Soil water seepage characteristic dynamic monitoring system and method based on multi-sensor fusion

Abstract

The invention discloses a dynamic monitoring system and method for soil water seepage characteristics based on multi-sensor fusion, and the system comprises a data collection module which collects the original data of soil moisture content, matrix potential and temperature through a plurality of integrated sensing units; the data preprocessing module is used for carrying out time synchronization and anomaly correction on the original data and generating water content, matrix potential and temperature data streams; the dynamic collaborative inversion module is used for resolving a soil hydraulic characteristic parameter group of each unit position in real time based on a space-time gradient change relation of a soil hydrodynamic model coupling data flow; and the space-time adaptive fusion module is used for carrying out space interpolation and time deduction by introducing a fusion algorithm taking a soil hydrodynamic model as a covariable according to the parameter groups and the temperature data streams of all the units, and generating a saturated permeability coefficient field and an unsaturated hydraulic conductivity parameter field of continuous space-time evolution. The spatial-temporal resolution and the real-time sensing capability of dynamic evolution of the soil moisture field are improved, and the requirements of precision agriculture and disaster early warning are met.

Description

Technical Field

[0001] This invention relates to the field of dynamic monitoring technology of soil infiltration characteristics, and in particular to a dynamic monitoring system and method for soil infiltration characteristics based on multi-sensor fusion. Background Technology

[0002] Research on soil moisture movement patterns is a crucial foundation for water-saving irrigation in agriculture, water resource management, and geological disaster early warning. Traditional soil infiltration characteristic monitoring mainly relies on single-type sensors or manual sampling and laboratory analysis. It uses point-based measurement equipment such as time-domain reflectometers and tensiometers to obtain parameters like soil moisture content and matrix potential, and then extrapolates the permeability coefficient using empirical models. These methods suffer from limitations in terms of single monitoring dimensions and insufficient spatial representativeness, making it difficult to accurately characterize the complex processes of moisture transport in heterogeneous soils. Existing technologies largely focus on static or quasi-static monitoring, with low sampling frequencies and significant response delays, failing to capture the rapid dynamic evolution of the soil moisture field under rainfall infiltration and irrigation effects, thus failing to meet the urgent real-time requirements of precision agriculture and disaster early warning. Therefore, there is an urgent need to develop multi-sensor collaborative monitoring and data fusion technologies to achieve high spatiotemporal resolution dynamic perception of soil infiltration characteristics.

[0003] Therefore, there is an urgent need to provide a technical solution to address the above problems. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a dynamic monitoring system and method for soil infiltration characteristics based on multi-sensor fusion.

[0005] In a first aspect, the present invention provides a dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion, the technical solution of which is as follows: The data acquisition module is used to simultaneously collect raw data on soil moisture content, soil matrix potential, and soil temperature at the location of each integrated sensing unit by multiple integrated sensing units deployed in the monitoring area. The data preprocessing module is used to perform time synchronization processing and outlier correction based on physical thresholds on the raw data of each integrated sensing unit, and generate spatiotemporally aligned water content data stream, matrix potential data stream and temperature data stream corresponding to each integrated sensing unit. The dynamic collaborative inversion module is used to couple the spatiotemporal gradient change relationship between the water content data stream and the matrix potential data stream corresponding to each integrated sensing unit based on the soil hydrodynamic model, and iteratively calculate the soil hydraulic characteristic parameter set of each integrated sensing unit in real time at a set frequency. The soil hydraulic characteristic parameter set includes core model parameters used to determine the saturated permeability coefficient and unsaturated hydraulic conductivity. The spatiotemporal adaptive fusion module is used to perform spatial interpolation and temporal extrapolation on the soil hydraulic characteristic parameter set of all integrated sensing unit locations and the corresponding temperature data stream by introducing the soil hydrodynamic model as a spatiotemporal covariate in the adaptive fusion algorithm. This process generates a saturated permeability coefficient field and an unsaturated hydraulic conductivity parameter field that cover the monitoring area and evolve continuously in spatiotemporal.

[0006] Furthermore, the multiple integrated sensing units are arranged in the monitoring area according to a non-uniform grid topology.

[0007] Furthermore, the non-uniform grid topology is adaptively encrypted based on the historical soil quality map or topographic elevation map of the monitoring area.

