A method and system for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution

Abstract

The present application discloses a method and system for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution. By collecting the time-series dataset of groundwater data in the target irrigation area located in arid and semi-arid regions, and extracting the time-series dataset related to the water level values of groundwater from it, and then introducing data processing and analysis algorithms based on artificial intelligence and deep learning at the backend to perform time-series aggregation analysis on these water level values, so as to capture the time-series multi-scale semantic feature representation of the groundwater level in the target irrigation area, and based on this groundwater level time-series feature representation, predict the estimated value of the decline of the groundwater level in the target irrigation area within the next 10 years. In this way, it is possible to predict the dynamic changes of the groundwater level based on the spatio-temporal dynamic evolution characteristics of the groundwater in the target irrigation area, and thereby determine whether it will affect the sustainable development of agricultural production and the sustainable utilization of water resources, providing more scientific and accurate support for water resource management and decision-making.

Description

Technical Field

[0001] This application relates to the field of feature analysis, and more specifically, to a method and system for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution. Background Art

[0002] Irrigation areas in arid and semi-arid regions usually face severe challenges in water and soil resource management, especially when groundwater is the main irrigation water source. Due to scarce precipitation, large evaporation, and limited natural recharge capacity in these regions, the dependence on groundwater is very high. However, long-term overexploitation has led to a series of problems such as declining groundwater levels and deteriorating water quality, which not only affect the sustainability of agricultural production but may also trigger a series of environmental and socio-economic problems. Therefore, it is particularly important to scientifically manage and rationally utilize the groundwater in irrigation areas.

[0003] Although traditional methods for analyzing groundwater in irrigation areas can provide some basic trend analyses, they have limitations in dealing with complex non-linear groundwater dynamics. Specifically, traditional analysis schemes are usually based on static or quasi-static models, assuming that the groundwater system remains unchanged over a long period. However, in reality, due to the influence of factors such as climate change and human activities, the dynamics of the groundwater system are very strong, making it difficult for static models to accurately reflect the actual situation. In addition, traditional groundwater analysis methods rely on statistical analysis and assume that the changes in groundwater data are linear, but in fact, the response of the groundwater system is often non-linear, especially under extreme conditions (such as long-term drought or overexploitation). This simplification may ignore important non-linear characteristics, thus underestimating or overestimating the importance of certain key processes and causing misjudgments in groundwater analysis.

[0004] Therefore, an optimized scheme for analyzing the characteristics of groundwater in irrigation areas is desired. Summary of the Invention

[0005] To solve the above technical problems, the present application is proposed. Embodiments of the present application provide a method and system for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution. By collecting a time-series dataset of groundwater data in a target irrigation area located in arid and semi-arid regions, and extracting a time-series dataset of water level values related to groundwater from it, and then introducing data processing and analysis algorithms based on artificial intelligence and deep learning in the backend to perform time-series aggregation analysis on these water level values, so as to capture the time-series multi-scale semantic feature representation of the groundwater level in the target irrigation area, and based on this groundwater level time-series feature representation, predict the estimated value of the decline of the groundwater level in the target irrigation area within the next 10 years. In this way, it is possible to predict the dynamic changes of the groundwater level based on the spatio-temporal dynamic evolution characteristics of the groundwater in the target irrigation area, and thereby determine whether it will affect the sustainable development of agricultural production and the sustainable utilization of water resources, providing more scientific and accurate support for water resource management and decision-making.

[0006] According to one aspect of the present application, there is provided a method for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution, which includes:

[0007] Obtain a time-series dataset of groundwater data in a target irrigation area, where the groundwater data includes water volume data and water quality data, the water volume data includes water level values and water volume values, and the water quality data includes salinity, pH value, total hardness, salt content, ionic composition, and nitrogen, phosphorus, and potassium;

[0008] Extract water level values from the time-series dataset of groundwater data in the target irrigation area to obtain a time-series dataset of water level values;

[0009] Perform singular spectrum decomposition on the time-series dataset of water level values to obtain a set of subsequences of the groundwater level in the target irrigation area;

[0010] Perform time-series feature extraction based on the groundwater level on each subsequence of the groundwater level in the target irrigation area in the set of subsequences of the groundwater level in the target irrigation area to obtain a set of time-series implicit coding features of the groundwater level in the target irrigation area;

[0011] Perform graph walk significant aggregation analysis on the set of time-series implicit coding features of the groundwater level in the target irrigation area to obtain time-series multi-scale semantic coding features of the groundwater level in the target irrigation area;

[0012] Perform groundwater feature analysis based on the time-series multi-scale semantic coding features of the groundwater level in the target irrigation area to determine the estimated value of the decline of the groundwater level in the target irrigation area within the next 10 years.

[0013] According to another aspect of the present application, there is provided a system for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution, which includes:

[0014] A data acquisition module for obtaining a time series dataset of groundwater data in a target irrigation area, where the groundwater data includes water volume data and water quality data, the water volume data includes water level values and water volume values, and the water quality data includes salinity, pH value, total hardness, salt content, ionic composition, and nitrogen, phosphorus, and potassium;

[0015] A water level value extraction module for extracting water level values from the time series dataset of groundwater data in the target irrigation area to obtain a time series dataset of water level values;

[0016] A singular spectrum decomposition module for performing singular spectrum decomposition on the time series dataset of water level values to obtain a set of subsequences of the groundwater level in the target irrigation area;

[0017] A time series feature extraction module for respectively performing time series feature extraction based on the groundwater level on each subsequence of the groundwater level in the target irrigation area in the set of subsequences of the groundwater level in the target irrigation area to obtain a set of time series implicit coding features of the groundwater level in the target irrigation area;

[0018] A significant aggregation analysis module for performing graph walk significant aggregation analysis on the set of time series implicit coding features of the groundwater level in the target irrigation area to obtain multi-scale semantic coding features of the time series of the groundwater level in the target irrigation area;

[0019] A groundwater feature analysis module for performing groundwater feature analysis based on the multi-scale semantic coding features of the time series of the groundwater level in the target irrigation area to determine an estimated value of the decline of the groundwater level in the target irrigation area in the next 10 years.

