A soil moisture dynamic monitoring system and method based on the Internet of Things

Abstract

The application discloses a kind of based on soil moisture dynamic monitoring system and method of Internet of Things, it is related to the technical field of monitoring connection, including real-time acquisition soil moisture data, using hilbert transform obtains phase feature and obtains phase difference between measuring points, carries out soil state evaluation;It is decomposed by wavelet transform, and periodic characteristic extraction is carried out using time-frequency joint analysis method, obtains abnormal measuring point positioning information;Based on abnormal measuring point positioning information, using Dijkstra algorithm isolates abnormal measuring point and selects shortest path, obtains water monitoring network structure;Based on water monitoring network structure dynamic adjustment wavelet transform decomposition parameter and periodic characteristic extraction parameter, output soil moisture dynamic monitoring report.The application dynamically selects shortest path by using Dijkstra algorithm, and by redundant path construction standby communication channel, improves data transmission stability.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of Internet of Things monitoring technology, and in particular to a soil moisture dynamic monitoring system and method based on Internet of Things. BACKGROUND

[0002] Soil moisture monitoring is an important technology used in agriculture, environmental monitoring and ecological research. Most current soil moisture monitoring methods rely on fixed-point sampling or use sensors such as time-domain reflectometry and frequency-domain reflectometry for single-point static measurement. However, with the development of Internet of Things technology, wireless sensor networks and remote data transmission technology are gradually being applied in the field of soil moisture monitoring for real-time data collection and transmission. Most current technologies use fixed thresholds or simple statistical analysis for anomaly detection, and use static network topology for data transmission to meet monitoring needs.

[0003] However, the current methods have limitations in dynamic environment adaptability. Traditional anomaly detection methods use fixed thresholds or single statistical indicators, which sometimes cannot effectively distinguish between device failures and real soil moisture anomalies, resulting in a high false positive rate. In addition, static network topology lacks dynamic optimization of link quality and node state, and in complex terrain or interference environments, data transmission delays and interruptions are likely to occur, affecting the real-time and reliability of monitoring. SUMMARY

[0004] In view of the above existing problems, the present application is proposed.

[0005] Therefore, the present application provides a soil moisture dynamic monitoring method based on Internet of Things to solve the problems of low anomaly detection accuracy caused by reliance on fixed thresholds and poor static network topology optimization and transmission delays and interruptions.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] In a first aspect, the present application provides a soil moisture dynamic monitoring method based on Internet of Things, which includes real-time collection of soil moisture data, use of Hilbert transform to obtain phase characteristics and obtain phase difference between measurement points, and soil state evaluation.

[0008] Time window data smoothing is used to real-time calibrate soil moisture data, and potential abnormal measurement points are analyzed in combination with soil state evaluation analysis, and an abnormal diagnosis report is output.

[0009] Wavelet transform is used for decomposition, and a time-frequency joint analysis method is used for periodic feature extraction to obtain abnormal measurement point positioning information.

[0010] Based on the abnormal measurement point positioning information, the abnormal measurement point is isolated and the shortest path is selected using Dijkstra algorithm, and the water monitoring network structure is obtained;

[0011] Based on the water monitoring network structure, the wavelet transform decomposition parameters and the periodic feature extraction parameters are dynamically adjusted, and the soil water dynamic monitoring report is output.

[0012] As a preferred scheme of the soil water dynamic monitoring method based on the Internet of Things, wherein: the Hilbert transform is used to obtain the phase feature and the phase difference between the measurement points, and the soil state evaluation is carried out, and the specific steps are as follows,

[0013] Based on the soil water data, the Hilbert transform is applied to obtain the instantaneous phase of the signal feature data, and the spatial distribution characteristics of the signal feature data are obtained through the phase difference operation between the measurement points, and the phase difference data set is integrated and generated;

[0014] Based on the phase difference data set, the water feature is extracted, the dry probability is obtained by using the weighted voting method and the threshold judgment method, and the reference threshold is set through the adaptive adjustment strategy, and the soil state evaluation result is output.

[0015] As a preferred scheme of the soil water dynamic monitoring method based on the Internet of Things, wherein: the soil water data is calibrated in real time by using time window data smoothing, and the potential abnormal measurement points are analyzed in combination with the soil state evaluation, and the abnormal diagnosis report is output, and the specific steps are as follows,

[0016] The measurement deviation of the soil water data is corrected by using time window data smoothing, and the calibrated soil water data is output.

[0017] Based on the soil state evaluation result and the calibrated soil water data, the difference analysis is carried out on the adjacent measurement point data by using the dynamic threshold method, and the abnormal measurement point is identified;

[0018] The abnormal index is obtained by using the statistical control method, and the position and the abnormal degree of the abnormal measurement point are identified by using the adjacent point grouping method, and the abnormal diagnosis report is output.

[0019] As a preferred scheme of the soil water dynamic monitoring method based on the Internet of Things, wherein: the decomposition is carried out by using the wavelet transform, and the periodic feature extraction is carried out by using the time-frequency joint analysis method, and the abnormal measurement point positioning information is obtained, and the specific steps are as follows,

[0020] The wavelet transform is used to carry out 5-layer wavelet packet decomposition on the measurement point signal, the energy entropy value of each sub-band is obtained, and the energy distribution abnormal sub-band frequency band is output by using the dynamic threshold method;

[0021] The time-frequency joint analysis method is used to capture the periodic characteristics in the sub-band frequency band with abnormal energy distribution, and output a periodic interference signal;

[0022] The abnormal phase difference is identified by using the change degree of the phase difference of adjacent measurement points in combination with the sub-band frequency band with abnormal energy distribution and the periodic interference signal, and the positioning information of the abnormal measurement point is obtained.

[0023] As a preferred scheme of the soil moisture dynamic monitoring method based on the Internet of Things, wherein: based on the positioning information of the abnormal measurement point, the Dijkstra algorithm is used to isolate the abnormal measurement point and select the shortest path, and the water monitoring network structure is obtained, and the specific steps are as follows,

[0024] The initial water monitoring network structure is constructed by the near neighbor connection method, and the threshold filtering is used for measurement point screening and link cleaning, and the purified initial water monitoring network structure is output;

[0025] Based on the purified initial soil water monitoring network structure, real-time link parameter measurement is performed, and the comprehensive quality is obtained, the threshold comparison method is used to find out the problematic link, and the link information table is output;

[0026] According to the link information table, the Dijkstra algorithm is used to obtain the shortest path from each measurement point to the gateway, and real-time performance monitoring is performed, and the path optimization parameter set is output;

[0027] The standby communication channel is constructed by using the redundant path, and the purified initial water monitoring network structure is dynamically adjusted by using the path optimization parameter set, and the water monitoring network structure is generated.

