A Smart Digital Rural Environmental Monitoring System and Method Based on Big Data

Abstract

This invention relates to the field of environmental monitoring technology, specifically to a smart digital rural environmental monitoring system and method based on big data. The system includes a stratified water potential acquisition module, a gradient inversion correction module, a path inversion modeling module, and a multidimensional diffusion modeling module. The stratified water potential acquisition module calculates and generates rural soil environmental monitoring data. This data includes stratified soil water potential distribution, soil moisture content gradient, and soil electrical conductivity characteristics. Abnormal soil water flow data includes vertical water potential difference, abnormal flow direction vector, and flow intensity level. The water infiltration path distribution map includes the infiltration path distribution structure, soil water potential accumulation intervals, and flow propagation direction indicators. The diffusion boundary space model includes the pollutant diffusion boundary range, water lateral migration area, and multidimensional diffusion rate distribution. This solution quantifies the abnormal distribution range, duration, and development trend, improving the accuracy of judging abnormal infiltration and diffusion processes.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, and in particular to a smart digital rural environmental monitoring system and method based on big data. Background Technology

[0002] The field of environmental monitoring technology involves the real-time sensing, collection, transmission, and analysis of air, water, soil, meteorological, and ecological elements in the natural environment. Its core lies in utilizing sensors, data acquisition terminals, communication networks, and information processing platforms to continuously monitor and manage environmental parameters digitally, thereby providing data support for pollution source supervision, ecological protection, and resource utilization. This field includes environmental data acquisition technology, data transmission and storage technology, statistical and analytical methods for monitoring data, and the construction and application of environmental information management systems. In recent years, with the development of informatization and intelligentization, environmental monitoring has gradually evolved towards automation, networking, and digitalization.

[0003] Traditional smart digital rural environmental monitoring systems refer to comprehensive information systems applied to environmental monitoring in rural areas. They primarily monitor and manage environmental indicators such as air quality, water quality, noise, and soil moisture in rural regions. Traditional systems collect data by deploying multiple types of sensor nodes and transmit the data to a centralized monitoring center via wired or wireless communication. The data is then processed and displayed by management terminals. These systems rely on manually setting sampling frequencies and monitoring points, and data processing often uses fixed threshold comparisons or simple statistical analysis methods. This makes it difficult to achieve dynamic correlation and trend identification of multi-source environmental data, and the coordination between monitoring devices is low, resulting in limited data utilization.

[0004] Current rural environmental monitoring relies on distributed sensor nodes for data collection, with data often limited to single-point parameter recording. The lack of correlation between monitoring depths makes it impossible to identify the potential energy relationships between different soil layers, hindering the understanding of vertical water flow patterns. Monitoring operations are maintained by manually set rules, resulting in a sampling rhythm disconnected from actual soil changes and difficulty in capturing rapid infiltration or covert backflow phenomena. Data processing focuses on comparing stable intervals, lacking the ability to identify interlayer differences, persistent anomalies, and dynamic shifts. This makes it difficult to determine the direction of water migration, define abnormal diffusion areas, and deviate from the true rhythm of soil changes, impacting the early detection and response to potential risks. Summary of the Invention

[0005] To address the shortcomings of existing rural environmental monitoring systems, which rely on distributed sensor nodes for data collection, resulting in single-point parameter recordings and a lack of correlation between monitoring depths, the current system fails to identify potential energy relationships between different soil layers and thus struggles to reveal longitudinal water flow patterns. Furthermore, the monitoring process relies on manually set rules, leading to a disconnect between sampling rhythms and actual soil changes, making it difficult to capture rapid infiltration or hidden backflow phenomena. Data processing focuses on comparing stable intervals, lacking the ability to identify interlayer differences, persistent anomalies, and dynamic shifts, making it difficult to determine water migration directions, define abnormal diffusion areas, and deviate from the true rhythm of soil changes, thus hindering the early detection and handling of potential risks. This invention provides a smart digital rural environmental monitoring system and method based on big data, with the technical solution as follows:

[0006] The smart digital village environmental monitoring system based on big data includes:

[0007] The stratified water potential acquisition module collects voltage signals, water conductivity signals and micro-pressure signals through sensor nodes, performs standardized calculations, binds the stratification identifiers to the depth of the acquisition nodes, calculates the slope angle, generates rural soil environmental monitoring data, and transmits it to the gradient inverse correction module.

[0008] The gradient inverse correction module calls the rural soil environment monitoring data, extracts the potential energy of adjacent layers, calculates the potential energy difference, and performs inverse correction when the potential energy of the lower layer is greater than that of the upper layer and the difference exceeds a preset threshold, generating soil water potential abnormal flow data, which is then transmitted to the path inverse modeling module.

[0009] The path inversion modeling module calls the soil water potential anomaly flow data, uses a random forest model to extract multi-level direction vectors and potential energy changes, accumulates the potential energy changes layer by layer, calculates the infiltration propagation path according to depth, generates a water infiltration path distribution map, and transmits it to the multi-dimensional diffusion modeling module.

[0010] The multidimensional diffusion modeling module calls the moisture infiltration path distribution map, uses a sensor fusion algorithm for real-time monitoring, calibration and analysis, combines spatial distribution characteristics to analyze the diffusion trends of moisture and pollutants, constructs a diffusion boundary spatial model, and transmits it to the anomaly detection and monitoring module.

[0011] As a further aspect of the present invention, the rural soil environmental monitoring data includes stratified soil water potential distribution, soil moisture content gradient and soil electrical conductivity characteristic values; the soil water potential abnormal flow data includes vertical water potential difference, abnormal water flow direction vector and water flow intensity level; the water infiltration path distribution map includes infiltration path distribution structure, soil water potential accumulation interval and water flow propagation direction marker; and the diffusion boundary space model includes pollutant diffusion boundary range, water lateral migration area and multidimensional diffusion rate distribution.

[0012] As a further aspect of the present invention, the stratified water potential acquisition module includes:

[0013] The signal acquisition submodule acquires the voltage signal, moisture conductivity signal and micro-pressure signal output by the layered sensing nodes, performs comparison based on the time synchronization benchmark, calculates the instantaneous difference value after normalizing the signal amplitude, filters out abnormal fluctuation points that exceed three times the preset standard deviation, and calculates the weighted average value after determining the weight based on the signal-to-noise ratio of the signal, generating a multi-source signal fusion value.

[0014] The node identification binding submodule collects node depth data for hierarchical indexing based on the multi-source signal fusion value, establishes a mapping table in ascending order of depth, binds signal, layer number and node spatial coordinate data, calculates the signal difference between adjacent nodes and analyzes the gradient change ratio, and generates the stratified water potential gradient coefficient.

[0015] The slope angle calculation submodule, based on the layered water potential gradient coefficient, calls the node spatial coordinate data, calculates the horizontal projection distance and vertical height difference, and calculates the slope angle using the arctangent function based on the ratio of the two, generating rural soil environmental monitoring data.

[0016] As a further aspect of the present invention, the gradient inversion correction module includes:

[0017] The data potential energy extraction submodule obtains the water potential and electrical conductivity of the deep layer in the rural soil environment monitoring data, performs hierarchical division and numbering on adjacent layer data, calculates the interlayer gradient change rate based on the depth and water potential distribution, and generates layer potential energy values.