[0020] Compared with the prior art, a method and system for analyzing groundwater characteristics in an irrigation area based on spatio-temporal evolution provided by the present application collects a time series dataset of groundwater data in a target irrigation area located in an arid and semi-arid area, extracts a time series dataset of water level values related to groundwater from it, and then introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the backend to perform time series aggregation analysis on these water level values, so as to capture the multi-scale semantic feature representation of the time series of the groundwater level in the target irrigation area, and based on this time series feature representation of the groundwater level, predict the estimated value of the decline of the groundwater level in the target irrigation area in the next 10 years. In this way, it is possible to predict the dynamic changes of the groundwater level based on the spatio-temporal dynamic evolution characteristics of the groundwater in the target irrigation area, and thereby determine whether it will affect the sustainable development of agricultural production and the sustainable utilization of water resources, providing more scientific and accurate support for water resource management and decision-making. Description of the Drawings

[0021] The above and other objects, features, and advantages of the present application will become more apparent by describing the embodiments of the present application in more detail with reference to the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0022] Figure 1 FIG. is a flowchart of a method for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution according to an embodiment of the present application;

[0023] Figure 2 FIG. is a schematic diagram of data flow of a method for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution according to an embodiment of the present application;

[0024] Figure 3 FIG. is a flowchart of sub-step S5 of a method for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution according to an embodiment of the present application;

[0025] Figure 4 FIG. is a block diagram of a system for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution according to an embodiment of the present application. Detailed Embodiments

[0026] Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0027] As shown in the present application and the claims, unless the context clearly indicates otherwise, words such as "a", "an", "one", and / or "the" are not specifically singular and may also include plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0028] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or the server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0029] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the operations before or below do not necessarily need to be executed precisely in order. On the contrary, according to the need, various steps can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or more steps can be removed from these processes.

[0030] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0031] It should be understood that since groundwater data includes various different types of data, such as water volume data such as water level values and water volume values, and water quality data such as salinity, pH value, total hardness, salt content, ionic total, and nitrogen, phosphorus, and potassium, these data have different temporal variation situations in different irrigation areas. At the same time, since the current arid and semi-arid regions mainly rely on groundwater for irrigation, but the problem of overexploitation of groundwater has been serious in recent years, the groundwater system shows complex spatio-temporal variation characteristics. If the existing irrigation mode is maintained, the groundwater level will drop within the next 10 years, affecting the sustainability of agricultural production. Therefore, if one wants to dynamically predict the groundwater in the irrigation area, it is necessary to conduct spatio-temporal evolution on the spatio-temporal variation characteristics of the groundwater in the irrigation area.

[0032] Based on this, in the technical solution of the present application, a method for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution is proposed. Figure 1 FIG. is a flowchart of a method for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution according to an embodiment of the present application. Figure 2 FIG. is a schematic diagram of data flow of a method for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution according to an embodiment of the present application. As Figure 1 and Figure 2 shown, the method for analyzing the characteristics of groundwater in an irrigation area based on spatio-temporal evolution according to an embodiment of the present application includes the steps of: S1, obtaining a time series dataset of groundwater data of a target irrigation area, the groundwater data including water volume data and water quality data, the water volume data including a water level value and a water volume value, and the water quality data including salinity, pH value, total hardness, salt content, ionic total, and nitrogen, phosphorus, and potassium; S2, extracting the water level value from the time series dataset of the groundwater data of the target irrigation area to obtain a time series dataset of the water level value; S3, performing singular spectrum decomposition on the time series dataset of the water level value to obtain a set of subsequences of the groundwater level in the target irrigation area; S4, respectively performing time series feature extraction based on the groundwater level on each subsequence of the groundwater level in the set of subsequences of the groundwater level in the target irrigation area to obtain a set of time series implicit coding features of the groundwater level in the target irrigation area; S5, performing graph walk significant aggregation analysis on the set of time series implicit coding features of the groundwater level in the target irrigation area to obtain time series multi-scale semantic coding features of the groundwater level in the target irrigation area; S6, performing groundwater feature analysis based on the time series multi-scale semantic coding features of the groundwater level in the target irrigation area to determine an estimated value of the drop of the groundwater level in the target irrigation area within the next 10 years.

[0033] In particular, in S1, a time series dataset of groundwater data for the target irrigation area is obtained. The groundwater data includes water quantity data and water quality data. The water quantity data includes water level values and water volume values. The water quality data includes salinity, pH value, total hardness, salt content, total ion content, and nitrogen, phosphorus, and potassium. It should be understood that the water level value in the water quantity data is an important indicator for measuring the abundance of groundwater resources. Long-term monitoring of the change trend of the water level can help identify whether there is overexploitation and evaluate the contribution of natural recharge (such as rainfall) to the groundwater system. By analyzing the water level changes, the future groundwater level dynamics can be predicted, providing a basis for water resource management and policy making. The water volume reflects the total amount of available groundwater resources within a certain period, which is crucial for evaluating irrigation needs and planning water allocation. Combining with the water level information, the water quantity data helps to more accurately understand the storage and flow characteristics of groundwater and support reasonable water intake strategies. In terms of water quality data, salinity refers to the total amount of dissolved solids in water. High salinity may indicate the presence of excessive minerals or pollutants, which can affect crop growth and cause soil salinization. Monitoring the salinity level can alert potential water quality deterioration problems and guide appropriate purification measures. The pH value indicates the acidity or alkalinity of the water body. Extreme pH values can have a negative impact on the absorption of nutrients by plant roots and may also corrode irrigation facilities. Maintaining an appropriate pH range is very important for maintaining a healthy agricultural ecosystem and extending the lifespan of the irrigation system. The total hardness is determined by the calcium and magnesium ion content. Excessive hardness will lead to hard water phenomena, affect irrigation efficiency, increase energy consumption, and may cause pipeline scaling. Understanding the total hardness helps to optimize irrigation methods and select crop varieties resistant to hard water suitable for local conditions. The salt content directly affects soil structure and plant health. Excessive salt can inhibit crop growth and even poison plants. Regular detection of the salt content can timely adjust the irrigation management plan to prevent land degradation. The presence and concentration distribution of specific ions (such as sodium, chlorine, etc.) have a significant impact on the soil and water environment. Excessive concentration of certain ions may cause ecological problems. By monitoring the ion composition, the interaction between groundwater and the surrounding environment can be better understood to ensure ecological environment safety. Nitrogen, phosphorus, and potassium are the main nutrients necessary for plant growth, but excessive fertilization may lead to eutrophication problems and pollute groundwater. Reasonably controlling the input amount of nitrogen, phosphorus, and potassium can not only improve fertilizer utilization efficiency but also reduce the pollution risk to groundwater. Therefore, the spatio-temporal evolution characteristics of groundwater in the irrigation area can be comprehensively grasped, providing solid data support for precision agriculture, environmental protection, and resource management.