[0028] As a preferred scheme of the soil moisture dynamic monitoring method based on the Internet of Things, wherein: based on the water monitoring network structure, the wavelet transform decomposition parameter and the periodic feature extraction parameter are dynamically adjusted, and the soil moisture dynamic monitoring report is output, and the specific steps are as follows,

[0029] Based on the water monitoring network structure, the measurement point density and the link stability score are obtained, and the health status of the water monitoring network structure is evaluated;

[0030] According to the health status of the water monitoring network structure, the wavelet base and the decomposition layer number are selected, and the wavelet packet energy spectrum analysis method is used to perform signal decomposition and time-frequency analysis, and the soil moisture abnormal feature classification is output;

[0031] The abnormal feature classification is analyzed by using the spatial interpolation, and the soil moisture dynamic monitoring report is output.

[0032] As a preferred scheme of the soil moisture dynamic monitoring method based on the Internet of Things, wherein: the wavelet basis and the decomposition layer number are selected according to the health status of the water monitoring network structure, the wavelet packet energy spectrum analysis method is used to perform signal decomposition and time-frequency analysis, and the soil moisture abnormal feature classification is output, and the specific steps are as follows,

[0033] The wavelet basis and the decomposition layer number are selected according to the health status of the water monitoring network structure, and the optimized wavelet transform is performed, and the optimized wavelet parameter combination is output.

[0034] Based on the optimized wavelet parameter combination, signal decomposition and time-frequency analysis are performed, abnormal patterns and periodic disturbances in the data are identified, and an abnormal feature classification is output.

[0035] In a second aspect, the present application provides a soil moisture dynamic monitoring system based on the Internet of Things, comprising a data acquisition module, a state evaluation module, an abnormal positioning module, a topology optimization module and a dynamic reporting module.

[0036] The data acquisition module is used for real-time acquisition of soil moisture data, uses Hilbert transform to obtain phase characteristics and obtains phase difference between measurement points, and performs soil state evaluation.

[0037] The state evaluation module is used for real-time calibration of soil moisture data by using time window data smoothing, and analyzes potential abnormal measurement points in combination with soil state evaluation, and outputs an abnormal diagnosis report.

[0038] The abnormal positioning module is used for decomposition by wavelet transform, and periodic feature extraction by time-frequency joint analysis method to obtain abnormal measurement point positioning information.

[0039] The topology optimization module is used for isolating abnormal measurement points and selecting the shortest path based on abnormal measurement point positioning information, and obtaining a water monitoring network structure.

[0040] The dynamic reporting module is used for dynamically adjusting wavelet transform decomposition parameters and periodic feature extraction parameters based on the water monitoring network structure, and outputting a soil moisture dynamic monitoring report.

[0041] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the computer program is executed by the processor to implement any step of the soil moisture dynamic monitoring method based on the Internet of Things according to the first aspect of the present application.

[0042] In a fourth aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement any step of the soil moisture dynamic monitoring method based on the Internet of Things according to the first aspect of the present application.

[0043] The application has the beneficial effects that: the abnormal measurement points are positioned through wavelet transform and time-frequency joint analysis, the periodic interference identification capability is improved, and the false positive rate is reduced; the Dijkstra algorithm is used to dynamically select the shortest path, and the standby communication channel is constructed through the redundant path, and the data transmission stability is improved. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0045] Fig. 1 Flow chart of the soil moisture dynamic monitoring method based on the Internet of Things.

[0046] Fig. 2 Schematic diagram of the soil moisture dynamic monitoring system based on the Internet of Things.

[0047] Fig. 3 Flow chart of positioning of abnormal measurement points.

[0048] Fig. 4 Flow chart of water monitoring network structure construction. DETAILED DESCRIPTION

[0049] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification.

[0050] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0051] Secondly, the "one embodiment" or "embodiment" referred to herein means that the specific features, structures or characteristics can be included in at least one implementation of the present application. "In one embodiment" appearing in different places in the specification does not mean the same embodiment, nor is it an independent or alternative embodiment that excludes other embodiments.

[0052] REFERENCE Figs. 1-4 For one embodiment of the present application, the embodiment provides a soil moisture dynamic monitoring method based on the Internet of Things, comprising the following steps:

[0053] S1, collecting soil moisture data in real time, using Hilbert transform to obtain phase characteristics and obtain phase difference between measurement points, and performing soil state evaluation.

[0054] The soil moisture data includes physical quantity data, environment-related data, signal characteristic data, and space-time dimension data.

[0055] Further, the physical quantity data includes volume water content, which is the proportion of water in unit volume of soil, weight water content, which is the ratio of water mass to dry soil mass, matric potential, which is soil suction, and conductivity, which is an indicator of salt content; the environment-related data includes soil temperature, atmospheric temperature and humidity, and rainfall; the signal characteristic data includes phase difference, dielectric constant, and wavelet energy entropy, the phase difference and dielectric constant are directly obtained through the phase of the radio frequency signal, and the wavelet energy entropy is obtained through wavelet decomposition; the space-time dimension data includes spatial data (measurement point GPS coordinates and soil layering data) and time data (sampling interval and data time effectiveness label).

[0056] Based on the soil moisture data, the instantaneous phase of the signal characteristic data is obtained by applying Hilbert transform, and the spatial distribution characteristics of the signal characteristic data are obtained by phase difference calculation between measurement points, and a phase difference dataset is generated by integration.

[0057] Specifically, the signal characteristic data is subjected to Z-value standardization and processed using moving average filtering to output smoothed time series signals; Hilbert transform is applied to directly analyze the smoothed time series signals to obtain the instantaneous phase of each measurement point; the instantaneous phase sequence time stamps of each measurement point are unified by linear interpolation, and the phase difference values between two measurement points are obtained by phase difference calculation to generate a phase difference sequence; the phase difference sequence is stored as a structured array, and anomalies are marked using a fixed threshold (absolute value of phase difference sequence greater than 1 radian) to generate a phase difference dataset.

[0058] It should be noted that the phase difference dataset includes time stamp, measurement point pair number: indicating two sensor measurement points participating in phase difference calculation, phase difference value: unit is radian, reflecting the soil moisture difference between two measurement points, environment-related data: temperature and rainfall associated by time stamp, etc.

[0059] Based on the phase difference dataset, the moisture feature is extracted, the dry probability is obtained by using weighted voting method and threshold determination method, and the reference threshold is set by adaptive adjustment strategy, and the soil state evaluation result is output.