[0018] The potential energy difference determination submodule calls the hierarchical potential energy value, compares the potential energy elements of the upper and lower layers and performs a judgment based on the potential energy difference threshold, identifies abnormal nodes where the lower layer potential energy is greater than the upper layer and the difference exceeds the threshold, and clusters to calculate the average potential energy difference to generate a cross-layer potential energy difference set.

[0019] The abnormal flow calculation submodule calls the interlayer potential energy difference set, derives the soil permeability coefficient based on the interlayer potential energy difference and gradient change characteristics, calculates the potential energy driving direction and water potential transmission rate, analyzes the time series distribution of abnormal paths based on the rate change trend, and generates abnormal soil water potential flow data.

[0020] As a further aspect of the present invention, the potential energy difference threshold is a value dynamically determined based on the average potential energy difference between the upper and lower layers within a preset range. When the difference between the potential energy of the lower layer and the potential energy of the upper layer exceeds the threshold, it is identified as an abnormal node.

[0021] The calculation of the interlayer gradient change rate refers to obtaining the change trend between depth and water potential by calculating the difference between data of adjacent layers, and determining the change rate between each layer based on the relative distribution relationship between depth and water potential.

[0022] As a further aspect of the present invention, the path inversion modeling module includes:

[0023] The soil water potential data integration submodule acquires the abnormal flow data of soil water potential, formats and groups the time series of different depth levels, detects the differences in flow rate and gradient, removes samples with discontinuous time, and performs normalization processing using the Min-Max normalization method to generate soil water potential distribution data.

[0024] The direction vector extraction submodule calls the soil water potential distribution data, uses a random forest model to perform feature splitting on the multi-layer water potential gradient data, extracts multi-dimensional potential energy change features and calculates direction vector weights, and clusters and adjusts the node directions according to feature importance to generate a multi-level potential energy direction array.

[0025] The infiltration path calculation submodule, based on the multi-level potential energy direction array, accumulates the node potential energy difference layer by layer to obtain the potential energy change sequence, calculates the multi-level potential energy cumulative gradient and determines the path connection relationship, statistically analyzes the distribution density and diffusion range, and generates a water infiltration path distribution map.

[0026] As a further aspect of the present invention, the multidimensional diffusion modeling module includes:

[0027] The moisture path receiving submodule collects the coordinate values ​​of path nodes and the moisture content values ​​of nodes based on the moisture infiltration path distribution map, compares them with the offset threshold and records the offset, calculates the moisture content gradient of adjacent nodes based on the distance between nodes, adjusts the moisture content values ​​of nodes according to the offset, and generates the path gradient distribution.

[0028] The moisture situation construction submodule calls the path gradient distribution, fuses multi-source sensor readings for time alignment and weighting, calculates gradient correction based on humidity deviation benchmark, superimposes correction to adjust path gradient distribution and calculates spatial humidity distribution to obtain spatial humidity distribution.

[0029] The diffusion boundary generation submodule analyzes the changes in spatial humidity and pollutant concentration based on the spatial humidity distribution and calculates the diffusion boundary threshold according to the concentration change benchmark value. It then calls the diffusion boundary threshold to divide the humidity distribution segment and records the diffusion edge position of the segment, generating a diffusion boundary spatial model.

[0030] The offset threshold is determined by calculating quantile statistics from the original sampling sequence of node moisture content values, and is a quantile point higher than the median;

[0031] The humidity deviation benchmark value is determined by calculating the statistical characteristics of the difference amplitude within the sliding window using the humidity difference sequence at the position corresponding to the path gradient distribution.

[0032] The concentration variation benchmark value is determined by calculating the statistical characteristics of the sliding window difference amplitude from the time series of pollutant concentration changes.

[0033] As a further embodiment of the present invention, the anomaly detection and monitoring module calls the diffusion boundary space model, extracts anomaly indicators and state duration, classifies and statistically analyzes anomaly types, calculates the proportion of anomaly areas to the total monitoring area, analyzes the development trend based on the trend of proportion change, calculates the response time window, and generates statistical results of anomaly indicators.

[0034] The statistical results of the abnormal indicators include statistics on the classification of abnormal types, the changing trend of the proportion of abnormal areas, and the threshold of the response time window.

[0035] As a further aspect of the present invention, the anomaly detection and monitoring module includes:

[0036] The abnormal feature extraction submodule calls the diffusion boundary space model to obtain abnormal indicators in the monitoring area, extracts the indicator values ​​and timestamps of the monitoring nodes that deviate from the normal range boundary, calculates the duration of the current time and the detection start time, performs cross-matching to classify abnormal types, and generates an abnormal indicator classification statistics set.

[0037] The abnormal situation assessment submodule calls the abnormal indicator classification statistics set, extracts the spatial coverage area value of the abnormality, calculates the ratio of the abnormal coverage area to the total monitoring area, obtains the abnormal area ratio coefficient, calculates the rate of change of the ratio coefficient between adjacent time points, and generates the abnormal spread trend quantity.

[0038] The response window calculation submodule extracts the diffusion rate parameter and the anomaly type weight coefficient based on the anomaly diffusion trend quantity, calculates the correlation value between the diffusion rate and the weight coefficient, analyzes the allowable response time to obtain the response time window interval, summarizes it with the anomaly indicator classification statistics set, and generates anomaly indicator statistical results.

[0039] Based on the same inventive concept, a smart digital village environmental monitoring method based on big data is also proposed. This method is executed based on the aforementioned smart digital village environmental monitoring system based on big data and includes the following steps:

[0040] S1: Collect voltage signals, moisture conductivity signals and micro-pressure signals through sensor nodes, perform standardized calculations, bind layered identifiers to the depth of the data collection nodes, calculate slope angles, and generate rural soil environmental monitoring data.

[0041] S2: Call the rural soil environment monitoring data, extract the potential energy of adjacent layers, calculate the potential energy difference, and when the potential energy of the lower layer is greater than that of the upper layer and the difference exceeds the preset threshold, perform the adjustment operation to generate abnormal soil water flow data.

[0042] S3: Call the soil water potential anomaly flow data, use the random forest model to extract multi-level direction vectors and potential energy changes, accumulate the potential energy changes layer by layer, calculate the infiltration propagation path according to depth, and generate a water infiltration path distribution map.

[0043] S4: Call the moisture infiltration path distribution map, use sensor fusion algorithm for real-time monitoring, calibration and analysis, combine spatial distribution characteristics to analyze the diffusion trend of moisture and pollutants, and construct a diffusion boundary spatial model;

[0044] S5: Call the diffusion boundary space model, extract abnormal indicators and state duration, classify and statistically analyze abnormal types, calculate the proportion of abnormal areas to the total monitoring area, analyze the development trend based on the trend of proportion change, calculate the response time window, and generate statistical results of abnormal indicators.

[0045] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0046] A longitudinal profile moisture state description is constructed by hierarchical potential energy correlation calculation. The potential energy difference between layers is extracted by deeply bound hierarchical identifiers and reverse calibration is triggered, so that abnormal flow trends can be identified in the early stage. Infiltration paths are generated by multi-level direction vectors and potential energy accumulation, realizing a continuous expression of the water migration trajectory within the profile. The diffusion model with boundary extrapolation capability is formed by calibrating by combining spatial distribution characteristics with multi-source sensor data. This allows for a quantitative description of the abnormal distribution range, duration and development trend, improving the accuracy of judgment on abnormal infiltration and diffusion processes and strengthening targeted monitoring and response capabilities. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the accompanying drawings without creative effort.

