequently, it is difficult to identify the interference of non-mandatory water intake or natural precipitation on the monitoring results. In addition, the spatial interweaving of heterogeneous crops within irrigation areas leads to a homogenization trend in optical observation signals, causing mixed pixel features to mask the water demand pulses of specific species.

[0004] Therefore, the technical problem to be solved by this invention is how to construct a full-cycle analysis mechanism that couples hydraulic propagation constraints with crop physiological growth characteristics to solve the problem of causal tracing in the scheduling and supervision process and the problem of prediction stability under complex working conditions. Summary of the Invention

[0005] To address the problems mentioned in the background section, the technical solution of this invention is as follows:

[0006] A full-cycle monitoring and prediction system for agricultural irrigation water based on multi-temporal remote sensing data, the system comprising:

[0007] The data acquisition module is used to acquire the flow scheduling sequence of the upstream gates in the target irrigation area and to acquire multi-temporal remote sensing image data of the target irrigation area containing geospatial attributes.

[0008] The feature extraction module is used to extract the reflectance response pulse sequence of target pixels from multi-temporal remote sensing image data;

[0009] The logic operation module is used to determine the expected water delivery delay based on the preset channel water delivery path length between the target pixel and the upstream gate, as well as the preset water delivery velocity parameters.

[0010] The correlation analysis module is used to perform cross-correlation calculations between the traffic scheduling sequence and the reflectivity response pulse sequence to determine the measured response delay.

[0011] The state recognition module is used to calculate the absolute value of the deviation between the measured response delay and the expected water delivery delay, and execute the following judgment logic: if the absolute value of the deviation is lower than the preset fluctuation time threshold, the target pixel is determined to be in a compliant irrigation state; if the absolute value of the deviation exceeds the fluctuation time threshold, the target pixel is determined to be in a non-mandatory water intake state; if no effective peak value of the reflectivity response pulse sequence is extracted, the target pixel is determined to be in an observation missing compensation state.

[0012] The periodic prediction module is used to map the waveform characteristics of the reflectivity response pulse sequence to the real-time water content characteristics of the target pixel when it is determined that the target pixel is in a compliant irrigation state, and output the water demand prediction command.

[0013] Preferably, the system also includes an environmental compensation module, which is used to acquire temperature, humidity, total radiation and precipitation data of the target irrigation area, and construct a surface energy balance model accordingly; when observations are missing in multi-temporal remote sensing image data, the environmental compensation module uses the reflectance evolution trend generated by the surface energy balance model to perform data interpolation on the reflectance response pulse sequence to correct the calculation deviation of the water demand prediction command.

[0014] Preferably, the feature extraction module includes a pixel decomposition unit, which is used to extract the slope of the response pulse of the water-sensitive band at different phenological stages, and combine the differences in crop growth rhythm within the target pixel to process the reflectivity response pulse sequence through a second-order derivative algorithm, so as to separate the sub-response components of a specific crop from the mixed signal of the target pixel.

[0015] Preferably, the cell decomposition unit calculates the sub-response weight ω for a specific crop using the following formula: Where n is the number of phenological observation samples, R iLet θ be the reflectance intensity of the i-th observed sample, t be the observation time, and θ be the reflectance intensity of the i-th observed sample. i This is a preset growth rhythm correction coefficient for a specific crop.

[0016] Preferably, the correlation analysis module is also used to acquire canopy temperature rise rate data of the target irrigation area, and to use the canopy temperature rise rate data to correct the measured response time delay during the crop canopy closure period, so as to identify and eliminate non-irrigation water fluctuation characteristics in the reflectivity response pulse sequence.

[0017] Preferably, the periodic prediction module is also used to calculate the cumulative water surplus / deficit of the target pixel within the historical observation period based on the real-time water content characteristics, and to adjust the irrigation quota parameters in the water demand prediction instruction based on the cumulative water surplus / deficit.