[0060] Specifically, based on the phase difference data set, the phase difference data of all measurement point pairs at the same time is extracted, and the weather data and the phase difference data are aligned and merged according to the time stamp to generate a structured data table containing time stamp, measurement point pair, phase difference, environmental parameter and the like; using a three-level feature extraction strategy to obtain the arithmetic mean value of all phase differences to reflect the overall wetting trend, the standard deviation of the phase difference to represent the uniformity of the water distribution in space, and the maximum phase difference value to locate the water anomaly area; taking 5 minutes as a sliding window, the first-order difference value of the phase difference is obtained to capture the rapid change event of water, and the compensation relationship between temperature and phase difference is established, such as the phase difference reference value decreases by 0.1 radian when the temperature increases by 10°C, and the phase difference decay period after the rainfall event is marked (the phase difference decreases at a rate of more than 0.05 radian / minute within 2 hours after the rainfall); according to the historical data stability of the measurement points, weights are assigned, such as 0.8 high weight is given to the measurement points with phase difference fluctuation less than 0.2 radian in the past 24 hours, the weight of the measurement points with instantaneous mutation is reduced to 0.5, and the continuously abnormal measurement points are temporarily excluded; and the phase difference of each measurement point pair is binary judged, such as the phase difference of sandy soil greater than 1.2 radian is recorded as dry, and the phase difference of clay soil greater than 0.8 radian is recorded as dry; the dry probability is output by the weighted dry vote ratio; the reference threshold is set by the adaptive adjustment strategy and initialized according to the soil type, such as the initial reference threshold of sandy soil is 1.2 radian, the initial reference threshold of loam is 1.0 radian, and the initial threshold of clay is 0.8 radian, the state is graded and output in real time, and the soil state evaluation result is output.

[0061] It should be noted that the real-time environmental compensation, the temperature compensation: the reference threshold decreases linearly with the increase of temperature (compensation coefficient 0.01 radian / °C), the rainfall compensation: when the rainfall in the past 6 hours is greater than 5mm, the reference threshold temporarily increases by 0.2 radian, the day and night compensation: the reference threshold automatically decreases by 0.15 radian at night (considering the weakening of plant transpiration); state grading: dry probability 0-0.3 is over-wet, dry probability 0.3-0.6 is suitable, dry probability 0.6-0.8 is mild drought, and dry probability greater than 0.8 is severe drought.

[0062] It should be noted that the dry probability formula is:

[0063] ;

[0064] Wherein, is the dry probability, , is the number of dry measurement points, , is the total number of measurement points, .

[0065] More preferably, the extraction accuracy of signal phase characteristics is improved by Hilbert transform, and the sensitivity to small changes in soil moisture is enhanced; the spatial heterogeneity detection capability is improved by using multi-measurement point phase difference, and the consistency of adjacent measurement point data is verified; the fault tolerance of abnormal data is enhanced by using weighted voting fusion algorithm, and the reliability of state evaluation results is improved.

[0066] S2, using time window data smoothing, real-time calibration of soil moisture data, combined with soil state evaluation analysis of potential abnormal measurement points, output abnormal diagnosis report.

[0067] Further, the measurement deviation of soil moisture data is corrected using time window data smoothing, and the calibrated soil moisture data is output.

[0068] Specifically, when deploying the monitoring sensor, two high-precision reference sensors are arranged, and when measuring the measurement point sensor, the high-precision reference sensor data is read synchronously to obtain real-time deviation, which is the measurement point sensor data minus the high-precision reference sensor data. The real-time deviation is used to obtain compensation data (measurement point sensor data minus real-time deviation), and preliminary calibration data is obtained; the measurement consistency cross-validation of spatial adjacent measurement points is used to identify measurement outliers, based on the preliminary calibration data, the measurement values of the three nearest measurement points and the median value of the measurement values of the three nearest measurement points are obtained, when the difference between the measurement point sensor data and the median value of the measurement values is greater than the reference threshold (such as 2%), it is marked as measurement outlier, and the median value is replaced, and the sensor data after spatial verification is output; using time window data smoothing, setting a 1-hour sliding window and obtaining the average value of the window data, taking the average value of the window data as the current value, outputting the time domain smoothed data; based on the time domain smoothed data, setting an empirical compensation coefficient based on the cumulative working time of the sensor, the empirical compensation coefficient is 0.015% / day, and applying the empirical compensation coefficient, outputting the calibrated soil moisture data.

[0069] It should be noted that the calibrated soil moisture data Preliminary calibration data Empirical compensation coefficient.

[0070] Based on the soil state evaluation results and the calibrated soil moisture data, a dynamic threshold method is used to analyze the differences between adjacent measurement points, and to identify abnormal measurement points.

[0071] Specifically, based on the soil state evaluation result and the calibrated soil moisture data, the calibrated soil moisture data (unit %) and the instantaneous phase value of the target measurement point and the three adjacent measurement points are read, and the absolute phase difference value of each adjacent point of the target measurement point is obtained; the base moisture difference threshold is selected according to the soil type, for example, 3.5% for sandy soil, 2.5% for loam, and 1.8% for clay, and the base moisture difference threshold is adjusted according to the average phase difference value (the average of the absolute phase difference value), when the average phase difference value is greater than or equal to 45 degrees, 150% of the base moisture difference threshold is used as the final moisture difference judgment threshold, when the average phase difference value is less than or equal to 15 degrees, 80% of the base moisture difference threshold is used as the final moisture difference judgment threshold, and when the average phase difference value is greater than 15 degrees and less than 45 degrees, the base moisture difference threshold is used as the final moisture difference judgment threshold; the absolute difference value of the moisture of the target measurement point and each adjacent point (the absolute value of the difference between the calibrated soil moisture data of the target measurement point and each adjacent point, which directly reflects the spatial difference of soil moisture content) is obtained, the number of difference values exceeding the final moisture difference judgment threshold is counted, and it is checked whether the maximum absolute phase difference value exceeds 60 degrees. When two or more final moisture difference judgment thresholds exceed the base moisture difference threshold, and the maximum absolute phase difference is greater than 60 degrees, the measurement point is determined to be abnormal.

[0072] It should be noted that the abnormal measurement point determination result is abnormal, to be observed and normal, and the abnormal type is equipment failure and soil moisture abnormality.

[0073] Based on the abnormal measurement point, the statistical control method is used to obtain the abnormal index, and the adjacent point grouping method is used to identify the position and abnormality degree of the abnormal measurement point, and an abnormal diagnosis report is output.

[0074] Specifically, based on the abnormal measurement point, the soil moisture data of the abnormal measurement point and the soil moisture difference value of the adjacent measurement point, the maximum phase difference and the abnormal duration are obtained, combined with three-level classification, to generate an abnormal point list; taking each abnormal measurement point as the center, search for other abnormal measurement points within a radius of 5 meters, and mark the points forming a group (3 or more abnormal measurement points as a group), and perform regional classification, no adjacent abnormal measurement point is an isolated abnormal point and is an equipment failure area, and the grouped abnormal measurement points are a soil abnormal area, and the equipment failure area and the soil abnormal area are labeled, and an abnormal area map is output; the number of abnormal measurement points is counted, and the execution suggestion is output combined with the condition judgment, for example, the abnormal measurement point is an isolated abnormal point and the level is red, suggesting replacing the sensor immediately, for example, the abnormal measurement point is a cluster and the level is yellow, for example, the abnormal measurement point is a grouped abnormal measurement point and the level is yellow, suggesting checking within 3 days, outputting a soil abnormal measurement point distribution map and processing suggestion, and outputting an abnormal diagnosis report.