[0048] Figure 1 This is a system schematic diagram of the present invention;

[0049] Figure 2 This is a schematic diagram of the system framework of the present invention;

[0050] Figure 3 This is a flowchart of the stratified water potential acquisition module in this invention;

[0051] Figure 4 This is a flowchart of the gradient inversion correction module in this invention;

[0052] Figure 5This is a flowchart of the path inversion modeling module in this invention;

[0053] Figure 6 This is a flowchart of the multidimensional diffusion modeling module in this invention;

[0054] Figure 7 This is a flowchart of the anomaly detection and monitoring module in this invention;

[0055] Figure 8 This is a flowchart of the method of the present invention. Detailed Implementation

[0056] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0057] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0058] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0059] In this embodiment of the invention, sometimes the subscript such as W1 is written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0060] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0061] This invention provides a smart digital rural environmental monitoring system based on big data, such as... Figure 1 and Figure 2 The diagram shown illustrates a smart digital rural environmental monitoring system based on big data. The system includes:

[0062] The stratified water potential acquisition module collects voltage signals, water conductivity signals and micro-pressure signals through sensor nodes, performs standardized calculations, binds the stratification identifiers to the depth of the acquisition nodes, calculates the slope angle, generates rural soil environmental monitoring data, and transmits it to the gradient inverse correction module.

[0063] The gradient inverse correction module calls rural soil environmental monitoring data, extracts potential energy of adjacent layers, calculates potential energy difference, and performs inverse correction when the potential energy of the lower layer is greater than that of the upper layer and the difference exceeds a preset threshold, generating abnormal soil water potential flow data, which is then transmitted to the path inverse modeling module.

[0064] The path inversion modeling module calls soil water potential anomaly flow data, uses a random forest model to extract multi-level direction vectors and potential energy changes, accumulates potential energy changes layer by layer, calculates infiltration propagation paths according to depth, generates a water infiltration path distribution map, and transfers it to the multi-dimensional diffusion modeling module.

[0065] The multidimensional diffusion modeling module calls the water infiltration path distribution map, uses sensor fusion algorithm for real-time monitoring, calibration and analysis, combines spatial distribution characteristics to analyze the diffusion trend of water and pollutants, constructs a diffusion boundary spatial model, and transmits it to the anomaly detection and monitoring module.

[0066] The anomaly detection and monitoring module calls the diffusion boundary space model to extract anomaly indicators and state duration, classifies and statistically analyzes anomaly types, calculates the proportion of anomaly areas to the total monitoring area, analyzes the development trend based on the trend of proportion changes, calculates the response time window, and generates statistical results of anomaly indicators.

[0067] Rural soil environmental monitoring data includes stratified soil water potential distribution, soil moisture content gradient, and soil electrical conductivity characteristics. Soil water potential anomaly flow data includes vertical water potential difference, abnormal water flow direction vector, and water flow intensity level. Water infiltration path distribution map includes infiltration path distribution structure, soil water potential accumulation interval, and water flow propagation direction marker. Diffusion boundary spatial model includes pollutant diffusion boundary range, water lateral migration area, and multidimensional diffusion rate distribution. Anomaly index statistical results include anomaly type classification statistics, anomaly area proportion change trend, and response time window threshold.

[0068] Specifically, such as Figure 2 and Figure 3 As shown, the stratified water potential acquisition module includes:

[0069] The signal acquisition submodule acquires the voltage signal, moisture conductivity signal and micro-pressure signal output by the layered sensing nodes, performs comparison based on the time synchronization benchmark, calculates the instantaneous difference value after normalizing the signal amplitude, filters out abnormal fluctuation points that exceed three times the preset standard deviation, and calculates the weighted average value after determining the weight based on the signal-to-noise ratio of the signal, generating a multi-source signal fusion value.

[0070] Based on the preset timing synchronization benchmark, namely the system timestamp 14:30:00.000 as the base point, it is allowed The fluctuation range in milliseconds applies to monitoring point A. , , The voltage, moisture conductivity, and micro-pressure signals collected by the deep stratified sensing nodes at 14:30:00.025 were subjected to consistency verification. Due to the deviation between the collection timestamp and the reference, milliseconds in the setting Data is received within a millisecond range, specifically as follows: Voltage signal output by the depth node Moisture conductivity signal and micro-pressure signal The three signals were then normalized, and the voltage signal range was set to... The range of the moisture conductivity signal is set to The range of the micro-pressure signal is set to The normalized amplitude calculation process is as follows: normalized voltage signal value Normalized value of water conductivity signal Normalized value of micro-pressure signal Next, retrieve the previous moment. normalized value , , Calculate the instantaneous difference value, i.e., the voltage signal difference. Differential moisture conductivity signal Micro-pressure signal differential Set the screening threshold for abnormal fluctuation points as follows: This threshold is statistically derived from the upper boundary of the distribution interval of the normalized difference values ​​of signals at 95% of adjacent time points in the original monitoring data. , , All less than Therefore, the currently collected signal was determined to be a valid data point. Next, a weighted average was calculated for the three normalized signals. The weights were set based on the correlation between different signals and soil water potential. The water conductivity signal directly reflects ion concentration and has the highest correlation; the micropressure signal reflects pore water pressure, followed by the voltage signal, which is a basic working parameter and has the lowest correlation. Therefore, their weighting coefficients were set as follows: water conductivity... micro pressure ,Voltage The sum of the three is Calculate the weighted average, specifically as follows:

[0071] 0.500×0.2+0.267×0.5+0.125×0.3=0.1000+0.1335+0.0375=0.271;

[0072] The calculation result is... Multi-source signal fusion value of deep nodes.

[0073] The node identification binding submodule collects node depth data for hierarchical indexing based on multi-source signal fusion values, establishes a mapping table in ascending order of depth, binds signal, layer number and node spatial coordinate data, calculates the signal difference between adjacent nodes and analyzes the gradient change ratio, and generates the stratified water potential gradient coefficient.

[0074] Based on the generated multi-source signal fusion values, for example, the system has collected the fusion values ​​of three nodes at different depths on the same vertical profile, as shown in Table 1. First, the depth data of the nodes is collected and then... , , The depth values ​​are indexed hierarchically, and a mapping table is established between layer numbers and node numbers. That is, layer 1 corresponds to node A001, layer 2 corresponds to node A002, and layer 3 corresponds to node A003. Then, the signal fusion value of the layer, the layer number, and the node spatial coordinate data retrieved from the equipment file are bound together to form a structured hierarchical data record. For example, the record for layer 1 is {signal: Layer number: 1, Spatial coordinates: ( , , Next, the signal difference between adjacent layer nodes is calculated. The signal difference between layer 2 and layer 1 is... The signal difference between layer 3 and layer 2 is Then analyze the rate of gradient change and calculate... and ratio A stable gradient change rate range was set, which was determined based on soil homogeneity experimental data. For loam with relatively uniform texture, the gradient change rate under steady-state seepage was within a certain range. Within the range, due to the calculated ratio Within this range, it indicates a stable soil moisture infiltration trend, with no obvious preferential flow or perched layer. Finally, the average signal difference between adjacent layers in this profile is calculated. As the stratified water potential gradient coefficient in this vertical direction.