[0034] In one example, first of all, reasonably selecting monitoring points in the target irrigation area is a crucial step. These monitoring points should be able to represent the groundwater characteristics of the entire irrigation area and cover different types of geological structures and land use patterns. Considering factors such as population distribution, agricultural activities, and industrial facilities in the irrigation area, selecting monitoring points with broad representativeness can ensure the representativeness of the data. To achieve this, researchers need to comprehensively consider multiple factors such as geographical location, geological conditions, and hydrogeological characteristics to ensure the scientific and reasonable selection of monitoring points. For example, setting up monitoring points near frequently irrigated farmland can help capture the changes in the groundwater level caused by agricultural water use; while setting up monitoring points in nature reserves far from the influence of human activities helps to understand the undisturbed groundwater dynamics. After determining the location of the monitoring points, various types of sensors and measurement devices are installed to collect groundwater data. At the selected monitoring points, water level gauges are usually installed to monitor the changes in the groundwater level in real time. These water level gauges are usually placed in wells and can accurately record the changes in the groundwater level over time. In addition, flow meters are installed to measure the flow velocity or flow rate of groundwater, which is crucial for evaluating the amount of groundwater resources. At the same time, a water quality analyzer equipped with conductivity probes, pH sensors, temperature sensors, etc. is also used to online detect multiple parameters such as salinity, pH value, total hardness, and salt content. For some more detailed ion components (such as sodium, chlorine, etc.) and the content of nitrogen, phosphorus, and potassium, they can be determined by regularly sampling and bringing them back to the laboratory. In this way, the quality status and its change trend of groundwater can be comprehensively grasped. In terms of selecting the appropriate sampling frequency and cycle, a reasonable sampling strategy must be set according to the specific research purpose and actual situation. For key indicators such as water level changes, high-frequency data records (such as once an hour) may be required to capture the rapidly changing trends. For some relatively stable water quality parameters, lower-frequency monthly or quarterly sampling can be selected. This flexible sampling arrangement can not only meet the needs of scientific research but also effectively utilize resources and avoid unnecessary waste. At the same time, considering the impact of seasonal changes on groundwater, the sampling plan should also fully consider the different characteristics of the rainy season and the dry season to ensure the integrity and continuity of the annual data.

[0035] In particular, in S2, water level values are extracted from the time series dataset of the groundwater data of the target irrigation area to obtain a time series dataset of water level values. As an important indicator to measure the abundance of groundwater resources, the water level value directly reflects the state of the groundwater system. By separately extracting the water level value, one can focus more on the change trend of this key variable, avoid the interference of other factors, and improve the accuracy of the analysis.

[0036] In one example, first, the raw data is cleaned. Denoising and outlier handling are important steps in data cleaning, including operations such as removing obviously incorrect data points, filling in missing values (through interpolation or other statistical methods), and smoothing curves to reduce the impact of random fluctuations, ensuring the accuracy and reliability of the data used for subsequent analysis. Especially for water level data, any values that deviate significantly from the normal range should be carefully examined to determine whether they are measurement errors or actual extreme events. In addition, due to possible measurement biases among different brand models of instruments, all water level data also need to go through a standardization and calibration process. This usually involves converting the raw readings into a unified standard unit and regularly calibrating the instrument using known standard samples to ensure the consistency and comparability of long-term monitoring results. Through strict internal quality control and external quality assurance measures, the data quality is further improved to ensure its scientific nature and reliability. Next, the fields related to water level are selected from the database. These fields may include, but are not limited to, timestamps, well numbers / location identifiers, and water level gauge readings. The timestamp records the specific time and date of each observation; the well number / location identifier is used to distinguish different monitoring points; and the water level gauge reading is the actual measured value of the groundwater level. To ensure that the extracted data only contains valid water level values, reasonable filtering conditions need to be set. For example, data records marked as invalid or in an uncertain state can be excluded; at the same time, the extracted data set can also be limited according to a specific time range or geographical area. This is done to obtain a more focused and representative water level change sequence. Then, the filtered water level values are arranged in chronological order to form a complete time series data set of water level values. Each data point contains a timestamp and the corresponding water level measurement value. In some cases where there are multiple monitoring points, multiple independent time series data sets can also be considered, corresponding to the locations of each monitoring point respectively. This helps to better capture local features and facilitates subsequent comparative analysis of differences between different locations. If obvious discontinuities or missing time periods are found in the time series data set, these gaps can be filled through interpolation methods (such as linear interpolation, spline interpolation, etc.). Reasonable interpolation not only maintains the continuity of the data but also helps to improve the effect of model training, especially important when applying machine learning algorithms.