[0018] Preferably, the data acquisition module includes a time synchronization unit, which is used to align the sampling time of the flow scheduling sequence with the acquisition time of the multi-temporal remote sensing image data, and to perform smoothing filtering on the flow scheduling sequence.

[0019] Preferably, the system also includes a compliance audit module, which records the deviation between the measured response delay and the expected water delivery delay when the target pixel is determined to be in a non-mandatory water extraction state, and generates an early warning work order for illegal water extraction in combination with geospatial attributes.

[0020] Preferably, the period prediction module is also used to extract the waveform attenuation rate of the reflectivity response pulse sequence, and convert the waveform attenuation rate into the deep soil permeation parameter of the target pixel according to the preset soil water holding parameters.

[0021] Preferably, the system also includes an output interface module for rendering compliant irrigation status, non-mandatory water intake status, and water demand prediction commands into raster layers with spatial attributes.

[0022] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0023] 1. In the full-cycle supervision of agricultural irrigation water use, by utilizing the physical time delay constraint between the scheduling command sequence and the surface reflectivity rebound pulse, the system establishes a coherence arbitration mechanism between hydraulic and physiological responses. By calculating the expected time delay of water flow propagation between the target area and the control gate, it achieves logical identification of compliant irrigation, precipitation interference, and non-mandatory water intake behavior. This shifts the supervision logic from simple surface moisture status monitoring to causal auditing of the compliance of water allocation operations, thereby ensuring the authenticity of resource scheduling data and the accuracy of supervision commands.

[0024] 2. The system integrates meteorological metadata such as atmospheric temperature and humidity, and solar radiation, and constructs a reflectivity attenuation factor based on the principle of energy balance. When optical observation data is missing due to weather factors, the system uses meteorological-driven virtual evolution characteristics to maintain the physical continuity of the moisture index curve, avoiding cumulative drift of the state prediction model at observation breakpoints. This enables the continuous operation of the regulatory logic under complex weather conditions and reduces the system's dependence on satellite payloads with high revisit cycles.

[0025] 3. By extracting the slope of the response pulse of the water-sensitive band at different phenological stages and combining it with the differences in growth rhythms of heterogeneous crops within the pixel, the system uses second-order differentiation and feature projection techniques to separate sub-response components for specific crops from low-resolution mixed pixel signals. This solves the problem of homogenization of water demand signals in fragmented farmland scenarios and improves the monitoring system's ability to finely identify the water demand characteristics of specific species without increasing the hardware observation resolution. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the connection relationship and logical control flow of the various functional modules of the system of the present invention;

[0027] Figure 2 This is a diagram illustrating the data interaction architecture and business application scenarios of the core platform for full-cycle regulatory prediction in this invention. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0029] This invention provides a full-cycle monitoring and prediction system for agricultural irrigation water based on multi-temporal remote sensing data. The system includes a data acquisition module, a feature extraction module, a logic operation module, a correlation analysis module, a state recognition module, and a cycle prediction module. The data acquisition module acquires the flow scheduling sequence G(t) of the upstream gates within the target irrigation area and obtains multi-temporal remote sensing image data of the target irrigation area containing geospatial attributes. The feature extraction module extracts the reflectance response pulse sequence R of the target pixels from the multi-temporal remote sensing image data. ' The logic operation module is used to determine the expected water delivery delay τ=L / v based on the preset channel water delivery path length L between the target pixel and the upstream gate, and the preset water delivery velocity parameter v. The correlation analysis module is used to connect the flow scheduling sequence G(t) with the reflectivity response pulse sequence R. ' (t) Perform cross-correlation calculations to determine the measured response delay ΔT. actThe state recognition module is used to calculate the measured response delay ΔT. act The absolute value of the deviation between the expected water delivery delay τ and |ΔT act -τ|, and compare it with the preset fluctuation time threshold ϵ to identify compliant irrigation status, non-mandatory water intake status, and observation gap compensation status. The periodic prediction module is used to predict the irrigation status under compliant status based on real-time moisture content characteristics and cumulative water surplus / deficit W. acc The system outputs water demand prediction commands. Specifically, the periodic prediction module establishes a nonlinear functional relationship between the peak rebound intensity of the reflectivity response pulse sequence R'(t) and the apparent thermal inertia (ATI) of the land surface, thus mapping waveform characteristics to water content. After determining compliant irrigation, the system extracts the slope k of the reflectivity rebound from the waveform trough to the steady state, based on the preset spectral-moisture calibration curve: θ act =f(k,R max The real-time moisture content characteristic θ is calculated. act This mapping process incorporates the growth rhythm correction coefficient θ of the phenological stage. i To remove the shading effect of crop canopy cover on soil spectral signals and ensure the physical consistency of mapping results, a physical time delay constraint mechanism is established between scheduling command pulses and reflectivity response pulses to achieve causal auditing of water allocation compliance, ensuring the authenticity of resource scheduling data and the accuracy of regulatory instructions.