[0008] Furthermore, the outlier correction based on physical thresholds performed by the data preprocessing module specifically includes: setting an upper threshold for the moisture content data stream based on the soil saturated moisture content, setting an upper and lower threshold for the matrix potential data stream based on the range of the soil water potential sensor, and replacing data points exceeding the corresponding thresholds with data generated based on time series linear interpolation.

[0009] Furthermore, the dynamic collaborative inversion module is based on the Richards equation as the soil hydrodynamic model, and the dynamic collaborative inversion module uses an ensemble Kalman filter algorithm to assimilate the water content data stream and the matrix potential data stream, and performs the real-time iterative solution.

[0010] Furthermore, the state vector of the ensemble Kalman filter algorithm includes the soil hydraulic characteristic parameter set, and the observation vector of the ensemble Kalman filter algorithm is composed of the water content data stream and the matrix potential data stream.

[0011] Furthermore, when the dynamic collaborative inversion module solves the soil hydraulic characteristic parameter set, the objective function F(θ) is defined as the weighted sum of squares of the difference between the observed data and the model prediction, and a spatiotemporal smoothing constraint term of the soil hydraulic characteristic parameter set is introduced. The expression for the objective function F(θ) is:

[0012] Where θ represents the set of soil hydraulic characteristic parameters to be solved, T represents the total number of time steps, and N represents the total number of integrated sensing units. and Let represent the matrix potential and volumetric water content observed by the i-th integrated sensing unit at time t, respectively. and These represent the corresponding values ​​obtained from simulations based on the soil hydrodynamic model and parameter θ. and λ represents the weighting coefficients of the matrix potential and water content observations, respectively. R(θ) represents the spatiotemporal smoothing constraint function for the parameter set θ, and λ is the weighting coefficient of the constraint term.

[0013] Furthermore, in the spatiotemporal adaptive fusion module, the soil hydrodynamic model as a spatiotemporal covariate is specifically manifested as follows: the spatiotemporal gradient field of soil moisture predicted by the soil hydrodynamic model is used as the input of the variogram model in the adaptive fusion algorithm.

[0014] Furthermore, when the spatiotemporal adaptive fusion module performs spatial interpolation and temporal extrapolation, the adaptive fusion algorithm used is the physically constrained spatiotemporal kriging algorithm; the physically constrained spatiotemporal kriging algorithm predicts the soil hydraulic characteristic parameters at time t0 for the location s0 without sensor units. This is obtained by solving the following Kriging equations:

[0015] Where K represents the total number of neighboring sensor unit observations used for prediction. Indicates position At any moment The observed value of a certain parameter in the set of soil hydraulic characteristic parameters. For corresponding Kriging weight coefficients, This indicates that the soil hydrodynamic model is located at... and time For the same predicted parameter value, β is the physical constraint coefficient corresponding to the model's predicted value. Indicates the observed parameters at the spatiotemporal point and Covariance between This indicates the relationship between model predictions and observed parameters at spatiotemporal points. and The covariance between them.

[0016] Secondly, the present invention provides a method for dynamic monitoring of soil infiltration characteristics based on multi-sensor fusion, the technical solution of which is as follows: The system uses multiple integrated sensing units deployed in the monitoring area to simultaneously collect raw data on soil moisture content, soil matrix potential, and soil temperature at the locations of each integrated sensing unit. The raw data of each integrated sensing unit is time-synchronized and outlier correction is performed based on physical thresholds to generate spatiotemporally aligned water content data stream, matrix potential data stream and temperature data stream corresponding to each integrated sensing unit. Based on the soil hydrodynamic model, the spatiotemporal gradient change relationship between the water content data stream and the matrix potential data stream corresponding to each integrated sensing unit is coupled, and the soil hydraulic characteristic parameter set of each integrated sensing unit is iteratively calculated in real time at a set frequency. The soil hydraulic characteristic parameter set includes core model parameters used to determine the saturated permeability coefficient and unsaturated hydraulic conductivity. Based on the soil hydraulic characteristic parameter set and corresponding temperature data stream at the locations of all integrated sensing units, an adaptive fusion algorithm is used to introduce the soil hydrodynamic model as a spatiotemporal covariate to perform spatial interpolation and temporal extrapolation on the soil hydraulic characteristic parameter set at the locations of all integrated sensing units. This results in the fusion of a continuously spatiotemporally evolving saturated permeability coefficient field and unsaturated hydraulic conductivity parameter field covering the monitoring area.