[0075] It should be noted that the three-level classification can be set as mild abnormality (blue) when the soil moisture difference value is 10-20% higher or 10-20% lower than the adjacent measurement point, the maximum phase difference is less than 45 degrees, and the duration is less than 12 hours, moderate abnormality (yellow) when the soil moisture difference value is 20-30% higher or 20-30% lower than the adjacent measurement point, the maximum phase difference is between 45 and 60 degrees, and the duration is between 12 and 24 hours, and severe abnormality (red) when the soil moisture difference value is more than 30% higher or more than 30% lower than the adjacent measurement point, the maximum phase difference is more than 60 degrees, and the duration is more than 24 hours.

[0076] Preferably, through the time window data smoothing method, the real-time data calibration is improved, the short-term noise interference is eliminated, and the dynamic correction capability of sensor measurement deviation is enhanced; using dynamic threshold difference analysis, the sensitivity of adjacent measurement point data difference is improved, and the accurate identification of abnormal points is optimized; using the adjacent point grouping statistical method, the clustering analysis capability of abnormal area is enhanced, and the reliability of abnormal point positioning is improved; through the three-level abnormal classification mechanism, the quantitative evaluation of abnormal degree is optimized, and the operability of the diagnosis report is improved.

[0077] S3, decomposing by wavelet transform, and using time-frequency joint analysis method to extract periodic characteristic, obtaining abnormal measurement point positioning information.

[0078] Further, based on the abnormal diagnosis report, using wavelet transform, the soil moisture data is decomposed by 5-layer wavelet packet, the energy entropy value of each sub-band is obtained, and the dynamic threshold method is used to output the energy distribution abnormal sub-band frequency.

[0079] Specifically, using standardization processing to eliminate the dimension difference between different sensors, by obtaining the mean and standard deviation of soil moisture data, the soil moisture data is normalized and adjusted, using sliding average filtering technology, setting an analysis window of 5-10 sampling points, smoothing high-frequency noise interference, for the environment with electromagnetic interference, using Kalman filter for dynamic noise reduction, outputting clean time domain signal; performing 5-layer wavelet packet decomposition on the clean time domain signal, generating 32 characteristic sub-bands, each sub-band corresponding to a specific frequency range, and recording the decomposition coefficients of each sub-band, generating a complete time-frequency analysis matrix; obtaining the energy distribution characteristics of each sub-band through the time-frequency analysis matrix, evaluating the energy complexity of each frequency band through information entropy algorithm, identifying abnormal energy distribution by dynamic threshold technology, outputting the sub-band frequency with abnormal energy distribution, for example, sensor failure is more than 200% increase in energy entropy value of full frequency band, mechanical vibration interference is more than 30% in 1-10Hz frequency band, electromagnetic interference is abnormal aggregation of energy in high frequency band greater than 20Hz, and soil structure change is double peak feature in 0.5-2Hz energy distribution curve.

[0080] It should be noted that the energy distribution anomaly type is sensor failure, mechanical vibration interference, electromagnetic interference and soil structure change; each sub-band corresponds to a specific frequency range, which refers to dividing the clean time domain signal according to frequency components by mathematical transformation; the abnormal energy distribution

[0081] Using the time-frequency joint analysis method, the periodic characteristics in the sub-band frequency band of the abnormal energy distribution are captured, and the periodic interference signal is output.

[0082] Specifically, based on the sub-band frequency band of the abnormal energy distribution, the abnormal frequency range of the sub-band frequency band of the abnormal energy distribution (such as 2-4Hz low frequency anomaly, 8-16Hz high frequency anomaly) is obtained, the initial signal feature data in the frequency range is extracted, the measurement unit difference is eliminated by using data standardization processing, and the initial signal feature data is preprocessed by using the median adjustment method to improve the comparability of the initial signal feature data. The preprocessed abnormal frequency band signal data is output; based on the preprocessed abnormal frequency band signal data, a variable time length analysis method is used to find out the energy fluctuation that obviously exceeds the normal level, and the similarity of the preprocessed abnormal frequency band signal data is directly obtained by using the feature similarity to find out the obvious repeated feature points to determine the reliable repeated period in the range of 0.1-10Hz. The periodic interference signal is output.

[0083] It should be noted that the periodic characteristics are energy fluctuation, repeated feature points and reliable repeated period; the energy fluctuation that obviously exceeds the normal level is more than twice the average value; the obvious repeated feature point is more than 30% of the maximum value.

[0084] Combining the sub-band frequency band of the abnormal energy distribution and the periodic interference signal, the abnormal phase difference is identified by using the change degree of the phase difference of adjacent measurement points, and the abnormal measurement point positioning information is obtained.

[0085] Specifically, the energy distribution abnormal sub-band frequency range (such as 2-4Hz low frequency abnormality) is matched with the periodic interference signal (such as 1.2Hz mechanical vibration) according to the collection time, an abnormal feature record table is established for each measurement point, including: abnormal frequency range, interference signal period value and signal collection time stamp, and the energy abnormal frequency band signal intensity trend of each measurement point is marked, and an abnormal feature correlation table and a set of abnormal energy abnormal frequency band signal waveform graphs of each measurement point are output; the data collected by the topological perception sensor is directly used to obtain the network topology structure, a network topology centralization method is used, the measurement point with the highest connectivity in the network topology structure is selected as the phase reference measurement point, and the signal fluctuation range value (which should be less than ±5%) of the energy abnormal frequency band of the phase reference measurement point in the past 24 hours is verified, the candidate measurement points that have occurred faults in the recent period (such as within 7 days) are excluded, and the reference measurement point is output; for each energy distribution abnormal sub-band frequency, the instantaneous phase angle of the energy abnormal frequency band signal of each measurement point is extracted, the phase deviation value relative to the reference measurement point is obtained, and the phase difference of each measurement point in each energy distribution abnormal sub-band frequency and the phase difference change amplitude of adjacent time periods (interval 5 minutes) are recorded, and the abnormal phase difference is identified, the measurement points with phase difference increase of more than 30% are marked, and the measurement points with multi-frequency band phase difference synchronous abnormality are identified; the abnormal point judgment standard is set as the main fault point and the secondary interference point, the main fault point simultaneously satisfies at least two energy distribution abnormal sub-band frequency phase difference abnormalities, the phase difference change rate is greater than 3 times the average value of adjacent measurement points, and the abnormality lasts for more than 3 cycles, the secondary interference point is a single frequency band phase difference sudden increase of more than 50° and located on the fault propagation path, and the abnormal measurement point positioning information is obtained.

