A method, device, medium and product for analyzing driving factors of cultivated land degradation

Abstract

The present application discloses a method, device, medium and product for analyzing driving factors of cultivated land degradation, which relates to the technical field of agricultural cultivated land data analysis. The method comprises obtaining a dataset for analyzing driving factors of cultivated land degradation in a target area; performing rasterization processing on the dataset to obtain raster data; performing temporal causal reasoning using a CCM model based on time series data to obtain a time-based causal direction and causal relationship value between each two explanatory variables; performing spatial causal reasoning using a GCCM model based on spatial data to obtain a space-based causal direction and causal relationship value between each two explanatory variables and between the explanatory variables and cultivated land degradation data; constructing a causal relationship graph based on the causal direction and causal relationship value based on time and space as known relationships; and using the causal relationship graph in an SCM model to obtain a result of analyzing cultivated land degradation. The present application can improve the accuracy of analyzing driving factors of cultivated land degradation.

Description

Technical Field

[0001] The present application relates to the technical field of agricultural land data analysis, and in particular to a method, equipment, medium and product for analyzing driving factors of land degradation based on causal inference. Background Art

[0002] Cultivated land is one of the main types of land cover and an important resource for ensuring food security, ecological security and sustainable development. Black soil refers to land with a black or dark humus topsoil layer. It is a high-quality land with good properties, high fertility and suitable for farming. Although it accounts for less than 7% of the world's total cultivated land area, it feeds 18% of the world's population and is a very scarce soil resource. However, with the long-term reclamation and high-intensity use of black soil by humans, the natural conditions for the formation of black soil have been completely changed. Frequent plowing and surface exposure during reclamation and planting have caused profound changes in the surface process and organic matter accumulation process of black soil, resulting in varying degrees of degradation of black soil quality. In recent years, the black soil in Northeast China has shown significant degradation phenomena such as "thinning", "slimming" and "hardening". Protecting cultivated land is to protect survival, development and the future. In order to avoid the serious threat of cultivated land degradation to the human living environment, only by fully understanding the mechanism of cultivated land degradation can we better configure prevention and control measures. Cultivated land degradation is affected by a variety of driving factors, and it is necessary to fully and accurately understand the causal relationship between driving factors and dependent variables. Therefore, in order to improve the causal analysis results of the driving factors of cultivated land degradation in complex dynamic systems, a method is needed that can directly reflect the spatiotemporal degradation analysis of cultivated land by multiple driving factors.

[0003] The process of cultivated land degradation is a complex dynamic system that is affected by multiple couplings of natural and human factors, including natural factors such as temperature, precipitation, soil type, soil texture, topography, land cover, and human factors such as GDP, population density, and transportation network. There is a complex coupling relationship between the influencing factors in the complex dynamic system, and the spatial distribution of cultivated land quality degradation also shows a certain degree of spatial correlation and spatial heterogeneity. Spatially, land in adjacent areas or under similar environmental conditions is prone to similar quality problems, and there are significant differences in the quality of cultivated land in different regions. Attribution analysis in complex dynamic fields is conducive to revealing the coupling effects of various driving factors on each other and on cultivated land degradation.

[0004] At present, there are three main methods for analyzing the driving factors of cultivated land degradation. The first is to use statistical models such as traditional linear regression analysis and principal component analysis. The second is to use machine learning models such as random forest regression and support vector machine regression. The third is to use economic models such as system dynamics model and economic equilibrium method. These models are suitable for prediction, and the relationship between each independent variable and dependent variable can be inferred through prediction. However, their analytical results generally do not consider the relationship between driving factors, which is easy to confuse the interaction of variables, which will lead to deviations in the analytical results of the driving factors of cultivated land degradation and make it difficult to propose direct and effective configuration and prevention measures for cultivated land degradation. Therefore, the above three methods have the following disadvantages: (1) linear correlation is prone to false correlation; (2) lack of consideration of spatial correlation and spatial heterogeneity; (3) when the correlation between two variables is significant and strong, it is impossible to distinguish the causal direction between the independent variable and the dependent variable; (4) the complex coupling relationship between multiple variables cannot be explored. These disadvantages directly lead to inaccurate and unreliable analytical results of the driving factors of cultivated land degradation.