[0017] The technical solution of this invention solves the problems of single dimension, insufficient spatial representativeness and response delay of traditional point measurement by using a multi-sensor collaborative monitoring architecture and real-time data fusion technology. It improves the spatiotemporal resolution and real-time perception capability of soil moisture field dynamic evolution, and meets the application needs of precision agriculture and disaster early warning.

[0018] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

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

[0020] 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 This is a schematic diagram of an embodiment of a dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to the present invention. Figure 2 This is a flowchart illustrating an embodiment of a method for dynamic monitoring of soil infiltration characteristics based on multi-sensor fusion according to the present invention. Detailed Implementation

[0021] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0022] Figure 1 A schematic diagram of an embodiment of a dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion provided by the present invention is shown. Figure 1 As shown, the system includes: The data acquisition module 110 is used to simultaneously collect raw data on soil moisture content, soil matrix potential and soil temperature at the location of each integrated sensing unit by means of multiple integrated sensing units deployed in the monitoring area.

[0023] The monitoring area refers to the spatial range within which a dynamic monitoring system for soil infiltration characteristics is deployed and data is collected; for example, a 50-hectare cornfield in a water-saving agricultural experimental zone in the Huang-Huai-Hai Plain. The integrated sensing unit refers to a hardware device that integrates soil moisture content sensors, soil matrix potential sensors, and soil temperature sensors, enabling the simultaneous acquisition of multiple physical quantities; for example, a sensor node buried 30 cm deep in a soil profile of a cornfield, capable of simultaneously measuring the volumetric moisture content, matrix potential, and temperature at that point. Soil moisture content refers to the volume ratio of water contained in a unit volume of soil; for example, the frequency domain reflectance sensor in the integrated sensing unit measures a soil volumetric moisture content of 0.32 m³ at a point in a cornfield. 3 / m 3 Soil matrix potential refers to the potential energy held by soil water due to adsorption and capillary action of the soil matrix; for example, the soil matrix potential at a point in a cornfield measured by a tensiometer in the integrated sensing unit is -25 kPa. Soil temperature refers to the internal temperature of the soil medium; for example, the soil temperature at a point in a cornfield measured by a digital temperature sensor in the integrated sensing unit is 18 °C. Raw data refers to electrical signals or numerical records directly acquired and output by the integrated sensing unit without any processing; for example, a set of raw measurement values ​​including device identification, timestamp, and soil moisture content, matrix potential, and temperature transmitted wirelessly from a sensing node in the field via a transmission module.

[0024] The data preprocessing module 120 is used to perform time synchronization processing and outlier correction based on physical thresholds on the raw data of each integrated sensing unit, and generate spatiotemporally aligned water content data stream, matrix potential data stream and temperature data stream corresponding to each integrated sensing unit.

[0025] Time synchronization processing refers to the data processing procedure of correcting and aligning data collected by different integrated sensing units at different times to a unified time coordinate axis; for example, all data reported by various sensing nodes distributed in a cornfield within a 10-second time window are marked and processed into corresponding data at whole-minute times. Outlier correction based on physical thresholds refers to the data cleaning operation of setting judgment limits based on the reasonable range of variation of soil physical parameters and identifying and replacing measurements exceeding the limits; for example, setting the upper limit of volumetric moisture content to 0.50 based on soil porosity, replacing an abnormal moisture content value of 0.55 reported by a certain node with 0.48 by linear interpolation based on normal data at adjacent times. Moisture content data stream refers to the soil moisture content data sequence formed after preprocessing, arranged continuously in chronological order, originating from the same integrated sensing unit; for example, an ordered sequence consisting of one volumetric moisture content data point generated per minute by sensing unit numbered Node_05 over 6 consecutive hours. A matrix potential data stream refers to a pre-processed, chronologically consecutive sequence of soil matrix potential data originating from the same integrated sensing unit; for example, an ordered sequence consisting of one matrix potential data point generated per minute by sensing unit Node_05 over a continuous 6-hour period. A temperature data stream refers to a pre-processed, chronologically consecutive sequence of soil temperature data originating from the same integrated sensing unit; for example, an ordered sequence consisting of one temperature data point generated per minute by sensing unit Node_05 over a continuous 6-hour period.