[0075] Table 1: Initial Data Table for Layered Sensing Nodes

[0076] Node number Multi-source signal fusion value Node depth (cm) Spatial coordinates X (m) Spatial coordinates Y (m) Spatial coordinates Z (m) A001 0.271 20 10.5 20.2 -0.2 A002 0.355 40 10.5 20.2 -0.4 A003 0.442 60 10.5 20.2 -0.6

[0077] As shown in Table 1, the table lists the numbers, fusion values, depths, and three-dimensional spatial coordinates of the three sensor nodes deployed along the vertical direction, where the Z coordinate represents the depth relative to the ground surface.

[0078] The slope angle calculation submodule, based on the stratified water potential gradient coefficient, calls the node spatial coordinate data, calculates the horizontal projection distance and vertical height difference, and uses the arctangent function to calculate the slope angle based on the ratio of the two, generating rural soil environmental monitoring data.

[0079] The stratified water potential gradient coefficients are calculated based on two different spatial locations, profile A and profile B, within the designated monitoring area. Specifically, profile A is located at... The water potential representation value for depth (i.e., the fused value of node A002) is: Section B is in the same The water potential of depth represents the value The system retrieves the spatial coordinate data of these two monitoring nodes (A002 and B002). The coordinates of A002 are ( , , The coordinates of B002 are ( ), , , First, calculate the horizontal projected distance between the two points. The calculation process involves adjusting the difference in the X coordinates. The difference between the square of the y coordinate and the y coordinate Sum of squares and then take the square root, that is... Then, the vertical elevation difference between the two points is calculated. Here, the vertical elevation difference does not refer to the difference in Z coordinates, but rather the vertical component of the equivalent hydraulic gradient characterized by the difference in water potential, that is, the difference in the representative values ​​of the water potential between the two points. This value represents the potential energy difference driving the water flow. Next, based on the ratio of the horizontal projection distance to the vertical height difference, the arctangent function is used to calculate the slope angle, i.e., to calculate... The result is approximately Radius, converted to degrees is The angle represents the degree of inclination of the water potential gradient in the soil layer at a specific depth in the horizontal direction. The final data set, which includes the slope angle value, monitoring point coordinates, depth, timestamp, and other information, together constitutes the rural soil environment monitoring data.

[0080] Specifically, such as Figure 2 and Figure 4 As shown, the gradient inverse correction module includes:

[0081] The data potential energy extraction submodule obtains the water potential and electrical conductivity of the deep layer in the rural soil environmental monitoring data, performs hierarchical division and numbering on adjacent layer data, calculates the interlayer gradient change rate based on the depth and water potential distribution, and generates the layer potential energy value.

[0082] Obtain rural soil environmental monitoring data from monitoring profile C, specifically from four depth layers ( , , , The water potential and conductivity data collected at timestamp 15:00:00 are shown in Table 2. The data from these four adjacent levels were then hierarchically divided and numbered. Depth is level 1. Level 2 Level 3 For level 4, the gradient change rate between layers is calculated based on the depth of multiple layers and the corresponding water potential values. First, the gradient change rate between level 2 and level 1 is calculated, and the water potential of level 2 is used. Water potential at level 1 and the depth of level 2 Depth of level 1 The calculation process involves the potential difference between the two water layers. Divide by the depth difference between the two layers The gradient change rate is obtained. Similarly, the gradient change rate between level 3 and level 2 is calculated by calling the water potential of level 3. With depth Calculate the water potential difference Divide by depth difference The gradient change rate is obtained. Finally, the gradient change rate between level 4 and level 3 is calculated, and the water potential of level 4 is called. With depth Calculate the water potential difference Divide by depth difference The gradient change rate is obtained. The three gradient rates of change , , Combined, the hierarchical potential energy value of the monitoring profile C is generated.

[0083] Table 2: Initial Data Table for Monitoring Profile C

[0084] Hierarchical numbering Node depth (cm) Soil water potential (kPa) Soil electrical conductivity (dS / m) 1 20 12.3 0.75 2 40 15.5 0.81 3 60 25.8 1.25 4 80 28.2 0.92

[0085] As shown in Table 2, this table presents the initial monitoring data of soil water potential and electrical conductivity collected at four different depth levels in monitoring profile C.

[0086] The potential energy difference determination submodule calls the hierarchical potential energy value, compares the potential energy elements of the upper and lower layers and performs a judgment based on the potential energy difference threshold, identifies abnormal nodes where the lower layer potential energy is greater than the upper layer and the difference exceeds the threshold, and clusters to calculate the average potential energy difference, generating a cross-layer potential energy difference set.

[0087] The generated hierarchical potential energy value is invoked, which is the gradient change rate between the three layers. , and The data is then processed in conjunction with the original water potential data. First, the water potential values ​​of adjacent upper and lower layers are compared. The water potential of layer 2 is... Water potential greater than level 1 The difference is Level 3 water potential Water potential greater than level 2 The difference is Level 4 water potential Water potential greater than level 3 The difference is Next, a potential energy difference threshold is set. This threshold is set based on a preset multiple of the average potential energy difference between adjacent layers in the current profile. First, the average potential energy difference is calculated, i.e. The preset multiplier is determined based on soil type and the stability of the original data. For the loam in this region, differences in its internal structure cause fluctuations in water potential; therefore, the preset multiplier is set to [value missing]. The potential energy difference threshold is then calculated as follows: The difference in water potential between multiple layers is compared with this threshold, and the difference between layer 2 and layer 1 is... Less than The result is considered normal; the difference between level 3 and level 2 is [value missing]. Greater than Furthermore, since the lower-level potential energy is greater than that of the upper-level potential energy, the node corresponding to level 3 is marked as an abnormal node. The difference between level 4 and level 3 is... Less than If the value is determined to be normal, then the potential energy difference values ​​of the nodes marked as abnormal will be clustered. Currently, there is only one abnormal difference value. Without the need for clustering calculations, this value can be directly used as the result to generate a set of cross-layer potential energy differences.

[0088] The abnormal flow calculation submodule calls the interlayer potential energy difference set, derives the soil permeability coefficient based on the interlayer potential energy difference and gradient change characteristics, calculates the potential energy driving direction and water potential transport rate, analyzes the time series distribution of abnormal paths based on the rate change trend, and generates abnormal soil water potential flow data.