[0030] To address the issue of unclear causal relationships between water conservancy dispatch instructions and surface physiological responses due to physical transmission delays in the supervision of large-scale irrigation areas, the data acquisition module acquires the flow dispatch sequence G(t), records the timestamps of gate opening and closing, and instantaneous flow values. Simultaneously, it acquires multi-temporal remote sensing image data of the target irrigation area, using vegetation index sequences acquired by satellite payloads as the data source. The time-series synchronization unit in the data acquisition module aligns the sampling time of the flow dispatch sequence G(t) with the acquisition time of the remote sensing image data, establishing a linear interpolation buffer with a step size of 600 seconds. For satellite image observations with a revisit period of 5 to 7 days, the spectral reflectance difference between two adjacent observation times is extracted. This is achieved by establishing a coarse adjustment interval of 600 seconds and an evolution step size of 60 seconds. The reflectivity pseudo-evolutionary incremental model uses a linear interpolation algorithm to reconstruct low-frequency remote sensing observations into a high-frequency fitted sequence with a sampling frequency of 1 Hz. This ensures precise alignment with the flow scheduling sequence with a sampling frequency of 1 Hz within the time window during cross-correlation calculations. A smoothing filter is applied to the flow scheduling sequence G(t) to reduce random noise during sensor acquisition. The logic operation module determines the expected water delivery delay τ according to the formula τ = L / v, where τ is the expected water delivery delay in seconds, L is the channel water delivery path length in meters, and v is the water delivery velocity parameter in m / s. The correlation analysis module calculates the correlation coefficient between the two sequences at different time displacements and selects the displacement corresponding to the maximum correlation coefficient as the measured response delay ΔT. act The state recognition module executes the judgment logic. If the absolute value of the deviation is lower than the preset fluctuation time threshold ϵ, the target pixel is determined to be in a compliant irrigation state. If the measured response delay ΔT act If the absolute value of the deviation from the expected water delivery delay τ exceeds the fluctuation time threshold ϵ, the target pixel is determined to be in a non-mandatory water intake state.