[0086] It should be noted that the connectivity refers to the number of adjacent measurement points directly communicating with the measurement point; the synchronous abnormality refers to frequency interval synchronous abnormality and space-time synchronous abnormality; the phase difference abnormality refers to that the absolute phase deviation of the measurement point relative to the reference reference exceeds the range; the phase difference change rate refers to the change amplitude of the phase difference per unit time; and the fault propagation path is the transmission direction of the abnormal phase difference in the spatial distribution of the sensor measurement points.

[0087] Preferably, the wavelet packet decomposition technology is used to improve the analysis accuracy of the signal frequency domain features, and the abnormal detection capability of the energy distribution of different frequency bands is enhanced; the time-frequency joint analysis method is used to improve the recognition rate of the periodic interference signal, and the distinguishing capability of the transient abnormality and the continuous abnormality is optimized; the dynamic threshold energy entropy detection is used to enhance the adaptive judgment of the abnormal frequency band, and the discrimination capability of the sensor fault and the real soil abnormality is improved; the multi-frequency band phase difference cooperative analysis is used to improve the spatial positioning accuracy of the abnormal measurement point, and the tracking effect of the fault propagation path is optimized.

[0088] S4, based on the abnormal measurement point positioning information, using Dijkstra algorithm to isolate the abnormal measurement point and select the shortest path, obtaining the water monitoring network structure.

[0089] Further, the initial soil moisture monitoring network structure is constructed by the near-neighbor connection method and the threshold filtering is used for measurement point screening and link cleaning, and the purified initial soil moisture monitoring network structure is output.

[0090] Specifically, based on the three-dimensional coordinates (longitude, latitude and altitude) of the measurement points, the initial connection of each measurement point and 3-6 nearest neighbors is established by the Delaunay triangulation algorithm, the triangular mesh meeting the empty circumscribed circle criterion is generated, and the abnormal connection (such as the triangular edge with a length exceeding the maximum distance) is deleted, and the initial network topology graph is output; based on the initial network topology graph, the default maximum communication distance is set to 50 meters (which can be adjusted according to the monitoring environment), and all communication links exceeding the distance are deleted, if the network has isolated measurement points, then the distance threshold is gradually relaxed (increased by 10 meters each time, the upper limit is 80 meters), until the network is connected, and the network subgraph is output; based on the network subgraph, the measurement points perform handshake communication, measure the wireless signal strength, delete the weak signal link with a wireless signal strength less than -75dBm, and test the link with a wireless signal strength between -75dBm and -70dBm multiple times; the link reliability is evaluated by real-time data transmission test, the data transmission is performed for 30 seconds, the packet reception rate is required to be greater than or equal to 90%, the average latency is required to be less than or equal to 100ms, and the latency jitter is required to be less than or equal to 15%, according to the test results, the quality weight is assigned to each link (the packet reception rate accounts for 40%, the average latency accounts for 30%, and the latency jitter accounts for 30%), and the unstable link is removed, and the stable link set and quality score table are output; the graph connectivity analysis method is used to cover the whole domain, whether there is a subnetwork that cannot be connected is checked, the existing measurement points are preferentially used to supplement the line (such as adjusting the connection relationship), the mobile relay measurement point (such as temporarily adding a UAV) is deployed when it cannot be connected, the hop number of any measurement point to the gateway is less than or equal to 5, and the purified initial soil moisture monitoring network structure is output.

[0091] It should be noted that the initial soil moisture monitoring network structure is the main channel, and the packet reception rate refers to the proportion of successfully received data packets per unit time.

[0092] Based on the purified initial soil moisture monitoring network structure, real-time link parameter measurement is performed, and the comprehensive quality is obtained, the threshold comparison method is used to find out the problematic link, and the link information table is output.

[0093] Specifically, the wireless communication active probe is used to let each measurement point send a test data packet to a nearby measurement point, and record the wireless signal strength, data packet reception success rate, transmission delay and delay fluctuation, and record the environment temperature and humidity at the time of measurement, and output an initial measurement data record table; using standardized quality score, the measurement data is converted into quality score, and each index is converted into percentage system, to obtain the comprehensive quality score, and output the link information table with quality score; the threshold comparison method is used to find out the problematic link, the signal strength of each index is set to be less than-85dBm, the packet success rate is less than 80%, and the delay is greater than 300ms, and all links exceeding the index are marked for repeated test confirmation, and the link information table is output.

[0094] It should be noted that the percentage conversion is performed on each index, which can be set to 100 points when the signal strength is-60dBm, and 0 points when the wireless signal strength is-90dBm (the wireless signal strength is directly obtained as the middle value), 100 points when the packet success rate is 100% received, and 90 points when the packet success rate is 90% received (decrease by 10 points for every 10% decrease, and the minimum is 0 points), 100 points when the delay is less than 50ms, and 0 points when the delay is greater than 200ms (the delay score decreases by a certain percentage when the delay is between 50ms and 200ms),

[0095] It should be noted that the comprehensive quality score formula is:

[0096] ;

[0097] Among them, is the comprehensive quality score, which represents the comprehensive performance score of a single communication link, and the higher the score, the better the quality; is the signal strength score, which represents the quantitative evaluation of the received power of the radio signal, ; is the packet reception rate score, which represents the standardized processing of the proportion of successfully received data packets, ; is the delay score, which represents the performance mapping of the end-to-end transmission delay, ; is the signal strength score weight, is the packet reception rate score weight; is the delay score weight; , which can be set to is 40%, is 35%, is 25%.

[0098] According to the link information table, the Dijkstra algorithm is used to obtain the shortest path from each measurement point to the gateway, and real-time performance monitoring is performed, and the path optimization parameter set is output.

[0099] Specifically, a weighted directed graph is constructed, link information is converted into a graph structure based on a link information table, each measurement point is taken as a graph node, an effective communication link is taken as a directed edge, the reciprocal of a link quality score is taken as an edge weight (the higher the quality score, the lower the weight value), and a gateway measurement point is marked as a terminal point, and a weighted directed graph data structure is output; Dijkstra's algorithm is used to obtain the optimal path of each measurement point to the graph node, the distance of the graph node is set to 0, the distance of other graph nodes is set to infinity, and each time the currently smallest unprocessed graph node is selected, the distance values of adjacent graph nodes are updated, and a termination condition is set as all reachable graph nodes being processed or a specified number of backup paths (such as 3) being found, and a shortest path table of each graph node to the gateway is output; a lightweight probe is used to track the actual transmission performance of the path, and end-to-end delay, packet arrival rate, and path hop count are monitored, abnormality detection is performed, path interruption is determined when three consecutive detections fail, an alarm is triggered when the delay suddenly increases by more than 50%, and a path performance real-time monitoring log is output; based on the path performance real-time monitoring log, an adaptive weight update is used, a weight update rule is established, the weight can be set to +0.1 for each 10ms increase in delay, and the weight can be set to +0.2 for each 5% increase in packet loss rate, and the network weight is automatically updated every hour and the abnormal path is immediately recalculated, and the updated edge weight is output; based on the updated edge weight matrix, the maximum allowed path hop count and delay tolerance and other parameters are obtained, and the path optimization parameter set is output, including the recommended path flow distribution ratio and the graph node maintenance priority.