[0037] Specifically, in step S3, singular spectrum decomposition is performed on the time series dataset of the water level values to obtain a set of subsequences of the groundwater level in the target irrigation area. It should be understood that since the change of the groundwater level usually includes long-term trends, such as slow rise or fall due to climate change or human activities. In addition, the groundwater level may also exhibit seasonal fluctuations or other periodic patterns. For example, the peak irrigation period, rainy season, etc. will cause periodic changes. Therefore, in the technical solution of this application, singular spectrum decomposition is further performed on the time series dataset of the water level values to obtain a set of subsequences of the groundwater level in the target irrigation area. Through the singular spectrum decomposition operation, the time series dataset of the water level values can be decomposed into multiple subsequences of the groundwater level in the target irrigation area (i.e., reconstructed time series) that are easier to understand and analyze, thereby reducing the complexity of the problem and facilitating the subsequent capture of time series dynamic feature information such as the long-term trend and periodic changes of the groundwater level in the target irrigation area, which helps to more accurately predict and estimate the decline of the groundwater level in the next 10 years.

[0038] Specifically, in step S4, temporal feature extraction based on the groundwater level is respectively performed on each subsequence of the groundwater level in the target irrigation area in the set of subsequences of the groundwater level in the target irrigation area to obtain a set of temporal implicit coding features of the groundwater level in the target irrigation area. That is, in the technical solution of this application, each subsequence of the groundwater level in the target irrigation area in the set of subsequences of the groundwater level in the target irrigation area is respectively input into a groundwater level temporal feature extractor based on an LSTM-RNN hybrid model to obtain a set of temporal implicit coding feature vectors of the groundwater level in the target irrigation area as the set of temporal implicit coding features of the groundwater level in the target irrigation area. It should be understood that the LSTM-RNN hybrid model combines the advantages of the two architectures, retaining both the ability of LSTM to handle long-term dependencies and the sensitivity of RNN to short-term local features. Therefore, through the processing of the groundwater level temporal feature extractor based on the LSTM-RNN hybrid model, LSTM and RNN can be respectively used to capture the long-term dependency relationship and short-term local temporal features of the groundwater level in the target irrigation area in the time dimension, so as to perceive and extract the temporal implicit correlation features of the groundwater level at different time scales, and then combine the multi-scale perception features of the groundwater level at these two scales to provide data support for subsequent groundwater level prediction.

[0039] Specifically, in step S5, graph-walking significant aggregation analysis is performed on the set of implicit encoding features of the target irrigation area's groundwater level time series to obtain the multi-scale semantic encoding features of the target irrigation area's groundwater level time series. Since each implicit encoding feature vector of the target irrigation area's groundwater level time series in the set of implicit encoding feature vectors contains the time series feature information regarding the groundwater level of the target irrigation area in different time periods, the groundwater level time series features in these different time periods represent the feature semantic information of different nodes. And the groundwater level of the target irrigation area shows dynamic change characteristics and trends in the time dimension. In order to be able to perform aggregation analysis on the groundwater level time series feature semantics of these different nodes to provide a basis for subsequent groundwater level prediction, in the technical solution of this application, further graph-walking significant aggregation analysis is performed on the set of implicit encoding features of the target irrigation area's groundwater level time series to obtain the multi-scale semantic encoding features of the target irrigation area's groundwater level time series. Through graph-walking significant aggregation analysis, it is possible to use graph-walking technology and self-correlation gating mechanism to perform aggregation analysis on the groundwater level time series features of different nodes in the discrete feature distribution to obtain a significant aggregation result of the discrete feature distribution. In a specific example of this application, as Figure 3 shown, step S5 includes: S51, based on the set of implicit encoding features of the target irrigation area's groundwater level time series, extracting the graph-walking topological features of the target irrigation area's groundwater level time series; S52, based on the graph-walking topological features of the target irrigation area's groundwater level time series, performing significant gating weighted aggregation on the set of implicit encoding features of the target irrigation area's groundwater level time series to obtain the multi-scale semantic encoding features of the target irrigation area's groundwater level time series.

[0040] Specifically, in S51, based on the set of target irrigation area groundwater level time-series implicit coding features, the target irrigation area groundwater level time-series graph walking topological features are extracted. That is, in the embodiments of the present application, first, each target irrigation area groundwater level time-series implicit coding feature vector in the set of target irrigation area groundwater level time-series implicit coding feature vectors is respectively input into a hyperbolic space mapper to obtain a set of target irrigation area groundwater level time-series implicit coding feature vectors after hyperbolic space mapping; furthermore, based on the set of target irrigation area groundwater level time-series implicit coding feature vectors after hyperbolic space mapping, graph walking topological features are extracted to obtain a target irrigation area groundwater level time-series graph walking topological feature matrix as the target irrigation area groundwater level time-series graph walking topological features. In the technical solution of the present application, the specific process of extracting graph walking topological features based on the set of target irrigation area groundwater level time-series implicit coding feature vectors after hyperbolic space mapping includes: calculating the Poincaré distance between any two target irrigation area groundwater level time-series implicit coding feature vectors after hyperbolic space mapping in the set of target irrigation area groundwater level time-series implicit coding feature vectors after hyperbolic space mapping to obtain a target irrigation area groundwater level time-series graph walking topological matrix; performing hole convolution coding on the target irrigation area groundwater level time-series graph walking topological matrix to obtain the target irrigation area groundwater level time-series graph walking topological feature matrix. By introducing hyperbolic space mapping and Poincaré distance calculation, the hierarchical structure and long-tail distribution in high-dimensional groundwater level time-series features can be better maintained, overcoming the problem that it is difficult to effectively measure different target irrigation area groundwater level time-series implicit coding feature vectors in a high-dimensional sparse space, and enhancing the discrimination ability and robustness of the target irrigation area groundwater level time-series feature representation. More specifically, based on the set of target irrigation area groundwater level time-series implicit coding features, the target irrigation area groundwater level time-series graph walking topological features are extracted with the following feature extraction formula; where the feature extraction formula is:

[0041]

[0042]

[0043]

[0044] Where, is the set of target irrigation area groundwater level time-series implicit coding feature vectors, , , , and are respectively the 1st, 2nd, th, th and th target irrigation area groundwater level time-series implicit coding feature vectors in the set of target irrigation area groundwater level time-series implicit coding feature vectors, and are the first weight matrix and the second weight matrix of the hyperbolic space respectively, is the corresponding time series implicit coding feature vector of the groundwater level in the target irrigation area after mapping in the hyperbolic space, is the corresponding time series implicit coding feature vector of the groundwater level in the target irrigation area after mapping in the hyperbolic space, the square of the vector one-norm, the inverse hyperbolic cosine function, is for calculating and the Poincaré distance between, is the eigenvalue at the position of in the target irrigation area groundwater level time series graph walk topological matrix, is the target irrigation area groundwater level time series graph walk topological matrix.