[0089] The generated set of cross-layer potential energy differences is invoked, which includes an anomalous potential energy difference. This value corresponds to level 2 ( ) and Level 3 ( Between these layers, based on this interlayer potential energy difference. The soil permeability coefficient is derived from the gradient variation characteristics of this profile, and the average water potential gradient of this profile is: The gradient between abnormal layers is It is much higher than the average, which indicates that in to There are preferred flow paths between depths; the baseline permeability coefficient for this soil type is set as follows. The coefficient is adjusted based on the magnitude by which the gradient exceeds the average value. The adjustment rule is that for every gradient exceeding the average value... Increased permeability The current gradient exceeds the value. Therefore, the permeability coefficient adjustment ratio is The derived permeability coefficient is Next, the driving direction of potential energy and the water potential transport rate are calculated. Since the water potential in level 3 is greater than that in level 2, the driving direction is determined to be from top to bottom. The water potential transport rate is estimated by multiplying the derived permeability coefficient by the water potential gradient. Finally, based on the rate values ​​of this layer in the original database over the past three hours (12:00, 13:00, 14:00) (respectively... , , ) and the currently calculated By comparing and analyzing the rate change trend, it was found that the rate decreased at 15:00. The jump marks this point in time as an anomalous event, and the anomalous level ( to ), Derivation of permeability coefficient ( ), Drive direction (down), Current transmission rate ( The data is integrated with the time series distribution (rate array from 12:00 to 15:00) to generate soil water potential anomaly flow data.

[0090] Specifically, such as Figure 2 and Figure 5 As shown, the path inversion modeling module includes:

[0091] The soil water potential data integration submodule acquires abnormal soil water potential flow data, formats and groups the time series at different depth levels, detects flow rate and gradient differences, removes discontinuous samples and performs normalization processing, and generates soil water potential distribution data.

[0092] Obtain the generated soil water potential anomaly flow data, specifically at monitoring profile C at 15:00. to Interlayer anomaly data records {anomaly layer: 40cm to 60cm, derived permeability coefficient:} Drive direction: downward, transmission rate: The time series is [0.15, 0.18, 0.21, 0.425], and combined with continuous time series data collected from other depth levels within the same time period, as shown in Table 3, the data from different depth levels are first grouped and statistically analyzed according to "depth level-time" to detect differences in flow rate and gradient. to Taking the data from layer 15:00 as an example, its flow rate is: The water potential gradient is , and abnormal layers to rate and gradient The rate difference is compared to the previous one. The gradient difference is Set the difference threshold as the rate change exceeds or gradient change exceeds Since both exceeded the threshold, the significance of the abnormal data was confirmed. Next, the temporal continuity of the data samples was checked, and a standard collection interval was set. If, for a certain level, the data immediately follows 16:00 after 14:00, the 16:00 data points are discarded due to time discontinuity. Normalization is then performed after confirming sample continuity. to Hierarchical rate time series For example, its minimum value is The maximum value is The rate value at 14:00 Normalization calculation is performed, the process is as follows: Perform this operation on all time-point data at each level to generate soil water potential distribution data.

[0093] Table 3: Multi-level Time Series Data Table

[0094] Time point Depth layer (cm) Flow rate (cm / hr) Water potential gradient (kPa / cm) 12:00 20-40 0.14 0.15 13:00 20-40 0.16 0.15 14:00 20-40 0.17 0.16 15:00 20-40 0.22 0.16 12:00 40-60 0.15 0.17 13:00 40-60 0.18 0.19 14:00 40-60 0.21 0.22 15:00 40-60 0.425 0.515

[0095] As shown in Table 3, this table lists the flow rate and water potential gradient monitoring data of two depth levels at four consecutive time points, which will be used for subsequent data integration and analysis.

[0096] The direction vector extraction submodule calls soil water potential distribution data, uses a random forest model to perform feature splitting on multi-layer water potential gradient data, extracts multi-dimensional potential energy change features and calculates direction vector weights, and clusters and adjusts node directions based on feature importance to generate a multi-level potential energy direction array.

[0097] The system calls upon the generated soil water potential distribution data, which contains normalized water potential gradient time series from multiple monitoring profiles at different depth levels. The system performs feature splitting on the data by constructing a series of decision trees. Specifically, it randomly extracts subsets of data samples and feature subsets; for example, it extracts feature {profile C in...}. to The gradient of the profile D in to The gradient of the gradient is used to find an optimal split point, such as determining the gradient of the profile C. to Is the gradient normalized value greater than ? Based on this judgment, the samples are divided into two groups, and this process is repeated until a complete decision tree is formed. By constructing hundreds of such decision trees, multidimensional potential energy change features are extracted and the direction vector weights of the features are calculated. The calculation logic of the weights is to count the number of times each feature (such as the gradient of a specific level) is used to split nodes in the decision tree and the amount of impurity reduction brought about by each split, and then sum the reductions by weight. For example, after calculation, profile C in to The importance score of the vertical gradient feature is While section C and the adjacent section D are in The importance of horizontal gradient features in depth is The total importance of the remaining features is The importance score is used as the weight of the direction vector. Node directions are clustered based on feature importance weights, with the vertical gradient weight being higher than the weight of the other features. Nodes (such as those in section C) to The nodes are classified into the "vertical priority flow" category, the nodes with higher horizontal gradient weights are classified into the "lateral penetration" category, and the directional vector weights of the nodes in the same category are averaged and adjusted to generate a multi-level potential energy direction array.

[0098] The infiltration path calculation submodule obtains the potential energy change sequence by accumulating the node potential energy difference layer by layer according to the multi-level potential energy direction array, calculates the multi-level potential energy cumulative gradient and determines the path connection relationship, statistically analyzes the distribution density and diffusion range, and generates a water infiltration path distribution map.

[0099] Based on the generated multi-level potential energy direction array, which indicates the main driving direction and intensity of water flow between monitoring nodes, the system accumulates the original potential energy difference (water potential difference value) between nodes layer by layer, starting from the surface layer (level 1), to obtain a sequence of potential energy changes. For example, the inter-layer water potential differences in profile C are as follows: , , Then the accumulated potential energy change sequence is: The unit is Then, the cumulative potential gradient of multiple layers is calculated, that is, the cumulative potential energy change of each layer is divided by the corresponding total depth, for example, in... Depth (end position of level 3), cumulative gradient is (depth difference is) The path connectivity is determined based on the direction vector weights calculated from the direction array, and a connection weight threshold is set. When the direction vector weights between two adjacent nodes (such as the vertical gradient weights between level 2 and level 3 in profile C) When the value exceeds the threshold, the two nodes are determined to be connected on the infiltration path. The system traverses the monitoring profile, connects the connected node pairs that meet the conditions, forming an infiltration path network. Finally, the network is statistically analyzed in farmland plots (divided into...). The distribution density and diffusion range in the grid are determined by calculating the number of paths passing through each grid. If a grid has more than 5 paths passing through it, it is defined as a high-density area with a risk level of "high". The diffusion range is the total area of ​​the grids containing the infiltration paths. The density and risk level are visualized and rendered on the map to generate a water infiltration path distribution map.

[0100] Specifically, such as Figure 2 and Figure 6 As shown, the multidimensional diffusion modeling module includes:

[0101] The moisture path receiving submodule collects the coordinate values ​​of path nodes and the moisture content values ​​of nodes based on the moisture infiltration path distribution map, compares them with the offset threshold and records the offset, calculates the moisture content gradient of adjacent nodes based on the distance between nodes, adjusts the moisture content values ​​of nodes according to the offset, and generates the path gradient distribution.