[0026] The dynamic collaborative inversion module 130 is used to couple the spatiotemporal gradient change relationship between the water content data stream and the matrix potential data stream corresponding to each integrated sensing unit based on the soil hydrodynamic model, and to iteratively calculate the soil hydraulic characteristic parameter set of the location of each integrated sensing unit in real time according to a set frequency. The soil hydraulic characteristic parameter set includes core model parameters used to determine the saturated permeability coefficient and the unsaturated hydraulic conductivity.

[0027] Among them, the soil hydrodynamic model refers to the physical mechanism mathematical model that describes and simulates the process of water transport in soil; for example, the Richards equation is used as the governing equation to characterize the movement of unsaturated water in maize field soil. The spatiotemporal gradient change relationship refers to the correlation and synergy between the rates of change of soil moisture content data and soil matrix potential data in the temporal and spatial dimensions; for example, during a rainfall event, there is a dynamic correlation between the rate of increase of moisture content at different points in the field over time and the rate of decrease of the absolute value of matrix potential over time. The set frequency refers to the time interval at which the dynamic collaborative inversion module is configured to initiate a parameter calculation task; for example, the system is set to perform a parameter inversion calculation every 5 minutes using the latest received data stream. The soil hydraulic characteristic parameter set refers to a set of key physical model parameters used to quantitatively characterize the water holding and hydraulic conductivity properties of specific soils; for example, in the van Genuchten-Mualem model, this includes the set of parameters such as saturated water content θs, residual water content θr, curve shape parameters α and n, and saturated hydraulic conductivity Ks. Saturated permeability coefficient refers to the rate at which water flows through soil under a unit hydraulic gradient when the soil is fully saturated; for example, the saturated hydraulic conductivity Ks of a certain point in a cornfield obtained through model inversion is 0.8 m / d. Unsaturated hydraulic conductivity refers to the functional relationship between the water-conducting capacity of soil in an unsaturated state and changes in soil moisture content or matrix potential; for example, the hydraulic conductivity K determined by the inverted model parameters is the mathematical function K(ψ) that varies with matrix potential ψ. Core model parameters are the key parameters that constitute the set of soil hydraulic characteristic parameters and directly determine the calculation of saturated permeability coefficient and the functional form of unsaturated hydraulic conductivity; for example, the parameters α, n, and Ks in the van Genuchten-Mualem model.

[0028] The spatiotemporal adaptive fusion module 140 is used to perform spatial interpolation and temporal extrapolation on the soil hydraulic characteristic parameter set of all integrated sensing unit locations and the corresponding temperature data stream by introducing the soil hydrodynamic model as a spatiotemporal covariate in the adaptive fusion algorithm, and to fuse and generate a saturated permeability coefficient field and an unsaturated hydraulic conductivity parameter field that cover the monitoring area and evolve continuously in spatiotemporal.

[0029] Among them, adaptive fusion algorithms refer to data integration calculation methods that can automatically adjust fusion weights or model structures based on the spatiotemporal statistical characteristics of input data and external physical constraints; for example, a spatiotemporal kriging interpolation algorithm that uses soil hydrodynamic model simulation results as auxiliary variables. Spatial interpolation and temporal extrapolation refer to the process of estimating parameter values ​​at unsampled spatial points within a monitoring area based on parameter data at known discrete spatial points, and predicting the future state of the parameters; for example, using parameters obtained from the inversion of 50 sensor nodes in a cornfield, estimating parameter values ​​at any location within the entire field, and predicting the parameter change trend over the next 30 minutes. Saturated permeability coefficient field refers to the spatial distribution map formed by continuously distributed saturated permeability coefficient values ​​within a two-dimensional or three-dimensional spatial range of the monitoring area; for example, a spatial distribution cloud map reflecting the gradual decrease in saturated hydraulic conductivity from east to west in a cornfield demonstration area. The unsaturated hydraulic conductivity parameter field refers to the spatial distribution pattern formed by the model parameters used to define the unsaturated hydraulic conductivity function at each point, which are continuously distributed within the spatial range of the monitoring area; for example, multiple spatial variation distribution maps of the van Genuchten model parameters α and n, respectively describing the same cornfield area.