[0045] Specifically, in S52, based on the temporal graph-walking topological features of the groundwater level in the target irrigation area, significant gating weighted aggregation is performed on the set of implicit coding features of the groundwater level time series in the target irrigation area to obtain the multi-scale semantic coding features of the groundwater level time series in the target irrigation area. That is, in the embodiment of the present application, first, based on the temporal graph-walking topological feature matrix of the groundwater level in the target irrigation area, context semantic enhancement processing is performed on the set of implicit coding feature vectors of the groundwater level time series in the target irrigation area after hyperbolic space mapping to obtain a set of context semantic enhanced feature vectors of the groundwater level time series in the target irrigation area; in the technical solution of the present application, the specific process of performing context semantic enhancement processing on the set of implicit coding feature vectors of the groundwater level time series in the target irrigation area after hyperbolic space mapping includes: inputting the temporal graph-walking topological feature matrix of the groundwater level in the target irrigation area and the set of implicit coding feature vectors of the groundwater level time series in the target irrigation area after hyperbolic space mapping into the global context walking encoder based on the graph convolutional neural network model to obtain a set of context semantic enhanced feature vectors of the groundwater level time series in the target irrigation area. That is, by constructing the temporal graph-walking topological matrix of the groundwater level in the target irrigation area related to the graph structure, a more comprehensive understanding of the internal relationship between the temporal features of the groundwater level at different nodes can be achieved. By using dilated convolution and the global context walking encoder, this network architecture can effectively capture the feature information of the temporal semantics of the groundwater level in the target irrigation area at different scales, improve the modeling ability of the long-range dependence relationship between features, and thus improve the learning and understanding effect of the temporal graph-walking topological matrix of the groundwater level in the target irrigation area. Further, based on the set of context semantic enhanced feature vectors of the groundwater level time series in the target irrigation area, significant gating weighted aggregation is performed on the set of implicit coding feature vectors of the groundwater level time series in the target irrigation area to obtain the multi-scale semantic coding features of the groundwater level time series in the target irrigation area.In the technical solution of this application, the specific process of performing significant gating weighted aggregation on the set of target irrigation area groundwater level time series implicit coding feature vectors includes: inputting each corresponding target irrigation area groundwater level time series context semantic enhancement feature vector and target irrigation area groundwater level time series implicit coding feature vector in the set of target irrigation area groundwater level time series context semantic enhancement feature vectors and the set of target irrigation area groundwater level time series implicit coding feature vectors into an autocorrelation gating unit to obtain a set of target irrigation area groundwater level time series autocorrelation gating significant confidence factors; inputting the set of target irrigation area groundwater level time series autocorrelation gating significant confidence factors into a normalization unit based on the Softmax function to obtain a sequence of target irrigation area groundwater level time series autocorrelation gating significant confidence weight factors; based on the sequence of target irrigation area groundwater level time series autocorrelation gating significant confidence weight factors, calculating the position-wise weighted sum of the set of target irrigation area groundwater level time series implicit coding feature vectors to obtain a target irrigation area groundwater level time series multi-scale semantic coding feature vector as the target irrigation area groundwater level time series multi-scale semantic coding feature. In particular, the autocorrelation gating mechanism ensures that the network can dynamically adjust the weight distribution according to the importance of the groundwater level time series features represented by each node, thereby highlighting the key node feature semantics in the aggregation process, suppressing irrelevant or noisy features, improving the quality of the final target irrigation area groundwater level time series multi-scale semantic coding feature representation, providing richer and more accurate input information for subsequent target irrigation area groundwater level prediction, and helping to formulate more reasonable water resource management strategies. More specifically, based on the target irrigation area groundwater level time series graph walk topological features, the following significant gating weighted aggregation formula is used to perform significant gating weighted aggregation on the set of target irrigation area groundwater level time series implicit coding features to obtain the target irrigation area groundwater level time series multi-scale semantic coding feature; where, the significant gating weighted aggregation formula is:.

[0046]

[0047]

[0048]

[0049]

[0050]

[0051] Among them, is the dilated convolution coding, is the target irrigation area groundwater level time series graph walk topological feature matrix, is the graph convolution coding, is the A time-series context semantic enhancement feature vector of the groundwater level in a target irrigation area, is a function, for position-wise subtraction, which calculates and the autocorrelation gated significant confidence factor between them, where and is the autocorrelation gated significant confidence factor of the time series of the groundwater level in the target irrigation area between them, represents the exponential function value with the natural constant e as the base, is the number of autocorrelation gated significant confidence factors of the time series of the groundwater level in the target irrigation area in the set of autocorrelation gated significant confidence factors of the time series of the groundwater level in the target irrigation area, is the corresponding autocorrelation gated significant confidence weight factor of the time series of the groundwater level in the target irrigation area, is the number of vectors in the set of the implicit coding feature vectors of the time series of the groundwater level in the target irrigation area, for position-wise dot product, is the multi-scale semantic coding feature vector of the time series of the groundwater level in the target irrigation area.