[0102] Use the generated water infiltration path distribution map to extract the paths. The key node coordinate data and soil volumetric moisture content values ​​of the nodes are shown in Table 4. First, statistical analysis is performed on this set of node moisture content data to determine the offset threshold. Then, the original sampling sequence of moisture content of the nodes in the monitoring area at the same time is collected. Sequence length Calculate the median for The quantiles used to determine high humidity anomalies are set as follows: Quantity, that is, the digit after sorting the sequence. Bit data ( ), corresponding value is Therefore, Set as the offset threshold, and then compare the moisture content of path nodes A001, A002, and A003 one by one with this threshold. Node A001 ( ) and A002 ( All values ​​are less than the threshold, and the record offset is... The moisture content of node A003 is Exceeding the threshold Calculate the offset Next, the moisture content gradient between adjacent nodes is calculated based on the vertical distance (i.e., depth difference) between nodes, and the distance between A002 and A001. Moisture content difference ,gradient The distance between A003 and A002 Original moisture content difference ,gradient At this point, the moisture content of node A003, which exhibits an anomaly, is adjusted based on the recorded offset, and an adjustment coefficient is set. (Based on historical sensor drift data), calculate the adjustment value. And based on the adjusted values, the gradient at A003 is recalculated, and the adjusted gradient is... Finally, the adjusted node moisture content and gradient values ​​are integrated to generate the path gradient distribution.

[0103] Table 4: Calculation Table of Moisture Content and Gradient at Path Nodes

[0104] Node number Depth (cm) Original moisture content (%) Offset threshold (%) Offset (%) Initial gradient (% / cm) A001 20 22.1 34.2 0.0 - A002 40 28.5 34.2 0.0 0.320 A003 60 36.2 34.2 +2.0 0.385

[0105] As shown in Table 4, this table lists the paths in detail. The monitoring data of the three depth nodes, the threshold comparison results, and the preliminary calculated gradient values ​​are shown, where a positive offset indicates that the threshold is exceeded.

[0106] The moisture situation construction submodule calls the path gradient distribution, fuses multi-source sensor readings for time alignment and weighting, calculates the gradient correction based on the humidity deviation benchmark, superimposes the correction to adjust the path gradient distribution and calculates the spatial humidity distribution to obtain the spatial humidity distribution.

[0107] The generated path gradient distribution, especially the adjusted gradient of node A003, is called. and moisture content Simultaneously, the high-precision TDR (Time Domain Reflectometer) sensor deployed at that location was acquired at the same timestamp. readings After time alignment, multi-source data fusion is performed, and the weights of the TDR sensors are set. Weights of path deduction values Calculate the fusion moisture content Next, the humidity deviation baseline value is calculated by selecting a sequence of differences between sensor readings and path projection values ​​from the past hour (sampling once every 10 minutes, for a total of 6 points). Set the sliding window size to 6, and calculate the average value of the difference within the window. and standard deviation The benchmark value is set as the mean plus twice the standard deviation, i.e. The actual difference at the current moment is This value is less than the baseline value. This indicates that the deviation is within the normal fluctuation range, therefore a small correction coefficient is assigned when calculating the gradient correction amount. If the deviation exceeds the reference value, the correction factor will increase exponentially. This correction amount will be added to the original gradient to obtain the final gradient. Finally, based on the corrected gradient and fused moisture content, the Kriging interpolation method is used to calculate the moisture content around the node. The spatial humidity distribution within the radius is generated, producing a spatial humidity distribution quantity that includes three-dimensional coordinates and corresponding humidity values.

[0108] The diffusion boundary generation submodule analyzes the changes in spatial humidity and pollutant concentration based on the spatial humidity distribution and calculates the diffusion boundary threshold according to the concentration change benchmark. It then calls the diffusion boundary threshold to divide the humidity distribution segment and records the diffusion edge position of the segment, generating a diffusion boundary spatial model.

[0109] Based on the generated spatial humidity distribution, the focus is on analyzing the area where node A003 is located. The relationship between the spatial humidity field (depth) and the temporal variation of nitrate nitrogen pollutant concentration at this location was obtained, and the time series of pollutant concentration at this node over the past four monitoring times was acquired. Calculate the concentration change difference sequence between adjacent time points. Set the sliding window to 3, calculate the statistical characteristics of the difference amplitude, including the mean. Standard deviation Based on the calculation rules for the concentration change benchmark value, the benchmark value is set as follows: This baseline value is the critical indicator for determining whether a pollutant has significantly diffused. The diffusion boundary threshold is calculated, and the latest concentration change at this point is... (Right now (close to the benchmark value) Based on the spatial humidity distribution, when the humidity is greater than And the rate of concentration change is close to or exceeds The system delineates the boundaries of the region, scans spatial data, and identifies coordinates. The concentration change at that location did not exceed the standard, but at the coordinate... The projected concentration change at the location is And the humidity is Therefore, the coordinate point is marked as the diffusion edge position. Connecting such edge points forms a closed or semi-closed surface, which is the physical boundary of the pollutant transported with water, thus generating a diffusion boundary space model.

[0110] Specifically, such as Figure 2 and Figure 7 As shown, the anomaly detection and monitoring module includes:

[0111] The abnormal feature extraction submodule calls the diffusion boundary space model to obtain abnormal indicators in the monitoring area, extracts the indicator values ​​and timestamps of monitoring nodes that deviate from the normal range boundary, calculates the duration of the current time and the detection start time, performs cross-matching to classify abnormal types, and generates an abnormal indicator classification statistics set.

[0112] The generated diffusion boundary space model is invoked, which defines the three-dimensional distribution range of pollutants in the soil. First, the data of monitoring nodes within this range at the current time (16:00) are traversed. Taking monitoring node N-45 as an example, its soil nitrate nitrogen concentration value is obtained. and soil moisture content A normal range of indicator boundaries is set, which is determined based on the statistical characteristics of the historical monitoring data of the land parcel over the past three years. The mean of the original data is [value missing]. The standard deviation is The upper limit of the normal range is set as the mean plus 2.5 times the standard deviation, i.e. Similarly, the normal upper limit for moisture content is set as follows: The concentration at node N-45 was compared. Exceeding the limit Moisture content Exceeding the limit The node was identified as having a dual anomaly. Its indicator value and current timestamp were then extracted, and continuous records of the node in the original database were retrieved to determine the initial time of the anomaly. Data showed that the concentration at this node first exceeded [a certain threshold] at 13:30 on that day. Calculate the duration between the current time and the detection start time, i.e. Next, the extracted abnormal indicators are cross-matched to classify the abnormality types. A classification rule table is set. If "concentration exceeds the standard and moisture content exceeds the standard", it is defined as "leaching and diffusion type". If "concentration exceeds the standard but moisture content is normal", it is defined as "drought accumulation type". Based on the data characteristics of N-45, it is classified as "leaching and diffusion type". This process is executed on abnormal nodes in the region. As shown in Table 5, the node number, coordinates, abnormal value, duration and classification type are integrated to generate an abnormal indicator classification statistical set.

[0113] Table 5: Statistical Table of Abnormal Node Indicators in the Monitoring Area

[0114] Node number Depth coordinates (cm) Nitrate nitrogen concentration (mg / L) Soil moisture content (%) Duration (h) Exception types N-45 60 15.6 35.4 2.5 Leaching and diffusion type N-48 60 14.8 34.2 2.0 Leaching and diffusion type N-52 40 13.5 18.5 4.5 drought accumulation type

[0115] As shown in Table 5, this table lists in detail the key abnormal nodes identified at the current monitoring time and their corresponding multiple indicator data, which serve as the basis for subsequent situation assessment.