[0030] The technical solution in this embodiment solves the problems of single dimension, insufficient spatial representativeness and response delay of traditional point measurement by using a multi-sensor collaborative monitoring architecture and real-time data fusion technology. It improves the spatiotemporal resolution and real-time perception capability of soil moisture field dynamic evolution, and meets the application needs of precision agriculture and disaster early warning.

[0031] In one alternative approach, the plurality of integrated sensing units are arranged in the monitoring area according to a non-uniform grid topology.

[0032] Among them, non-uniform grid topology refers to the spatial arrangement of integrated sensing units in the monitoring area according to non-equal spacing and irregular pattern; for example, denser sensing units are deployed at the irrigation inlet of cornfields and in low-lying areas prone to water accumulation, while sparser sensing units are deployed in flat and homogeneous areas.

[0033] Among the above-mentioned optional methods, the integrated sensing units are further deployed through a non-uniform grid topology structure, with denser deployment in key monitoring areas and sparse deployment in areas with gentle changes, thereby optimizing the spatial distribution of sensors and improving monitoring efficiency and data representativeness.

[0034] In one alternative approach, the non-uniform grid topology is adaptively densified based on a historical soil quality map or topographic elevation map of the monitored area.

[0035] Among them, the historical soil quality map refers to a map or digital layer reflecting the spatial distribution of soil types, texture classifications, and other attributes obtained from past surveys or investigations in the monitoring area; for example, the spatial distribution map of sandy loam and clay loam soils already drawn based on historical survey data in this agricultural experimental area. The topographic elevation map refers to a map or digital elevation model data describing the surface undulations and elevation changes in the monitoring area; for example, a high-precision digital elevation model map of the cornfield area obtained through lidar scanning. Adaptive densification refers to the process of dynamically adjusting the distribution density of sensor units based on auxiliary map information; for example, in the sandy soil area shown on the historical soil quality map, the preset 50 m deployment spacing is automatically adjusted to 30 m.

[0036] Among the above-mentioned optional methods, the non-uniform grid is further adaptively densified based on historical soil quality maps or topographic elevation maps, and the monitoring density is increased in areas with complex soil properties or drastic topographic undulations, so that the monitoring network layout is more in line with the actual spatial variation characteristics.

[0037] In one alternative approach, the outlier correction based on physical thresholds performed by the data preprocessing module 120 specifically includes: setting an upper threshold for the moisture content data stream based on the soil saturated moisture content, setting an upper and lower threshold for the matrix potential data stream based on the range of the soil water potential sensor, and replacing data points exceeding the corresponding thresholds with data generated based on time series linear interpolation.

[0038] Soil saturated water content refers to the maximum water content when all soil pores are filled with water; for example, the saturated volumetric water content of the clay loam soil in this cornfield is approximately 0.48 m³. 3 / m 3 This value is used as the upper limit threshold for judging abnormal moisture content during data preprocessing.

[0039] In the above-mentioned optional methods, outlier correction is further performed on the moisture content and matrix potential data through physical thresholds, replacing outlier data that exceed the soil physical properties or sensor range with linear interpolation values, thereby improving the continuity and reliability of the data stream.

[0040] In one alternative approach, the dynamic collaborative inversion module 130 is based on the Richards equations as the soil hydrodynamic model, and the dynamic collaborative inversion module uses an ensemble Kalman filter algorithm to assimilate the water content data stream and the matrix potential data stream, and performs the real-time iterative solution.

[0041] The Richards equation refers to a partial differential equation that combines Darcy's law and the law of conservation of mass to describe the movement of soil moisture in the unsaturated zone. For example, the dynamic collaborative inversion module uses a one-dimensional vertical Richards equation to simulate the infiltration process of water in a maize field soil profile. The ensemble Kalman filter algorithm is a data assimilation algorithm that performs state estimation and uncertainty analysis by maintaining and updating a set of system states. For example, this algorithm uses 200 different sets of soil hydraulic characteristic parameters to characterize uncertainty and iteratively updates these sets by assimilating continuous data streams of water content and matric potential observations to solve for the optimal parameters.

[0042] Among the above-mentioned optional methods, Richards equations are further adopted as the hydrodynamic model, and the ensemble Kalman filter algorithm is combined to assimilate the water content and matrix potential data streams, so as to realize the dynamic fusion of multi-source observation information and the real-time updating of model parameters, thereby enhancing the timeliness of the inversion process.