[0100] It should be noted that when updating the network weight, all measurement points report the current signal strength (the larger the value, the better), the data packet reception success rate (percentage), and the transmission speed (milliseconds), and the reciprocal of the calculated link quality score is used for each connection line, and the weight is increased by 0.1 for each 10ms slower than usual, and the weight is increased by 0.2 for each 5% more data packets lost than usual, so that the weight is between 0.1 and 10.

[0101] A redundant path is used to construct a backup communication channel, and the path optimization parameter set is used to dynamically adjust the initial water monitoring network structure after purification, and a water monitoring network structure is generated.

[0102] Specifically, using the K-shortest path algorithm, 1-2 backup communication paths are established for each measurement point, starting from the adjacent graph nodes of the main path, finding non-overlapping alternative routes, and requiring the backup communication path to overlap with the main path less than or equal to 2 graph nodes, the link quality score to be greater than or equal to 80% of the main path, the path hop count to be less than or equal to the main path plus 2, and the switching priority of each backup path to be recorded. Output backup path comparison table; based on the backup path comparison table, use the dual-channel probe to detect each weekly timing test, send test data packets in parallel, the main path continuously sends 10 packets (interval 100 ms), the backup path continuously sends 5 packets (interval 200 ms), and adds the judgment standard that the backup path needs to successfully receive more than or equal to 4 packets, and the time delay is less than or equal to 150% of the main path. Output backup path test report; use incremental topology optimization to optimize network layout based on real-time data, automatically identify monitoring blind areas through interpolation analysis, deploy mobile relay graph nodes on demand, mark graph nodes with quality scores less than 60 for three consecutive days, automatically migrate communication load to backup paths, and increase 1-2 dedicated relay links for high-load graph nodes (daily forwarding volume >1000 packets). Output network structure adjustment scheme; based on the network structure adjustment scheme, directly use the load balancing algorithm to reasonably allocate communication resources, with the main path occupying 60% of the communication time slot, the backup path occupying 30%, and 10% reserved for emergency broadcasts. Low-power graph nodes (less than 30%) are automatically downgraded to terminal nodes, and high-power graph nodes (more than 70%) preferentially undertake relay tasks, outputting a time slot allocation table and a graph node role adjustment list; combine the topology synthesis to integrate the main path, the backup path comparison table, the network structure adjustment scheme, the time slot allocation table, and the graph node role adjustment list to obtain the graph node configuration parameters and generate the soil moisture monitoring network structure.

[0103] Preferably, the Dijkstra shortest path algorithm is used to improve the real-time optimization capability of data transmission paths and enhance the dynamic isolation effect of abnormal nodes; link quality dynamic evaluation is used to improve the reliability of communication links and optimize network load balancing performance; redundant path backup is used to enhance network fault tolerance and improve data transmission stability in complex environments; dynamic topology structure adjustment is used to improve the coverage capability of monitoring blind areas and optimize the load distribution of energy-limited nodes.

[0104] S5, based on the soil moisture monitoring network structure, dynamically adjusting wavelet transform decomposition parameters and periodic feature extraction parameters, outputting a soil moisture dynamic monitoring report.

[0105] Further, based on the soil moisture monitoring network structure, the measurement point density and link stability score are obtained to evaluate the health status of the soil moisture monitoring network structure.

[0106] Specifically, the spatial distribution statistics and the adjacent node count are adopted, the position coordinate data of all measurement points are acquired through the Internet of Things gateway, each sensor graph node reports its own GPS coordinate to the gateway at a time (such as every hour), the monitoring area is divided into standard grids (such as 10m*10m square grids), for each grid, the number of sensor nodes in the grid is counted, and the global average density (total node number to total area) and the local density (node number in a single grid) are acquired, the grid with a node number of 0 is marked as a monitoring blind area, and the grid with a density lower than 50% of the average density is marked as a low coverage area; each communication link is evaluated through communication quality testing and historical data analysis, the gateway sends a test data packet to all graph nodes every 5 minutes, the response time, signal strength and packet loss of each node are recorded, the communication link is divided into three levels, which can be set as a high-quality link with a packet loss rate less than 5% and a response time less than 100ms, a normal link with a packet loss rate of 5-20% and a response time of 100-500ms, and a poor link with a packet loss rate greater than 20% and a response time greater than 500ms, when the link is continuously tested for 3 times and is all poor, the link is marked as an unstable link, when more than 2 unstable links appear in the last 24 hours, the link is marked as a high-risk link, the real-time state and quality level of all links are listed, and a link quality table is output; a weighted scoring method is adopted for comprehensive evaluation, the density score of the measurement point can be set as 0-100 points according to the coverage rate (100 points for no blind area), the network health degree is the sum of the density score and the link stability score, when the network health degree is greater than 80 points, the network is marked as healthy, when the network health degree is 60-79 points, the network is marked as sub-healthy, and when the network health degree is less than 60 points, the network is marked as unhealthy, and the health condition of the water content monitoring network structure is output.

[0107] It should be noted that the expression for obtaining the link stability score is:

[0108] ;

[0109] Among them, is the link stability score, is a quantitative index of link stability; is a packet loss weight coefficient, representing the data reliability weight, which can be set as ; is a time delay weight coefficient, representing the real-time evaluation weight, which can be set as , 0.40; is a jitter weight coefficient, representing the channel stability weight, which can be set as 0.20 0.30, and , and must satisfy the normalization condition ( ); For transmission delay, it represents the end-to-end transmission delay, For delay jitter, it represents the standard deviation of the delay.

[0110] It should be noted that the wavelet base and the number of decomposition layers are selected according to the health status of the water monitoring network structure, and the wavelet packet energy spectrum analysis method is used to perform signal decomposition and time-frequency analysis, and output the soil moisture abnormal feature classification.

[0111] According to the health status of the water monitoring network structure, the wavelet base and the number of decomposition layers are selected, and the optimized wavelet transform is performed, and the optimized wavelet parameter combination is output.

[0112] Specifically, a three-level classification method is used, the water monitoring network state level is divided according to the health status of the water monitoring network structure, the health is excellent, the sub-health is good, and the unhealthy is poor, and the network state level table is output; a wavelet base feature matching table is established, and the optimal wavelet base is selected according to the network state, the optimal wavelet base is selected for the optimal network state, the frequency resolution is high, the wavelet base is selected for the good network state, the frequency resolution is medium, and the wavelet base is selected for the poor network state, and the frequency resolution is low; the best decomposition layer number is determined by using the bandwidth demand analysis method, the effective frequency band of the soil moisture signal is 0.1Hz-10Hz, and each layer of decomposition is divided into two halves, the number of layers is calculated according to the rules of 6 layers of decomposition (minimum resolution 0.1Hz) for the optimal state, 5 layers of decomposition (minimum resolution 0.2Hz) for the good state, and 4 layers of decomposition (minimum resolution 0.4Hz) for the poor state, and the network state is re-evaluated every 6 hours, and the state is updated when the state change exceeds 10%, and the final decomposition layer number is output; the symmetric extension method is used to reduce the boundary distortion, and the middle 90% of the data is intercepted, the poor link data (packet loss rate>15%) is first subjected to median filtering (window=5 sampling points), and then subjected to wavelet threshold denoising, and the optimized wavelet parameter combination is output.