[0116] The abnormal situation assessment submodule calls the abnormal indicator classification statistics set, extracts the spatial coverage area value of the abnormality, calculates the ratio of the abnormal coverage area to the total monitored area, obtains the abnormal area ratio coefficient, calculates the rate of change of the ratio coefficient between adjacent time points, and generates the abnormal spread trend quantity.

[0117] The generated anomaly indicator classification statistics set is invoked. First, the spatial coverage of the high-risk category "leaching and diffusion type" is extracted. Then, using a gridded calculation method, the entire monitoring area is divided into... There are 10 cell grids, each with an area of ​​1. The total monitoring area is The number of grid cells containing anomalous nodes and those exceeding the value limit after interpolation was counted, and it was calculated that at the current time of 16:00, the "leaching and diffusion type" anomaly covered [a certain area / region]. Each grid cell is used to extract the spatial coverage area value of anomalies. Calculate the ratio of the abnormal coverage area to the total monitored area, i.e. Obtain the percentage coefficient of abnormal areas Next, retrieve the percentage coefficient record from the previous monitoring time of 15:00, at which time the number of covered grids was... Number, percentage coefficient Calculate the rate of change of the proportion coefficient at adjacent time points, with a time interval of . rate of change This positive value indicates that the anomalous area is expanding. Simultaneously, analyzing the change in the proportion of "drought-accumulated" anomalies, if the rate of change is negative, it indicates that this type is receding or transforming into another type. The current proportion coefficients of multiple types are then analyzed. Historical proportion coefficient and rate of change By combining these vectors, a vector reflecting the speed and direction of pollution diffusion is constructed, generating a trend quantity of abnormal diffusion patterns.

[0118] The response window calculation submodule extracts the diffusion rate parameter and the anomaly type weight coefficient based on the abnormal diffusion trend, calculates the correlation value between the diffusion rate and the weight coefficient, analyzes the allowable response time to obtain the response time window interval, summarizes it with the anomaly index classification statistics set, and generates anomaly index statistical results.

[0119] Based on the generated abnormal diffusion trend, the diffusion rate parameter is extracted. In addition to anomaly types, a weighting coefficient is set for "leaching and diffusion type" anomalies. This coefficient is determined based on the potential hazard of pollutants to groundwater infiltration. Referring to environmental risk assessment standards, nitrate nitrogen has an extremely high migration risk under supersaturated water conditions, therefore a weighting coefficient is set for it. The weighting coefficient for "drought accumulation type" is set to... Calculate the correlation between the diffusion rate and the weighting coefficient, i.e., the risk growth index. This index reflects the weighted rate of environmental risk escalation. Next, the allowable response time is analyzed to obtain the response time window interval, and the maximum tolerable anomaly percentage threshold for the monitoring area is set as follows: (Right now (area), currently accounting for . The remaining safety buffer is Dividing this buffer amount by the risk growth index yields the expected time to reach the critical state. Starting from the current time 16:00, the calculated response deadline is... Therefore, the determined response time window interval is Finally, the information in this window is summarized with the specific node locations and concentration values ​​in the abnormal indicator classification statistics set to clarify which areas need to be prioritized for processing within this time window, and abnormal indicator statistics results are generated.

[0120] Please see Figure 8 Based on the same inventive concept, a smart digital rural environmental monitoring method based on big data is also proposed. This method is implemented using the aforementioned smart digital rural environmental monitoring system based on big data and includes the following steps:

[0121] S1: Collect voltage signals, moisture conductivity signals, and micro-pressure signals through sensor nodes, perform standardized calculations, bind layered identifiers to the depth of the data collection nodes, calculate slope angles, and generate rural soil environmental monitoring data.

[0122] S2: Call rural soil environmental monitoring data, extract potential energy of adjacent layers, calculate potential energy difference, and perform adjustment operation when the potential energy of the lower layer is greater than that of the upper layer and the difference exceeds the preset threshold to generate abnormal soil water flow data.

[0123] S3: Call soil water potential anomaly flow data, use random forest model to extract multi-level direction vectors and potential energy changes, accumulate potential energy changes layer by layer, calculate infiltration propagation paths according to depth, and generate water infiltration path distribution map.

[0124] S4: Call the water infiltration path distribution map, use sensor fusion algorithm for real-time monitoring, calibration and analysis, combine spatial distribution characteristics to analyze the diffusion trend of water and pollutants, and construct a diffusion boundary spatial model;

[0125] S5: Call the diffusion boundary space model to extract abnormal indicators and state duration, classify and statistically analyze abnormal types, calculate the proportion of abnormal areas to the total monitoring area, analyze the development trend based on the trend of proportion change, calculate the response time window, and generate statistical results of abnormal indicators.

[0126] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A smart digital rural environmental monitoring system based on big data, characterized in that: include: The stratified water potential acquisition module collects voltage signals, water conductivity signals and micro-pressure signals through sensor nodes, performs standardized calculations, binds the stratification identifiers to the depth of the acquisition nodes, calculates the slope angle, generates rural soil environmental monitoring data, and transmits it to the gradient inverse correction module. The gradient inverse correction module calls the rural soil environment monitoring data, extracts the potential energy of adjacent layers, calculates the potential energy difference, and performs inverse correction when the potential energy of the lower layer is greater than that of the upper layer and the difference exceeds a preset threshold, generating soil water potential abnormal flow data, which is then transmitted to the path inverse modeling module. The path inversion modeling module calls the soil water potential anomaly flow data, uses a random forest model to extract multi-level direction vectors and potential energy changes, accumulates the potential energy changes layer by layer, calculates the infiltration propagation path according to depth, generates a water infiltration path distribution map, and transmits it to the multi-dimensional diffusion modeling module. The multidimensional diffusion modeling module calls the moisture infiltration path distribution map, uses a sensor fusion algorithm for real-time monitoring, calibration and analysis, combines spatial distribution characteristics to analyze the diffusion trend of moisture and pollutants, constructs a diffusion boundary spatial model, and transmits it to the anomaly detection and monitoring module. The rural soil environmental monitoring data includes stratified soil water potential distribution, soil moisture content gradient, and soil electrical conductivity characteristic values. The soil water potential abnormal flow data includes vertical water potential difference, abnormal water flow direction vector, and water flow intensity level. The water infiltration path distribution map includes infiltration path distribution structure, soil water potential accumulation interval, and water flow propagation direction marker. The diffusion boundary spatial model includes pollutant diffusion boundary range, water lateral migration area, and multidimensional diffusion rate distribution.

2. The smart digital rural environmental monitoring system based on big data according to claim 1, characterized in that, The stratified water potential acquisition module includes: The signal acquisition submodule acquires the voltage signal, moisture conductivity signal and micro-pressure signal output by the layered sensing nodes, performs comparison based on the time synchronization benchmark, calculates the instantaneous difference value after normalizing the signal amplitude, filters out abnormal fluctuation points that exceed three times the preset standard deviation, and calculates the weighted average value after determining the weight based on the signal-to-noise ratio of the signal, generating a multi-source signal fusion value. The node identification binding submodule collects node depth data for hierarchical indexing based on the multi-source signal fusion value, establishes a mapping table in ascending order of depth, binds signal, layer number and node spatial coordinate data, calculates the signal difference between adjacent nodes and analyzes the gradient change ratio, and generates the stratified water potential gradient coefficient. The slope angle calculation submodule, based on the layered water potential gradient coefficient, calls the node spatial coordinate data, calculates the horizontal projection distance and vertical height difference, and calculates the slope angle using the arctangent function based on the ratio of the two, generating rural soil environmental monitoring data.