[0043] In one alternative approach, the state vector of the ensemble Kalman filter algorithm includes the set of soil hydraulic characteristic parameters, and the observation vector of the ensemble Kalman filter algorithm is composed of the water content data stream and the matrix potential data stream.

[0044] In the above-mentioned optional approach, the soil hydraulic characteristic parameter set is further used as the state vector, and the water content and matrix potential data stream is used as the observation vector to construct an ensemble Kalman filter framework, so as to achieve synergistic optimization of model parameters and system state and improve the physical consistency of inversion.

[0045] In one alternative approach, when the dynamic collaborative inversion module 130 solves the soil hydraulic characteristic parameter set, the objective function F(θ) is defined as the weighted sum of squares of the difference between the observed data and the model prediction, and a spatiotemporal smoothing constraint term of the soil hydraulic characteristic parameter set is introduced.

[0046] The expression for the objective function F(θ) is:

[0047] Where θ represents the set of soil hydraulic characteristic parameters to be solved, T represents the total number of time steps, and N represents the total number of integrated sensing units. and Let represent the matrix potential and volumetric water content observed by the i-th integrated sensing unit at time t, respectively. and These represent the corresponding values ​​obtained from simulations based on the soil hydrodynamic model and parameter θ. and λ represents the weighting coefficients of the matrix potential and water content observations, respectively. R(θ) represents the spatiotemporal smoothing constraint function for the parameter set θ, and λ is the weighting coefficient of the constraint term.

[0048] In the above-mentioned optional approach, a spatiotemporal smoothing constraint term is further introduced into the objective function to constrain the spatiotemporal continuity of parameters while minimizing the difference between observation and simulation, thereby avoiding excessive oscillations in the parameter solution process and enhancing the physical rationality of the inversion results.

[0049] In one alternative approach, in the spatiotemporal adaptive fusion module 140, the soil hydrodynamic model as a spatiotemporal covariate is specifically manifested as follows: the spatiotemporal gradient field of soil moisture predicted by the soil hydrodynamic model is used as the input of the variogram model in the adaptive fusion algorithm.

[0050] Among them, the spatiotemporal gradient field of soil moisture refers to the spatial rate of change of soil moisture parameters and their distribution over time, calculated by soil hydrodynamic models; for example, the spatiotemporal distribution map of the horizontal and vertical transport rates and directions of water in a maize field after irrigation, predicted by the model simulation. The variogram model refers to a mathematical model used in geostatistics to quantify the spatial autocorrelation of regionalized variables as a function of distance; for example, the exponential model used in Kriging interpolation to describe the spatial correlation structure of saturated hydraulic conductivity parameters.

[0051] In the above-mentioned optional approach, the spatiotemporal gradient field predicted by the soil hydrodynamic model is further used as the input of the variogram function, so that the adaptive fusion algorithm incorporates physical mechanism constraints, thereby improving the response capability and prediction accuracy of spatial interpolation to the dynamic evolution of soil moisture.

[0052] In one optional embodiment, when the spatiotemporal adaptive fusion module 140 performs spatial interpolation and temporal extrapolation, the adaptive fusion algorithm used is the physically constrained spatiotemporal kriging algorithm; the physically constrained spatiotemporal kriging algorithm predicts the soil hydraulic characteristic parameters at time t0 for the location s0 where no sensing units are deployed. This is obtained by solving the following Kriging equations:

[0053] Where K represents the total number of neighboring sensor unit observations used for prediction. Indicates position At any moment The observed value of a certain parameter in the set of soil hydraulic characteristic parameters. For corresponding Kriging weight coefficients, This indicates that the soil hydrodynamic model is located at... and time For the same predicted parameter value, β is the physical constraint coefficient corresponding to the model's predicted value. Indicates the observed parameters at the spatiotemporal point and Covariance between This indicates the relationship between model predictions and observed parameters at spatiotemporal points. and The covariance between them.

[0054] Among them, the physical constraint spatiotemporal kriging algorithm refers to an advanced data fusion algorithm that uses the prediction results of physical mechanism models as soft constraints and embeds them into the classical spatiotemporal kriging interpolation framework. For example, when estimating the parameters of the location where no sensing units are deployed, the algorithm simultaneously considers the observations of neighboring nodes and the predictions of Richards equations for that location, and obtains the final estimate by weighted solution.