[0113] It should be noted that the necessary parameters selected for the poor state are the basic time domain parameters, the frequency domain parameters of the low-frequency energy proportion (0.1-1Hz frequency band), and the spatial correlation parameters.

[0114] Based on the optimized wavelet parameter combination, signal decomposition and time-frequency analysis are performed, abnormal patterns and periodic disturbances in the data are identified, and abnormal feature classification is output.

[0115] Specifically, based on the optimized wavelet parameter combination (such as Db4 wavelet + 5 layer decomposition), the signal is decomposed, the initial soil moisture signal is normalized, and the selected boundary extension method is applied, 5 layer decomposition is performed, 16 sub-bands are obtained, the energy entropy of each sub-band is obtained, the sub-band with energy entropy greater than 1.5 times the average entropy is selected as the abnormal frequency band, and the abnormal frequency band is reconstructed; combined with wavelet scale spectrum and Hilbert transform, the reconstructed signal is continuously wavelet transformed and the time-frequency energy heat map is generated, the Hilbert transform is performed on the abnormal frequency band to extract the instantaneous amplitude and phase, and the instantaneous frequency is obtained, the instantaneous frequency sudden increase greater than 20% (may be mechanical vibration) and the fixed frequency fluctuation lasting more than 3 minutes (may be irrigation interference) are marked, and the interference event marker list is output; the moving average μ and the standard deviation σ of the recent 7-day data at the same period are obtained, the abnormal threshold is set to mild abnormality μ±2σ and severe abnormality μ±3σ, and the feature morphology analysis is performed, the drought lasts for a long time, the low frequency energy rises, the waterlogging rises suddenly, the full band energy rises, the salinization rises and falls slowly, and the specific frequency band resonates, and the signal segment exceeding the abnormal threshold is output. The abnormal feature classification table; the periodic component is extracted by using the synchronous average method, the marked interference signal is subjected to autocorrelation analysis and the significant period (such as peak interval = 5 minutes) is extracted, the signal segment is intercepted with the period as the length, the average value of each segment is obtained to obtain the interference template, and the abnormal feature classification is generated.

[0116] The abnormal feature classification is analyzed by using spatial interpolation, and a soil moisture dynamic monitoring report is output.

[0117] Specifically, based on the abnormal feature classification, the geographic information spatial indexing method is combined to generate an abnormal feature data set with geographic coordinates, the node ID, longitude, latitude, abnormal type, abnormal degree and time stamp are extracted from the abnormal feature classification, the R-tree index is used to accelerate spatial query and generate spatial partitions according to 100m×100m grid, the abnormal degree is quantified as 1-5 levels (1 = slight, 5 = severe), and the time is aligned to the nearest whole point time window; the inverse distance weighted interpolation method is used, the search radius parameter is dynamically adjusted (default 50m, expanded to 100m in sparse areas), the obstacle is treated as superimposed farmland ditch, the road vector layer is used as a barrier, each type of abnormality is independently interpolated, and the interpolation is obtained grid by grid, 10% of the nodes are reserved for verification, and the soil moisture abnormal spatial distribution raster is output; the spatio-temporal change detection of time series grid algebra is used, the current and past 3 hours of interpolation results are loaded, the grid change quantity of each grid is obtained, the area with grid change quantity greater than 0.2 is the diffusion area, and the area with grid change quantity less than -0.2 is the recession area, the soil moisture change heat map is output, and the soil moisture dynamic monitoring report is output.

[0118] It should be noted that the abnormality degree is quantified as 1 point for every 5% exceeding the abnormal threshold amplitude (such as 4 points for exceeding 20%), and the abnormal threshold amplitude is the difference between the sensor real-time reading (soil volume water content) and the set reference threshold divided by the range; the abnormal duration is 0.5 points per hour (such as 1.5 points for 3 hours); the proportion of abnormal graph nodes within 10m is 1 point for every 20% (such as 3 points for 60%).

[0119] Preferably, the network health assessment improves the real-time performance of network state perception, enhances the dynamic assessment capability of node density and link quality, uses wavelet parameter adjustment to improve the accuracy of signal decomposition and optimize the use efficiency of computing resources, uses time-frequency joint feature extraction optimization to enhance the identification capability of periodic interference and improve the accuracy of abnormal feature classification, and uses spatial interpolation analysis to improve the spatial distribution accuracy of soil moisture anomalies.

[0120] The embodiment also provides a soil moisture dynamic monitoring system based on Internet of Things, comprising: a data acquisition module for acquiring soil moisture data in real time, using Hilbert transform to obtain phase features and phase differences between measurement points, and performing soil state evaluation; a state evaluation module for calibrating soil moisture data in real time by using time window data smoothing, and combining soil state evaluation to analyze potential abnormal measurement points and output abnormal diagnosis report; an abnormal positioning module for decomposing by wavelet transform and extracting periodic features by using time-frequency joint analysis method to obtain abnormal measurement point positioning information; a topology optimization module for isolating abnormal measurement points and selecting the shortest path based on abnormal measurement point positioning information by using Dijkstra algorithm to obtain water monitoring network structure; and a dynamic report module for dynamically adjusting wavelet transform decomposition parameters and periodic feature extraction parameters based on water monitoring network structure to output soil moisture dynamic monitoring report.

[0121] The embodiment also provides a computer device suitable for the soil moisture dynamic monitoring method based on Internet of Things, comprising: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the soil moisture dynamic monitoring method based on Internet of Things proposed in the above embodiment.

[0122] The computer device can be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected by a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be achieved by WIFI, an operator network, NFC (Near Field Communication) or other technologies. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0123] The embodiment also provides a storage medium having a computer program stored thereon, the program being executed by a processor to implement the method for monitoring soil moisture dynamics based on the Internet of Things as proposed in the above embodiment. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0124] To sum up, the application has the following advantages: wavelet transform and time-frequency joint analysis are used to locate abnormal measurement points, improve the ability to identify periodic interference, and reduce the false alarm rate; Dijkstra algorithm is used to dynamically select the shortest path, and a standby communication channel is constructed through a redundant path to improve the stability of data transmission.

[0125] It should be noted that the above embodiments are only used to illustrate the technical solutions of the application rather than limit the application. Although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the application can be modified or replaced by equivalents without departing from the spirit and scope of the application, and all modifications or replacements should be covered in the scope of the claims of the application.