3. The smart digital rural environmental monitoring system based on big data according to claim 2, characterized in that, The gradient inverse correction module includes: The data potential energy extraction submodule obtains the water potential and electrical conductivity of the deep layer in the rural soil environment monitoring data, performs hierarchical division and numbering on adjacent layer data, calculates the interlayer gradient change rate based on the depth and water potential distribution, and generates layer potential energy values. The potential energy difference determination submodule calls the hierarchical potential energy value, compares the potential energy elements of the upper and lower layers and performs a judgment based on the potential energy difference threshold, identifies abnormal nodes where the lower layer potential energy is greater than the upper layer and the difference exceeds the threshold, and clusters to calculate the average potential energy difference to generate a cross-layer potential energy difference set. The abnormal flow calculation submodule calls the interlayer potential energy difference set, derives the soil permeability coefficient based on the interlayer potential energy difference and gradient change characteristics, calculates the potential energy driving direction and water potential transmission rate, analyzes the time series distribution of abnormal paths based on the rate change trend, and generates abnormal soil water potential flow data.

4. The smart digital rural environmental monitoring system based on big data according to claim 3, characterized in that, The potential energy difference threshold is a value dynamically determined based on the average potential energy difference between the upper and lower layers within a preset range. When the difference between the potential energy of the lower layer and the potential energy of the upper layer exceeds the threshold, it is identified as an abnormal node. The calculation of the interlayer gradient change rate refers to obtaining the change trend between depth and water potential by calculating the difference between data of adjacent layers, and determining the change rate between each layer based on the relative distribution relationship between depth and water potential.

5. The smart digital rural environmental monitoring system based on big data according to claim 3, characterized in that, The path inversion modeling module includes: The soil water potential data integration submodule acquires the abnormal flow data of soil water potential, formats and groups the time series of different depth levels, detects the differences in flow rate and gradient, removes samples with discontinuous time, and performs normalization processing using the Min-Max normalization method to generate soil water potential distribution data. The direction vector extraction submodule calls the soil water potential distribution data, uses a random forest model to perform feature splitting on the multi-layer water potential gradient data, extracts multi-dimensional potential energy change features and calculates direction vector weights, and clusters and adjusts the node directions according to feature importance to generate a multi-level potential energy direction array. The infiltration path calculation submodule, based on the multi-level potential energy direction array, accumulates the node potential energy difference layer by layer to obtain the potential energy change sequence, calculates the multi-level potential energy cumulative gradient and determines the path connection relationship, statistically analyzes the distribution density and diffusion range, and generates a water infiltration path distribution map.

6. The smart digital rural environmental monitoring system based on big data according to claim 5, characterized in that, The multidimensional diffusion modeling module includes: The moisture path receiving submodule collects the coordinate values ​​of path nodes and the moisture content values ​​of nodes based on the moisture infiltration path distribution map, compares them with the offset threshold and records the offset, calculates the moisture content gradient of adjacent nodes based on the distance between nodes, adjusts the moisture content values ​​of nodes according to the offset, and generates the path gradient distribution. The moisture situation construction submodule calls the path gradient distribution, fuses multi-source sensor readings for time alignment and weighting, calculates gradient correction based on humidity deviation benchmark, superimposes correction to adjust path gradient distribution and calculates spatial humidity distribution to obtain spatial humidity distribution. The diffusion boundary generation submodule analyzes the changes in spatial humidity and pollutant concentration based on the spatial humidity distribution and calculates the diffusion boundary threshold according to the concentration change benchmark value. It then calls the diffusion boundary threshold to divide the humidity distribution segment and records the diffusion edge position of the segment, generating a diffusion boundary spatial model.

7. The smart digital rural environmental monitoring system based on big data according to claim 1, characterized in that, The anomaly detection and monitoring module calls the diffusion boundary space model to extract anomaly indicators and state duration, classifies and statistically analyzes anomaly types, calculates the proportion of anomaly areas to the total monitoring area, analyzes the development trend based on the trend of proportion changes, calculates the response time window, and generates statistical results of anomaly indicators. The statistical results of the abnormal indicators include statistics on the classification of abnormal types, the changing trend of the proportion of abnormal areas, and the threshold of the response time window.

8. The smart digital rural environmental monitoring system based on big data according to claim 7, characterized in that, The anomaly detection and monitoring module includes: The abnormal feature extraction submodule calls the diffusion boundary space model to obtain abnormal indicators in the monitoring area, extracts the indicator values ​​and timestamps of the monitoring nodes that deviate from the normal range boundary, calculates the duration of the current time and the detection start time, performs cross-matching to classify abnormal types, and generates an abnormal indicator classification statistics set. The abnormal situation assessment submodule calls the abnormal indicator classification statistics set, extracts the spatial coverage area value of the abnormality, calculates the ratio of the abnormal coverage area to the total monitoring area, obtains the abnormal area ratio coefficient, calculates the rate of change of the ratio coefficient between adjacent time points, and generates the abnormal spread trend quantity. The response window calculation submodule extracts the diffusion rate parameter and the anomaly type weight coefficient based on the anomaly diffusion trend quantity, calculates the correlation value between the diffusion rate and the weight coefficient, analyzes the allowable response time to obtain the response time window interval, summarizes it with the anomaly indicator classification statistics set, and generates anomaly indicator statistical results.

9. A smart digital rural environmental monitoring method based on big data, characterized in that, The implementation of the smart digital rural environmental monitoring system based on big data according to any one of claims 1-8 includes the following steps: S1: Collect voltage signals, moisture conductivity signals and micro-pressure signals through sensor nodes, perform standardized calculations, bind layered identifiers to the depth of the data collection nodes, calculate slope angles, and generate rural soil environmental monitoring data. S2: Call the rural soil environment monitoring data, extract the potential energy of adjacent layers, calculate the potential energy difference, and when the potential energy of the lower layer is greater than that of the upper layer and the difference exceeds the preset threshold, perform the adjustment operation to generate abnormal soil water flow data. S3: Call the soil water potential anomaly flow data, use the random forest model to extract multi-level direction vectors and potential energy changes, accumulate the potential energy changes layer by layer, calculate the infiltration propagation path according to depth, and generate a water infiltration path distribution map. S4: Call the moisture infiltration path distribution map, use sensor fusion algorithm for real-time monitoring, calibration and analysis, combine spatial distribution characteristics to analyze the diffusion trend of moisture and pollutants, and construct a diffusion boundary spatial model; S5: Call the diffusion boundary space model, extract abnormal indicators and state duration, classify and statistically analyze abnormal types, calculate the proportion of abnormal areas to the total monitoring area, analyze the development trend based on the trend of proportion change, calculate the response time window, and generate statistical results of abnormal indicators.

Images