[0031] The environmental compensation module acquires temperature, humidity, total radiation, and precipitation data for the target irrigation area, constructs a surface energy balance model, and generates a reflectance attenuation weighting coefficient K by utilizing the mapping relationship between the reference crop evapotranspiration ET0 calculated based on the Penman-Monteith principle and the historical reflectance attenuation gradient when remote sensing image data has observational gaps. dec The system performs data interpolation on the reflectance response pulse sequence R'(t). Specifically, it uses the Penman-Monteith principle to calculate the daily reference crop evapotranspiration ET0 during the phenological period, and establishes a functional mapping between ET0 and the reflectance decrease gradient under no-irrigation conditions through a linear regression model, thereby determining K. dec The system uses the measured value of the last time phase before the interruption as the initial value, and the reflectivity attenuation weighting coefficient K is used to determine the dynamic value. dec Virtual water evolution features are generated iteratively along the time axis. The pixel decomposition units in the feature extraction module are obtained through formulas. Calculate the sub-response weight ω for a specific crop, where ω is the sub-response weight for the specific crop, n is the number of phenological observation samples, and R0 is the weight of the phenological observation samples. i Let θ be the reflectance intensity of the i-th sample, t be the observation time, and θ be the reflectance intensity of the i-th sample. i To obtain the preset growth rhythm correction coefficient for a specific crop, the cycle prediction module calculates the water demand command based on the water balance equation and obtains the real-time moisture content characteristic θ. act And retrieve the lower limit of suitable moisture content θ for crops. low The periodic prediction module determines the cumulative water surplus / deficit W by calculating the difference between the cumulative evapotranspiration and effective precipitation of the target pixel within the historical observation period. acc When the real-time moisture content characteristic θ act Below the lower limit of suitable moisture content for crops θ low At that time, the system will calculate the preset standard irrigation quota and the accumulated water surplus / deficit W. acc Perform the summation operation and output the water demand prediction instruction, where W acc The cumulative moisture gain or loss is expressed in mm, θ act For real-time moisture content characteristics, θ low This is the lower limit of the suitable moisture content for crops; this embodiment involves the pre-calibration of hydraulic parameters during the initial operation of the system. The system selects a benchmark canal section in the target irrigation area and performs a controlled water distribution test during a period without natural precipitation, recording the opening time T of the upstream gate. start Simultaneously using a surface apparent thermal inertial sensor to monitor the time T when moisture reaches the target pixel. end Determine the measured transmission time ΔT test =T end -T start The water conveyance velocity parameter v is calculated based on the channel water conveyance path length L: v = L / (ΔT) test The system continuously executes five sets of controlled water distribution tests and statistically analyzes ΔT. test The standard deviation of the distribution is σ, and the fluctuation time threshold ϵ is set to 3σ. The calibration process establishes the physical benchmark required for the system to determine compliance, and eliminates the static calculation deviation introduced by the differences in canal roughness and geological seepage.

[0032] Example 1: In a large-scale irrigation area containing a three-level canal system, the upstream gate receives gate scheduling instructions and generates a flow scheduling sequence G(t). Since the water distribution target pixel is located at the end of the canal, the length L of the canal water conveyance path between the target pixel and the upstream gate is 8400m, and the water conveyance velocity parameter v is set to 0.7m / s. The logic operation module calculates the expected water conveyance delay τ as 12000s according to the formula τ=L / v, where τ is the expected water conveyance delay in seconds, L is the length of the canal water conveyance path in meters, and v is the water conveyance velocity parameter in meters per second. When the system faces short-term heavy rainfall interference caused by local meteorological factors, the surface reflectance index exhibits non-instructional fluctuations. The monitoring method based on a single time phase cannot separate environmental noise from irrigation behavior.

[0033] The feature extraction module extracts the reflectance response pulse sequence R from multi-temporal remote sensing image data containing geospatial attributes acquired by the satellite payload. ' The correlation analysis module compares the flow scheduling sequence G(t) with the reflectivity response pulse sequence R(t). ' (t) Perform cross-correlation calculations to determine the measured response delay ΔT act When the waveform transitions in the flow scheduling sequence G(t) are related to the reflectivity response pulse sequence R... ' When the spectral rebound phases in (t) are physically aligned on the time axis, the cross-correlation function generates a peak. The measured response delay ΔT is extracted by detecting the position of this peak. act The time interval is 12180s. To eliminate multiple interference peaks caused by natural precipitation or illegal pumping, the system searches only for extreme points within ±20% of the expected water delivery delay in the discrete correlation coefficient sequence generated by the cross-correlation function. If multiple peaks with correlation coefficients exceeding 0.75 appear within this predetermined time window, the system extracts the peak with the steepest slope and duration closest to the water distribution command duration as the main peak of the measured response delay, thereby eliminating random moisture fluctuation signals without physical causal correlation. The state recognition module then determines the response delay based on the measured response delay ΔT. act The absolute value of the deviation from the calculated expected water delivery delay τ |ΔT act -τ| is 180s. Since this deviation value is lower than the preset fluctuation time threshold ϵ, i.e., 600s, the system judges the spectral rebound event as controlled and compliant irrigation and uses it as input to the water evolution model. This excludes random precipitation signals that do not have physical coherence. The periodic prediction module outputs water demand prediction instructions, realizing causal auditing of irrigation facts under complex meteorological conditions.