[0052] Specifically, in S6, groundwater feature analysis is performed based on the multi-scale semantic coding features of the time series of the groundwater level in the target irrigation area to determine an estimated value of the decline of the groundwater level in the target irrigation area in the next 10 years. In the technical solution of this application, the multi-scale semantic coding feature vector of the time series of the groundwater level in the target irrigation area is input into a groundwater feature analysis module based on a decoder to obtain an analysis result, and the analysis result is used to represent the estimated value of the decline of the groundwater level in the target irrigation area in the next 10 years. That is to say, decoding regression is performed using the multi-scale aggregation representation of the time series of the groundwater level in the target irrigation area to predict the estimated value of the decline of the groundwater level in the target irrigation area in the next 10 years. In this way, it is possible to predict the dynamic changes of the groundwater level based on the spatio-temporal dynamic evolution characteristics of the groundwater in the target irrigation area, and thereby determine whether it will affect the sustainable development of agricultural production and the sustainable utilization of water resources, providing more scientific and accurate support for water resource management and decision-making.

[0053] Considering that each target irrigation area groundwater level time-series implicit coding feature vector in the set of target irrigation area groundwater level time-series implicit coding feature vectors represents the time-series periodic coding features of the target irrigation area groundwater level within different time-series windows, when performing feature sequence aggregation analysis based on graph walk self-correlation gating, the heterogeneity of the time-series periodic features within different time-series windows will lead to the lack of instance decision-making of the aggregation features of the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors, thus affecting the accuracy of the analysis results obtained by inputting them into the groundwater feature analysis module based on the decoder.

[0054] Based on this, before inputting the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors into the groundwater feature analysis module based on the decoder, first optimize the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors, including the steps of:

[0055] Calculate the distance between each pair of eigenvalues of the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors, such as the L2 distance, and take the square root of the distance to obtain the target irrigation area groundwater level time-series multi-scale semantic coding distance representation matrix, that is

[0056]

[0057] where represents the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors, and respectively represent the i-th and j-th eigenvalues of the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors, represents the distance between the i-th and j-th eigenvalues of the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors, represents the target irrigation area groundwater level time-series multi-scale semantic coding distance representation matrix;

[0058] Obtain the target irrigation area groundwater level time-series multi-scale semantic coding self-correlation matrix of the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors as row vectors, that is ;

[0059] Multiply the target irrigation area groundwater level time-series multi-scale semantic coding feature vectors by the target irrigation area groundwater level time-series multi-scale semantic coding distance representation matrix to obtain the target irrigation area groundwater level time-series multi-scale semantic coding first-level mapping vector, that is ;

[0060] Multiply the first-level mapping vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area by the matrix product of the distance representation matrix of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area and the self-correlation matrix of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area to obtain the multi-level mapping vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area ;

[0061] Dot-multiply the multi-level mapping vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area by the correlation eigenvector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area, which is composed of the eigenvalues of the self-correlation matrix of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area, to obtain an optimized feature vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area, where interpolation or zero-padding is performed when the eigenvalues are insufficient

[0062] That is, through the linear target mapping representation of the similarity distance representation matrix based on the feature vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area, the complete similarity instantiation of the self-correlation of the feature vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area is performed for the quadratic target mapping representation based on the multi-level distribution hierarchy, and the negative influence factor of the association mismatch is compensated by the associated fusion kernel bias, so as to improve the eigenvalue instance decision degree of the feature vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area under the similarity constraint, that is, the significance degree of the eigenvalue as an instance for the decoding regression decision, and improve the accuracy of the analysis result obtained by inputting the feature vector of the time-series multi-scale semantic encoding of the groundwater level in the target irrigation area into the groundwater feature analysis module based on the decoder. In this way, it is possible to predict the dynamic change of the groundwater level decline in the next 10 years based on the spatio-temporal dynamic evolution characteristics of the groundwater in the target irrigation area, and determine whether it will affect the sustainable development of agricultural production and the sustainable utilization of water resources, providing more scientific and accurate support for water resource management and decision-making

[0063] In summary, the method for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution according to the embodiments of the present application is elucidated. It collects the time-series dataset of groundwater data in the target irrigation area located in arid and semi-arid regions, extracts the time-series dataset of the water level values related to groundwater from it, and then introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the backend to perform time-series aggregation analysis on these water level values, so as to capture the time-series multi-scale semantic feature representation of the groundwater level in the target irrigation area, and predict the estimated value of the groundwater level decline in the target irrigation area in the next 10 years based on this time-series groundwater level feature representation. In this way, it is possible to predict the dynamic change of the groundwater level based on the spatio-temporal dynamic evolution characteristics of the groundwater in the target irrigation area, and determine whether it will affect the sustainable development of agricultural production and the sustainable utilization of water resources, providing more scientific and accurate support for water resource management and decision-making

[0064] Furthermore, a system for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution is also provided.

[0065] Figure 4 It is a block diagram of a system for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution according to an embodiment of the present application. As Figure 4 shown, the system 300 for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution according to an embodiment of the present application includes: a data acquisition module 310, configured to acquire a time-series dataset of groundwater data of a target irrigation area, where the groundwater data includes water volume data and water quality data, the water volume data includes a water level value and a water volume value, and the water quality data includes salinity, pH value, total hardness, salt content, ionic total, and nitrogen, phosphorus, and potassium; a water level value extraction module 320, configured to extract the water level value from the time-series dataset of the groundwater data of the target irrigation area to obtain a time-series dataset of the water level value; a singular spectrum decomposition module 330, configured to perform singular spectrum decomposition on the time-series dataset of the water level value to obtain a set of subsequences of the groundwater level in the target irrigation area; a time-series feature extraction module 340, configured to perform time-series feature extraction based on the groundwater level on each subsequence of the groundwater level in the target irrigation area in the set of subsequences of the groundwater level in the target irrigation area to obtain a set of time-series implicit coding features of the groundwater level in the target irrigation area; a significant aggregation analysis module 350, configured to perform graph-walking significant aggregation analysis on the set of time-series implicit coding features of the groundwater level in the target irrigation area to obtain a set of time-series multi-scale semantic coding features of the groundwater level in the target irrigation area; and a groundwater characteristic analysis module 360, configured to perform groundwater characteristic analysis based on the set of time-series multi-scale semantic coding features of the groundwater level in the target irrigation area to determine an estimated value of the decline of the groundwater level in the target irrigation area within the next 10 years.