Claims

1. A method for dynamic monitoring of soil moisture based on the Internet of Things, characterized in that: include, Soil moisture data is collected in real time, and phase characteristics are obtained using Hilbert transform to acquire the phase difference between measurement points for soil condition assessment. The soil moisture data includes physical quantity data, environmental correlation data, signal characteristic data, and spatiotemporal dimension data. Soil moisture data is calibrated in real time using time window data, and potential abnormal measurement points are analyzed in conjunction with soil condition assessment to output an anomaly diagnosis report; The abnormal measurement point location information is obtained by decomposing the sample using wavelet transform and extracting periodic features using a time-frequency joint analysis method. Based on the location information of abnormal measurement points, the Dijkstra algorithm is used to isolate abnormal measurement points and select the shortest path to obtain the moisture monitoring network structure. Based on the dynamic adjustment of wavelet transform decomposition parameters and periodic feature extraction parameters of the water monitoring network structure, a dynamic soil moisture monitoring report is output.

2. The method for dynamic monitoring of soil moisture based on the Internet of Things as described in claim 1, characterized in that: The specific steps for using Hilbert transform to obtain phase characteristics and acquire the phase difference between measurement points to assess soil condition are as follows. Based on soil moisture data, Hilbert transform is applied to obtain the instantaneous phase of signal feature data, and the spatial distribution characteristics of signal feature data are obtained by phase difference calculation between measurement points. The data are then integrated to generate a phase difference dataset. Moisture features are extracted based on the phase difference dataset, and the dryness probability is obtained by weighted voting and threshold determination. The baseline threshold is set by an adaptive adjustment strategy, and the soil condition assessment results are output.

3. The method for dynamic monitoring of soil moisture based on the Internet of Things as described in claim 2, characterized in that: The process involves smoothing and calibrating soil moisture data in real time using time window data, and combining this with soil condition assessment to analyze potential abnormal measurement points and output an anomaly diagnosis report. The specific steps are as follows: The measurement bias of soil moisture data is corrected by smoothing the data using time window data, and the calibrated soil moisture data is output. Based on the soil condition assessment results and calibrated soil moisture data, the dynamic threshold method was used to analyze the differences between adjacent measurement points and identify abnormal measurement points. Statistical control methods are used to obtain abnormal indicators, and the location and degree of abnormality of the measurement points are identified by the nearest neighbor grouping method, and an abnormality diagnosis report is output.

4. The method for dynamic monitoring of soil moisture based on the Internet of Things as described in claim 3, characterized in that: The process involves decomposing the data using wavelet transform and extracting periodic features using a time-frequency joint analysis method to obtain the location information of abnormal measurement points. The specific steps are as follows: Wavelet transform is used to perform 5-level wavelet packet decomposition on the measurement point signal to obtain the energy entropy value of each sub-band, and the dynamic threshold method is used to output the sub-band frequency band with abnormal energy distribution. Using a time-frequency joint analysis method, periodic features are captured in sub-band frequencies with abnormal energy distribution, and periodic interference signals are output. By combining the sub-band frequency bands with abnormal energy distribution with periodic interference signals, abnormal phase differences are identified by utilizing the degree of change in phase difference between adjacent measurement points, and abnormal measurement point location information is obtained.

5. The method for dynamic monitoring of soil moisture based on the Internet of Things as described in claim 4, characterized in that: The process involves using the Dijkstra algorithm to isolate abnormal measurement points and select the shortest path based on the location information of these points, thereby obtaining the moisture monitoring network structure. The specific steps are as follows: An initial moisture monitoring network structure was constructed using the nearest neighbor connection method, and threshold filtering was used to screen measurement points and clean up links, resulting in a purified initial moisture monitoring network structure. Based on the purified initial soil moisture monitoring network structure, real-time link parameters are measured and comprehensive quality is obtained. The threshold comparison method is used to identify problematic links and output a link information table. Based on the link information table, the Dijkstra algorithm is used to obtain the shortest path from each measurement point to the gateway, and real-time performance monitoring is performed to output a set of path optimization parameters. A backup communication channel is constructed using redundant paths, and the initial moisture monitoring network structure after purification is dynamically adjusted using the path optimization parameter set to generate the moisture monitoring network structure.

6. The method for dynamic monitoring of soil moisture based on the Internet of Things as described in claim 5, characterized in that: The method involves dynamically adjusting wavelet transform decomposition parameters and periodic feature extraction parameters based on the water monitoring network structure to output a dynamic soil moisture monitoring report. The specific steps are as follows: Based on the moisture monitoring network structure, the density of measurement points and link stability scores are obtained to assess the health status of the moisture monitoring network structure. Based on the health status of the water monitoring network structure, wavelet basis and decomposition layer number are selected, and wavelet packet energy spectrum analysis is used to perform signal decomposition and time-frequency analysis to output the classification of soil moisture anomaly characteristics. Spatial interpolation analysis is used to classify abnormal features and output a dynamic monitoring report on soil moisture.

7. The method for dynamic monitoring of soil moisture based on the Internet of Things as described in claim 6, characterized in that: The process involves selecting wavelet bases and decomposition levels based on the health status of the water monitoring network structure, and then using wavelet packet energy spectrum analysis to perform signal decomposition and time-frequency analysis to output a classification of soil moisture anomaly characteristics. The specific steps are as follows. Based on the health status of the water monitoring network structure, the wavelet basis and decomposition level are selected, and the wavelet transform is optimized to output the optimized wavelet parameter combination. Based on the optimized wavelet parameter combination, signal decomposition and time-frequency analysis are performed to identify abnormal patterns and periodic interference in the data, and anomaly feature classification is output.

8. A soil moisture dynamic monitoring system based on the Internet of Things (IoT), based on the soil moisture dynamic monitoring method based on the IoT as described in any one of claims 1 to 7, characterized in that: It includes a data acquisition module, a status assessment module, an anomaly location module, a topology optimization module, and a dynamic reporting module; The data acquisition module is used to collect soil moisture data in real time, use Hilbert transform to obtain phase characteristics and obtain the phase difference between measurement points, and perform soil condition assessment. The soil condition assessment module is used to smooth and calibrate soil moisture data in real time using time window data, and analyze potential abnormal measurement points in conjunction with soil condition assessment to output an anomaly diagnosis report. The anomaly localization module is used to decompose the anomaly through wavelet transform and extract periodic features using a time-frequency joint analysis method to obtain the location information of the anomaly measurement point. The topology optimization module is used to isolate abnormal measurement points and select the shortest path based on the location information of abnormal measurement points, thereby obtaining the moisture monitoring network structure. The dynamic reporting module is used to dynamically adjust the wavelet transform decomposition parameters and periodic feature extraction parameters based on the water monitoring network structure, and output a dynamic soil moisture monitoring report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the IoT-based dynamic soil moisture monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the IoT-based dynamic soil moisture monitoring method according to any one of claims 1 to 7.

Images