[0034] Example 2: This section provides an experimental verification procedure for an agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data. In a heterogeneous crop planting experimental area with a total area of ​​333.3 hectares, the data acquisition module retrieves the flow scheduling sequence G(t) of the No. 3 diversion gate from the flow monitoring terminal, simultaneously acquiring temperature, humidity, and total radiation parameters collected by the automatic weather station. The sampling frequency is set to 1Hz. The feature extraction module acquires satellite remote sensing images with a spatial resolution of 10m and actively performs pixel masking on consecutive temporal phases in the time-series processing chain to simulate observational gaps. The logic operation module determines the expected water delivery delay τ as 5561.6s based on the measured channel water conveyance path length L of 4560.5m and the water flow velocity parameter v of 0.82m / s. The correlation analysis module analyzes the reflectivity response pulse sequence R with 15% noise superimposed. ' (t) is smoothed, and the measured response delay ΔT is determined by extracting the envelope peak value of the cross-correlation function. act For specific data, please refer to Table 1.

[0035] Table 1: Record of Response Characteristics of Water Injection Events

[0036] According to the data shown in Table 1, the absolute value of the deviation in sample group A of this invention is 120.8s, which is lower than the fluctuation time threshold ϵ. The system identifies this event as effective irrigation driven by the scheduling command. However, in control group B, the occurrence phase of natural precipitation is much earlier than the gate action, and the absolute value of the deviation far exceeds the threshold. Based on this, the status recognition module generates an abnormal early warning work order. Sample group A uses the environmental compensation module to calculate the reflectivity attenuation weighting coefficient K. dec After performing data interpolation, the phase deviation between the virtual evolutionary characteristics and the regressed measured values ​​is 7.2%. The period prediction module predicts the period based on the waveform attenuation rate D. rate The calculated deep soil permeability parameter is 1.15 mm / d, based on which the irrigation quota parameter in the water demand forecasting instruction is reduced by 4.8%.

[0037] Example 3: This example combines Figures 1 to 2 An explanation of the agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data, such as... Figure 1As shown, the data acquisition module acquires the upstream gate flow sequence and remote sensing images containing geospatial attributes, transmits the flow scheduling sequence to the logic operation module, and simultaneously transmits multi-temporal remote sensing images to the feature extraction module. The logic operation module calculates the expected water delivery delay based on the channel water conveyance path length and flow velocity, and transmits it to the state recognition module. The feature extraction module is responsible for extracting the surface reflectance response pulse sequence of the target pixel. This sequence, together with the flow scheduling sequence (dashed path) directly from the data acquisition module, is collected by the correlation analysis module. The correlation analysis module uses the flow scheduling sequence and the reflectance response pulse sequence to perform cross-correlation to determine the measured response delay. The state recognition module receives the expected water delivery delay and the measured response delay and compares the deviation between the measured delay and the expected delay to identify compliant irrigation or non-mandatory water intake status. If the system determines that it is a compliant irrigation status, it activates the cycle prediction module to map the real-time water content characteristics and outputs a water demand prediction command.