[0066] As described above, the system 300 for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution according to an embodiment of the present application can be implemented in various wireless terminals, such as a server with an algorithm for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution. In a possible implementation, the system 300 for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution according to an embodiment of the present application can be integrated into a wireless terminal as a software module and / or a hardware module. For example, the system 300 for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution can be a software module in the operating system of the wireless terminal, or can be an application program developed for the wireless terminal; of course, the system 300 for analyzing the characteristics of groundwater in irrigation areas based on spatio-temporal evolution can also be one of many hardware modules of the wireless terminal.

[0067] Alternatively, in another example, the irrigation district groundwater feature analysis system 300 based on spatio-temporal evolution and the wireless terminal may also be separate devices, and the irrigation district groundwater feature analysis system 300 based on spatio-temporal evolution can be connected to the wireless terminal through a wired and / or wireless network and transmit interaction information in accordance with a predefined data format.

[0068] In a specific example of this application, the collection and collation of technical data on water and soil resources in the irrigation district are carried out. Using methods such as statistics, the spatio-temporal evolution characteristics of the quantity and quality of surface and groundwater resources are analyzed, the dynamic changes in the supply and water use structures and water use efficiency are analyzed, and the changing trend of the degree of water resource development and utilization is revealed; the dynamic analysis model of land use is used to study the spatio-temporal dynamic evolution characteristics of land use and cover in the irrigation district in the past 10 years from three aspects: degree, quantity, and type transfer; the water conveyance processes of irrigation and drainage in the irrigation district are understood, the change in the irrigation water use coefficient in the irrigation district is monitored, and a multi-dimensional digital base plate based on environmental element data (precipitation, evapotranspiration, temperature, runoff), basic data of the irrigation district (water diversion volume, drainage volume, canal system buildings), and real-time monitoring data is constructed to provide support for the construction of a digital irrigation district.

[0069] This application intends to balance step by step at the irrigation area - irrigation district scale, analyze the spatio-temporal evolution characteristics of the quantity of surface and groundwater resources, analyze the dynamic changes in the supply and water use structures and water use efficiency, and reveal the changing trend of the degree of water resource development and utilization; analyze the dynamic changes in the supply and water use structures and water use efficiency, and reveal the changing trend of the degree of water resource development and utilization; measure the water quality at representative points in the irrigation district through field sampling and laboratory experiments, analyze the spatio-temporal variation law and evolution characteristics of surface and groundwater quality in the entire irrigation district, and provide a basis and data support for the rational development and sustainable utilization of water resources in the irrigation district and the optimal allocation of water resources.

[0070] The embodiments of the present disclosure have been described above. The above description is exemplary and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art in the technical field without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technologies in the market, or to enable other ordinary technical personnel in the technical field to understand the disclosed embodiments.

Claims

1. A method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution, characterized in that: include: Acquire a time series data set of groundwater data of a target irrigation area, wherein the groundwater data includes water quantity data and water quality data, wherein the water quantity data includes water level value and water quantity value, and the water quality data includes mineralization, pH value, total hardness, salt content, ion composition, and nitrogen, phosphorus, and potassium; Extracting water level values ​​from a time series data set of groundwater data in the target irrigation area to obtain a time series data set of water level values; Performing singular spectrum decomposition on the time series data set of the water level value to obtain a set of groundwater level subsequences of the target irrigation area; Extracting time series features based on groundwater level for each target irrigation area groundwater level subsequence in the set of target irrigation area groundwater level subsequences to obtain a set of target irrigation area groundwater level time series implicit coding features; The set of implicit coding features of the target irrigation area groundwater level time series is subjected to graph walk significant aggregation analysis to obtain the target irrigation area groundwater level time series multi-scale semantic coding features, including: extracting the target irrigation area groundwater level time series graph walk topology features based on the set of implicit coding features of the target irrigation area groundwater level time series; based on the target irrigation area groundwater level time series graph walk topology features, performing significant gated weighted aggregation on the set of implicit coding features of the target irrigation area groundwater level time series to obtain the target irrigation area groundwater level time series multi-scale semantic coding features; Based on the multi-scale semantic coding characteristics of the groundwater level time series in the target irrigation area, groundwater characteristic analysis is performed to determine the estimated value of the groundwater level drop in the target irrigation area in the next 10 years.

2. The method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution according to claim 1 is characterized in that: Performing groundwater level-based time series feature extraction on each target irrigation area groundwater level subsequence in the set of target irrigation area groundwater level subsequences to obtain a set of target irrigation area groundwater level time series implicit coding features, including: inputting each target irrigation area groundwater level subsequence in the set of target irrigation area groundwater level subsequences into a groundwater level time series feature extractor based on an LSTM-RNN hybrid model to obtain a set of target irrigation area groundwater level time series implicit coding feature vectors as the set of target irrigation area groundwater level time series implicit coding features.

3. The method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution according to claim 2 is characterized in that: Based on the set of implicit coding features of the groundwater level time series of the target irrigation area, the wandering topological features of the groundwater level time series graph of the target irrigation area are extracted, including: Inputting each target irrigation area groundwater level time series implicit coding feature vector in the set of target irrigation area groundwater level time series implicit coding feature vectors into a hyperbolic space mapper to obtain a set of target irrigation area groundwater level time series implicit coding feature vectors after hyperbolic space mapping; Based on the set of implicitly encoded feature vectors of the target irrigation area groundwater level time series after the hyperbolic space mapping, the graph walk topology features are extracted to obtain the target irrigation area groundwater level time series graph walk topology feature matrix as the target irrigation area groundwater level time series graph walk topology features.