[0055] Among the above-mentioned optional methods, a physical constraint spatiotemporal kriging algorithm is further adopted. The model prediction value is used as a constraint to solve the kriging equation system, thereby achieving optimal weighted fusion of observation data and physical model, and improving the accuracy and spatiotemporal continuity of parameter prediction for unmeasured points.

[0056] To better illustrate the technical solution of this embodiment, the following example is used for complete explanation: Figure 2 This diagram illustrates a flowchart of an embodiment of a method for dynamic monitoring of soil infiltration characteristics based on multi-sensor fusion provided by the present invention. Figure 2 As shown, it includes the following steps: S1. Through multiple integrated sensing units deployed in the monitoring area, the raw data of soil moisture content, soil matrix potential and soil temperature at the location of each integrated sensing unit are collected simultaneously. S2. Perform time synchronization processing and outlier correction based on physical thresholds on the raw data of each integrated sensing unit to generate spatiotemporally aligned water content data stream, matrix potential data stream and temperature data stream corresponding to each integrated sensing unit. S3. Based on the soil hydrodynamic model, the spatiotemporal gradient change relationship between the water content data stream and the matrix potential data stream corresponding to each integrated sensing unit is coupled, and the soil hydraulic characteristic parameter set of each integrated sensing unit is iteratively calculated in real time according to the set frequency. The soil hydraulic characteristic parameter set includes the core model parameters used to determine the saturated permeability coefficient and the unsaturated hydraulic conductivity. S4. Based on the soil hydraulic characteristic parameter set and corresponding temperature data stream of all integrated sensing units, an adaptive fusion algorithm is introduced by introducing the soil hydrodynamic model as a spatiotemporal covariate to perform spatial interpolation and temporal extrapolation on the soil hydraulic characteristic parameter set of all integrated sensing units, and fused to generate a saturated permeability coefficient field and an unsaturated hydraulic conductivity parameter field that cover the monitoring area and evolve continuously in spatiotemporal.

[0057] The technical solution in this embodiment solves the problems of single dimension, insufficient spatial representativeness and response delay of traditional point measurement by using a multi-sensor collaborative monitoring architecture and real-time data fusion technology. It improves the spatiotemporal resolution and real-time perception capability of soil moisture field dynamic evolution, and meets the application needs of precision agriculture and disaster early warning.

[0058] Furthermore, the system provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the system can be divided into different functional modules according to the actual situation to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0059] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.

[0060] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.

[0061] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion, characterized in that, The system includes: The data acquisition module is used to simultaneously collect raw data on soil moisture content, soil matrix potential, and soil temperature at the location of each integrated sensing unit by multiple integrated sensing units deployed in the monitoring area. The data preprocessing module is used to perform time synchronization processing and outlier correction based on physical thresholds on the raw data of each integrated sensing unit, and generate spatiotemporally aligned water content data stream, matrix potential data stream and temperature data stream corresponding to each integrated sensing unit. The dynamic collaborative inversion module is used to couple the spatiotemporal gradient change relationship between the water content data stream and the matrix potential data stream corresponding to each integrated sensing unit based on the soil hydrodynamic model, and iteratively calculate the soil hydraulic characteristic parameter set of each integrated sensing unit in real time at a set frequency. The soil hydraulic characteristic parameter set includes core model parameters used to determine the saturated permeability coefficient and unsaturated hydraulic conductivity. The spatiotemporal adaptive fusion module is used to perform spatial interpolation and temporal extrapolation on the soil hydraulic characteristic parameter set of all integrated sensing unit locations and the corresponding temperature data stream by introducing the soil hydrodynamic model as a spatiotemporal covariate in the adaptive fusion algorithm. This process generates a saturated permeability coefficient field and an unsaturated hydraulic conductivity parameter field that cover the monitoring area and evolve continuously in spatiotemporal.

2. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 1, characterized in that, The multiple integrated sensing units are arranged in the monitoring area according to a non-uniform grid topology.

3. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 2, characterized in that, The non-uniform grid topology is adaptively encrypted based on the historical soil quality map or topographic elevation map of the monitoring area.

4. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 1, characterized in that, The outlier correction based on physical thresholds performed by the data preprocessing module specifically includes: setting an upper threshold for the moisture content data stream based on the soil saturated moisture content, setting an upper and lower threshold for the matrix potential data stream based on the range of the soil water potential sensor, and replacing data points exceeding the corresponding thresholds with data generated based on time series linear interpolation.

5. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 1, characterized in that, The dynamic collaborative inversion module is based on the Richards equation as the soil hydrodynamic model. The dynamic collaborative inversion module uses an ensemble Kalman filter algorithm to assimilate the water content data stream and the matrix potential data stream, and performs the real-time iterative solution.

6. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 5, characterized in that, The state vector of the ensemble Kalman filter algorithm includes the soil hydraulic characteristic parameter set, and the observation vector of the ensemble Kalman filter algorithm is composed of the water content data stream and the matrix potential data stream.

7. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 1 or 5, characterized in that, When the dynamic collaborative inversion module solves the soil hydraulic characteristic parameter set, the objective function F(θ) is defined as the weighted sum of squares of the difference between the observed data and the model prediction, and a spatiotemporal smoothing constraint term of the soil hydraulic characteristic parameter set is introduced. The expression for the objective function F(θ) is: Where θ represents the set of soil hydraulic characteristic parameters to be solved, T represents the total number of time steps, and N represents the total number of integrated sensing units. and Let represent the matrix potential and volumetric water content observed by the i-th integrated sensing unit at time t, respectively. and These represent the corresponding values ​​obtained from simulations based on the soil hydrodynamic model and parameter θ. and λ represents the weighting coefficients of the matrix potential and water content observations, respectively. R(θ) represents the spatiotemporal smoothing constraint function for the parameter set θ, and λ is the weighting coefficient of the constraint term.

8. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 1, characterized in that, In the spatiotemporal adaptive fusion module, the soil hydrodynamic model, as a spatiotemporal covariate, specifically manifests as follows: the spatiotemporal gradient field of soil moisture predicted by the soil hydrodynamic model is used as the input of the variogram model in the adaptive fusion algorithm.

9. The dynamic monitoring system for soil infiltration characteristics based on multi-sensor fusion according to claim 8, characterized in that, When the spatiotemporal adaptive fusion module performs spatial interpolation and temporal extrapolation, the adaptive fusion algorithm used is the physically constrained spatiotemporal kriging algorithm; the physically constrained spatiotemporal kriging algorithm predicts the soil hydraulic characteristic parameters at time t0 for the location s0 without sensor units. This is obtained by solving the following Kriging equations: Where K represents the total number of neighboring sensor unit observations used for prediction. Indicates position At any moment The observed value of a certain parameter in the set of soil hydraulic characteristic parameters. For corresponding Kriging weight coefficients, This indicates that the soil hydrodynamic model is located at... and time For the same predicted parameter value, β is the physical constraint coefficient corresponding to the model's predicted value. Indicates the observed parameters at the spatiotemporal point and Covariance between This indicates the relationship between model predictions and observed parameters at spatiotemporal points. and The covariance between them.

10. A method for dynamic monitoring of soil infiltration characteristics based on multi-sensor fusion, characterized in that, The method includes: The system uses multiple integrated sensing units deployed in the monitoring area to simultaneously collect raw data on soil moisture content, soil matrix potential, and soil temperature at the locations of each integrated sensing unit. The raw data of each integrated sensing unit is time-synchronized and outlier correction is performed based on physical thresholds to generate spatiotemporally aligned water content data stream, matrix potential data stream and temperature data stream corresponding to each integrated sensing unit. Based on the soil hydrodynamic model, the spatiotemporal gradient change relationship between the water content data stream and the matrix potential data stream corresponding to each integrated sensing unit is coupled, and the soil hydraulic characteristic parameter set of each integrated sensing unit is iteratively calculated in real time at a set frequency. The soil hydraulic characteristic parameter set includes core model parameters used to determine the saturated permeability coefficient and unsaturated hydraulic conductivity. Based on the soil hydraulic characteristic parameter set and corresponding temperature data stream at the locations of all integrated sensing units, an adaptive fusion algorithm is used to introduce the soil hydrodynamic model as a spatiotemporal covariate to perform spatial interpolation and temporal extrapolation on the soil hydraulic characteristic parameter set at the locations of all integrated sensing units. This results in the fusion of a continuously spatiotemporally evolving saturated permeability coefficient field and unsaturated hydraulic conductivity parameter field covering the monitoring area.

Images