[0038] like Figure 2 As shown, the full-cycle monitoring and forecasting core computing platform is located at the central hub. Its input end receives image data streams generated by spatial attribute acquisition from multi-temporal remote sensing satellites, scheduling command streams generated by flow scheduling execution from upstream water distribution gates, and environmental parameter streams involving energy balance parameters provided by surface environmental monitoring stations. The full-cycle monitoring and forecasting core computing platform integrates time-series synchronization, causal auditing, missing data compensation, and water demand forecasting functional units. On the output end, it sends the compliance status visualization results to the monitoring dashboard through raster layer rendering, and transmits the violation warning information to the mobile operation and maintenance terminal through work order push.

[0039] Example 4: In a fragmented cultivated area scenario involving interwoven winter wheat and weed communities, the data acquisition module obtains a discrete flow scheduling sequence G[k] and a reflectivity response pulse sequence R. ' [k] To address the phase blurring challenge caused by non-uniform sampling of remote sensing images, the correlation analysis module performs cross-correlation calculations based on a discrete sliding window. The system calculates the cross-correlation based on the discrete sampling step size T. s Calculate the correlation components within a sliding window of length N. Where C[m] is the cross-correlation function value when the displacement index is m, G[k] is the flow scheduling amplitude of the sampling step size k, and R ' [km] represents the amplitude of the reflectivity response after displacement, N is the total number of samples involved in the calculation, and the system searches for the displacement index m corresponding to the maximum value of the relevant component C[m]. max And according to the formula ΔT act =m max ⋅T s Extract the measured response delay ΔT act , where ΔT actThe measured response delay is expressed in seconds (s) and meters (m). max T is the displacement index corresponding to the cross-correlation maxima. s The discrete sampling step size is measured in seconds. This algorithm path achieves accurate reconstruction of physical time consumption through phase alignment of discrete signals.

[0040] The standard phenological curve is obtained by decomposing the pixel into units and the reciprocal of the absolute value of the second derivative of the physiological index over time is calculated to determine the growth rhythm correction coefficient θ for different phenological stages. i The sub-response weight ω is then adjusted to separate the water demand pulse of a specific species from the mixed signal. The periodic prediction module determines the waveform attenuation rate D by extracting the logarithmic domain descent slope of the reflectivity response pulse sequence from the rebound peak point to the regression steady-state point. rate Specifically, the waveform attenuation rate D rate The calculation process is as follows: The periodic prediction module locks the reflectivity response pulse sequence R'(t) at the peak point (t) after irrigation triggering. peak ,R peak '), and the steady-state point (t) after water transport reaches dynamic equilibrium. steady ,R steady By performing a natural logarithmic transformation on the reflectivity amplitude, a linear decay regression model on the time axis is established, D rate This is the absolute value of the slope of the logarithmic field curve in the decay interval, and its calculation formula is: This attenuation rate eliminates absolute amplitude interference and purely reflects the dynamic rate of vertical infiltration and evaporation of surface water in the soil profile. A permeability mapping model P is also introduced. loss =Φ⋅exp(D rate / S cap Output deep soil permeability parameters, where P loss This refers to deep soil permeability parameters, expressed in mm / d, D. rate S represents the waveform attenuation rate. cap The preset soil water holding parameter refers to the physical constant of soil field water holding capacity in the area where the target pixel is located, reflecting the resistance characteristics of soil pores to water. Φ is a physical calibration constant with a value of 1.02. The environmental compensation module executes an iterative interpolation procedure based on feedback residual correction. When the optical payload produces continuous observation gaps, the system uses a weighted coefficient K based on reflectivity attenuation. dec By performing convolution operations to generate virtual evolution trajectories and performing dynamic recalibration based on the phase residuals after satellite image regression, a closed-loop regulatory logic for the entire lifecycle under sensor failure conditions is realized.