4. The method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution according to claim 3 is characterized in that: Based on the set of implicitly encoded feature vectors of the target irrigation area groundwater level time series after the hyperbolic space mapping, the graph walk topology features are extracted to obtain the target irrigation area groundwater level time series graph walk topology feature matrix as the target irrigation area groundwater level time series graph walk topology features, including: Calculate the Poincare distance between any two implicitly coded feature vectors of the target irrigation area groundwater level time series after hyperbolic space mapping in the set of implicitly coded feature vectors of the target irrigation area groundwater level time series after hyperbolic space mapping to obtain a walk topology matrix of the target irrigation area groundwater level time series graph; The walk topology matrix of the target irrigation area groundwater level time series graph is subjected to dilated convolution coding to obtain the walk topology feature matrix of the target irrigation area groundwater level time series graph.

5. The method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution according to claim 4 is characterized in that: Based on the wandering topological features of the target irrigation area groundwater level time series graph, a set of implicit coding features of the target irrigation area groundwater level time series is significantly gated and weighted aggregated to obtain multi-scale semantic coding features of the target irrigation area groundwater level time series, including: Based on the walk topology feature matrix of the target irrigation area groundwater level time series graph, a set of implicitly encoded feature vectors of the target irrigation area groundwater level time series after the hyperbolic space mapping is subjected to contextual semantic enhancement processing to obtain a set of contextual semantic enhancement feature vectors of the target irrigation area groundwater level time series; Based on the set of contextual semantic enhancement feature vectors of the target irrigation area groundwater level time series, a significant gated weighted aggregation is performed on the set of implicit coding feature vectors of the target irrigation area groundwater level time series to obtain the multi-scale semantic coding features of the target irrigation area groundwater level time series.

6. The method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution according to claim 5 is characterized in that: Based on the walk topology feature matrix of the target irrigation area groundwater level time series graph, a set of implicitly encoded feature vectors of the target irrigation area groundwater level time series after the hyperbolic space mapping is subjected to contextual semantic enhancement processing to obtain a set of contextual semantic enhancement feature vectors of the target irrigation area groundwater level time series, including: The walk topological feature matrix of the target irrigation area groundwater level time series graph and the set of implicitly encoded feature vectors of the target irrigation area groundwater level time series after hyperbolic space mapping are input into the global context walk encoder based on the graph convolutional neural network model to obtain the set of context semantic enhancement feature vectors of the target irrigation area groundwater level time series.

7. The method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution according to claim 6 is characterized in that: Based on the set of semantically enhanced feature vectors of the target irrigation area groundwater level time series context, a significant gated weighted aggregation is performed on the set of implicitly encoded feature vectors of the target irrigation area groundwater level time series to obtain multi-scale semantic encoding features of the target irrigation area groundwater level time series, including: Inputting each corresponding set of the target irrigation area groundwater level time series context semantic enhancement feature vector and the target irrigation area groundwater level time series implicit coding feature vector in the set of the target irrigation area groundwater level time series context semantic enhancement feature vector and the target irrigation area groundwater level time series implicit coding feature vector into the autocorrelation gating unit to obtain a set of target irrigation area groundwater level time series autocorrelation gating significant confidence factors; Inputting the set of the time series autocorrelation gated significant confidence factors of the groundwater level in the target irrigation area into a normalization unit based on the Softmax function to obtain a sequence of the time series autocorrelation gated significant confidence weight factors of the groundwater level in the target irrigation area; Based on the sequence of autocorrelation gated significant confidence weight factors of the target irrigation area groundwater level time series, the position-weighted sum of the set of implicit coding feature vectors of the target irrigation area groundwater level time series is calculated to obtain the multi-scale semantic coding feature vector of the target irrigation area groundwater level time series as the multi-scale semantic coding feature of the target irrigation area groundwater level time series.

8. The method for analyzing groundwater characteristics in irrigation areas based on spatiotemporal evolution according to claim 7 is characterized in that: Based on the multi-scale semantic coding characteristics of the time series of groundwater levels in the target irrigation area, groundwater characteristic analysis is performed to determine the estimated value of the groundwater level drop in the target irrigation area within the next 10 years, including: inputting the multi-scale semantic coding feature vector of the time series of groundwater levels in the target irrigation area into a decoder-based groundwater characteristic analysis module to obtain an analysis result, and the analysis result is used to represent the estimated value of the groundwater level drop in the target irrigation area within the next 10 years.

9. A groundwater characteristic analysis system for irrigation areas based on spatiotemporal evolution, characterized in that: include: A data acquisition module is used to acquire a time series data set of groundwater data of a target irrigation area, wherein the groundwater data includes water quantity data and water quality data, wherein the water quantity data includes water level value and water quantity value, and the water quality data includes mineralization, pH value, total hardness, salt content, ion composition, and nitrogen, phosphorus, and potassium; A water level value extraction module, used to extract water level values ​​from a time series data set of groundwater data of the target irrigation area to obtain a time series data set of water level values; A singular spectrum decomposition module, used for performing singular spectrum decomposition on the time series data set of the water level value to obtain a set of subsequences of groundwater levels in the target irrigation area; A time series feature extraction module, used for extracting time series features based on groundwater level for each target irrigation area groundwater level subsequence in the set of target irrigation area groundwater level subsequences to obtain a set of implicit coding features of the time series of groundwater level in the target irrigation area; A significant aggregation analysis module is used to perform a graph walk significant aggregation analysis on the set of implicit coding features of the target irrigation area groundwater level time series to obtain multi-scale semantic coding features of the target irrigation area groundwater level time series; The groundwater characteristic analysis module is used to perform groundwater characteristic analysis based on the multi-scale semantic coding characteristics of the groundwater level time series of the target irrigation area to determine the estimated value of the groundwater level drop in the target irrigation area in the next 10 years.

Images