[0041] Example 5: In the scenario of monitoring and deploying a newly built irrigation district, the system executes a preliminary hydraulic parameter calibration process to determine the water conveyance velocity parameter v and the fluctuation time threshold ϵ. During the baseline period, technicians perform controlled experimental water distribution at the upstream gate and record the gate opening time T. start Simultaneously, the ground-based telemetry and control terminal is used to extract the water arrival signal and obtain the measured transmission time ΔT of the water flow reaching the target pixel. test The system calculates the water conveyance path length L based on the formula v=L / ΔT. test Determine the water conveyance velocity parameter v for this canal section at a specific water level, where v is the water conveyance velocity parameter in m / s, L is the canal water conveyance path length in m, and ΔT test The actual transmission time is measured in seconds. At the same time, the system statistically analyzes the distribution dispersion of transmission time through multiple consecutive pre-distribution water tests. Three times the standard deviation of the observed deviation component is set as the fluctuation time threshold ϵ. This procedure establishes the initial calibration benchmark for system operation through controlled experiments.

[0042] When the system faces a scenario involving the reconstruction of a water evolution model during periods of observational gaps, the environmental compensation module applies the reflectivity attenuation weighting coefficient K. dec The mapping process involves retrieving historical meteorological sequences within the phenological cycle and synchronous multi-temporal remote sensing image data. An algorithm is used to calculate the daily reference crop evapotranspiration (ET0), and the reflectance response pulse sequence R under conditions of no water input is simultaneously extracted. ' The descent gradient of [k] is used to establish a mapping function between the reference crop evapotranspiration ET0 and the reflectance attenuation through a linear regression model, thereby determining the reflectance attenuation weighting coefficient K. dec The initial parameters, where K dec R is the reflectance attenuation weighting coefficient, ET0 is the reference crop evapotranspiration in mm / d. ' [k] represents the reflectance response pulse sequence. After this mapping relationship is established, when an observation breakpoint occurs in the optical remote sensing data, the system uses this mapping function to drive the surface energy balance model to calculate the real-time reflectance attenuation weighting coefficient K. dec By applying it to the time-series evolution model, adaptive recalibration of the regulatory logic for different planting structures is achieved.

[0043] Example 6: In a large irrigation district deployment scenario with multi-level branch canal systems, the system executes a standardized topology calibration procedure to determine the channel water conveyance path length L between the target pixel and the upstream gate. The data acquisition module retrieves the geographic database of the irrigation district vector network and uses the spatial network analysis unit to search for the hydraulic connectivity path from the gate node to the geometric centroid of the target pixel. The geometric cumulative length of this path in the projected coordinate system is set as the channel water conveyance path length L. The logic operation module executes an adaptive procedure for velocity parameters based on the dynamic characteristics of water level. The system monitors the instantaneous water level height H of the upstream gate in real time at the canal head and retrieves the preset water level and velocity lookup table to determine the current water conveyance velocity parameter v. The water level and velocity lookup table is established by selecting multiple characteristic water points during the pre-debugging period and performing polynomial fitting, thus eliminating the hydraulic conduction calculation deviation caused by changes in the channel cross-section.

[0044] When the system faces phase offset sampling conditions from heterogeneous sensors, the correlation analysis module executes a time axis alignment procedure based on characteristic rising edges. The system detects the flow transition point triggered by the first water allocation command in the flow scheduling sequence G[k] and uses this as the global sampling reference time. It calculates the discrete phase deviation between the acquisition time of multi-temporal remote sensing image data and the global sampling reference time. The correlation analysis module uses a linear interpolation algorithm to perform a linear interpolation on the reflectivity response pulse sequence R. ' [k] is resampled, and its sampling step size is aligned to the sampling step size of the traffic scheduling sequence G[k], where G[k] is the traffic scheduling sequence in m. 3 / s, R ' [k] represents the reflectivity response pulse sequence. By executing this phase alignment procedure, the state recognition module calculates the measured response delay ΔT. act The random fluctuation component introduced by the system clock asynchrony was removed, reducing the measured deviation of irrigation compliance judgment to within 120.0s, and outputting compliance judgment instructions.

[0045] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. A full-cycle monitoring and forecasting system for agricultural irrigation water based on multi-temporal remote sensing data, characterized in that the system... include: The data acquisition module is used to acquire the flow scheduling sequence of the upstream gates in the target irrigation area and to acquire multi-temporal remote sensing image data of the target irrigation area containing geospatial attributes. The feature extraction module is used to extract the reflectance response pulse sequence of target pixels from multi-temporal remote sensing image data; The logic operation module is used to determine the expected water delivery delay based on the preset channel water delivery path length between the target pixel and the upstream gate, as well as the preset water delivery velocity parameters. The correlation analysis module is used to perform cross-correlation calculations between the traffic scheduling sequence and the reflectivity response pulse sequence to determine the measured response delay. The state recognition module is used to calculate the absolute value of the deviation between the measured response delay and the expected water delivery delay, and execute the following judgment logic: if the absolute value of the deviation is lower than the preset fluctuation time threshold, the target pixel is determined to be in a compliant irrigation state; if the absolute value of the deviation exceeds the fluctuation time threshold, the target pixel is determined to be in a non-mandatory water intake state; if no effective peak value of the reflectivity response pulse sequence is extracted, the target pixel is determined to be in an observation missing compensation state. The periodic prediction module is used to map the waveform characteristics of the reflectivity response pulse sequence to the real-time water content characteristics of the target pixel when it is determined that the target pixel is in a compliant irrigation state, and output the water demand prediction command.

2. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The system also includes an environmental compensation module, which is used to acquire temperature, humidity, total radiation and precipitation data of the target irrigation area and construct a surface energy balance model accordingly. When observations are missing in multi-temporal remote sensing image data, the environmental compensation module uses the reflectance evolution trend generated by the surface energy balance model to perform data interpolation on the reflectance response pulse sequence to correct the calculation deviation of the water demand prediction command.

3. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The feature extraction module includes a pixel decomposition unit, which is used to extract the slope of the response pulse of the water-sensitive band at different phenological stages. Combined with the differences in crop growth rhythm within the target pixel, the reflectivity response pulse sequence is processed by a second-order derivative algorithm to separate the sub-response components of a specific crop from the mixed signal of the target pixel.

4. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 3, characterized in that, The pixel decomposition unit calculates the sub-response weight ω for a specific crop using the following formula: Where n is the number of phenological observation samples, R i Let θ be the reflectance intensity of the i-th observed sample, t be the observation time, and θ be the reflectance intensity of the i-th observed sample. i This is a preset growth rhythm correction coefficient for a specific crop.

5. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The correlation analysis module is also used to acquire canopy temperature rise rate data of the target irrigation area, and to use the canopy temperature rise rate data to correct the measured response delay during the crop canopy closure period, so as to identify and eliminate non-irrigation water fluctuation characteristics in the reflectivity response pulse sequence.

6. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The periodic prediction module is also used to calculate the cumulative water surplus / deficit of target pixels within the historical observation period based on real-time water content characteristics, and to adjust the irrigation quota parameters in the water demand prediction instruction based on the cumulative water surplus / deficit.

7. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The data acquisition module includes a timing synchronization unit, which is used to align the sampling time of the flow scheduling sequence with the acquisition time of the multi-temporal remote sensing image data, and to perform smoothing filtering on the flow scheduling sequence.

8. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The system also includes a compliance audit module, which records the deviation between the measured response delay and the expected water delivery delay when the target cell is determined to be in a non-mandatory water intake state, and generates an early warning work order for illegal water intake based on geospatial attributes.

9. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The periodic prediction module is also used to extract the waveform attenuation rate of the reflectivity response pulse sequence and convert the waveform attenuation rate into the deep soil permeation parameter of the target pixel according to the preset soil water holding parameters.

10. The agricultural irrigation water full-cycle monitoring and prediction system based on multi-temporal remote sensing data according to claim 1, characterized in that, The system also includes an output interface module, which renders compliant irrigation status, non-mandatory water intake status, and water demand prediction commands into raster layers with spatial attributes.

Images