[0005] In summary, the existing methods for analyzing the driving factors of cultivated land degradation generally have the problem of low accuracy. Based on this, how to improve the accuracy of the analysis of the driving factors of cultivated land degradation has become a technical problem that needs to be urgently solved in this field. Summary of the invention

[0006] The purpose of this application is to provide a method, equipment, medium and product for analyzing the driving factors of cultivated land degradation, which can improve the accuracy of the analysis of the driving factors of cultivated land degradation.

[0007] To achieve the above objectives, this application provides the following solutions:

[0008] In a first aspect, the present application provides a method for analyzing driving factors of cultivated land degradation, and the method for analyzing driving factors of cultivated land degradation comprises:

[0009] Obtain a dataset of driving factors for cultivated land degradation in a target area; the dataset includes cultivated land degradation data and corresponding explanatory variable data, the explanatory variable data are data of various explanatory variables affecting cultivated land degradation in the target area, and the explanatory variables include natural factor variables and human factor variables.

[0010] The cultivated land degradation driving factor analysis data set is subjected to rasterization processing to obtain raster data, wherein the raster data includes time series data and spatial data in raster format.

[0011] According to the time series data, the CCM model is used to perform temporal causal reasoning to obtain the time-based causal direction and causal relationship value between every two explanatory variables.

[0012] According to the spatial data, the GCCM model is used to perform spatial causal reasoning to obtain the spatially based causal direction and causal relationship value between every two of the explanatory variables and between the explanatory variables and the cultivated land degradation data.

[0013] A causal relationship graph is constructed based on the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value as known relationships.

[0014] According to the cause-effect relationship diagram, the cause-effect relationship diagram is used in the SCM model to perform farmland degradation analysis on the target area to obtain a farmland degradation analysis result.

[0015] Optionally, the cultivated land degradation driving factor analysis dataset is subjected to rasterization processing to obtain raster data, specifically including:

[0016] From the cultivated land degradation driving factor analysis data set, cultivated land degradation data and explanatory variable data within the same time range are selected.

[0017] The cultivated land degradation data and all explanatory variable data are uniformly converted into raster data with the same spatial resolution, and the pixel positions of the raster data are completely consistent.

[0018] Optionally, according to the time series data, a CCM model is used to perform temporal causal reasoning to obtain a time-based causal direction and a causal relationship value between every two of the explanatory variables, specifically including:

[0019] The time series data are grouped in pairs according to the types of the explanatory variables to obtain multiple groups of time series data with different explanatory variables.

[0020] Multiple groups of time series data of different explanatory variables are sequentially input into the CCM model, and the CCM model is used to perform causal inference on the time series data of each group of different explanatory variables in two cross-mapping directions to obtain the causal relationship value with causal direction between the two explanatory variables in each group of time series data; the CCM model is used to generate attractor reconstruction by time delay embedding on a time series, and the simplex projection algorithm is used to estimate the concurrent value of another time series to determine the causal relationship with direction between the two explanatory variables corresponding to each group of time series data.

[0021] Optionally, multiple groups of time series data with different explanatory variables are sequentially input into the CCM model, and the CCM model is used to perform causal inference on the time series data of each group of different explanatory variables in two cross-mapping directions to obtain a causal relationship value with a causal direction between two of the explanatory variables in each group of time series data, specifically including:

[0022] Each of the time series data is normalized respectively, and the Person correlation coefficient between every two of the time series data is calculated.

[0023] Create an index vector for each of the time series data, and form a lagged coordinate vector based on the time series data as a reconstructed manifold M x , in the reconstructed manifold M x Based on the variable positioning, the lag coordinate vector is Y'(t)|M x Represents , and determines the time lag E+1 of the nearest neighbor.

[0024] Calculate the nearest neighbor time index of each explanatory variable X, and use the nearest neighbor time index of the explanatory variable X to determine the nearest neighbor time index of the explanatory variable Y.

[0025] The lagged coordinate vector formed by the time series data corresponding to the explanatory variable Y is used as the reconstructed manifold M y , in the reconstructed manifold M y Based on the variable positioning, the lag coordinate vector is X'(t)|M y Represents , and determines the time lag E+1 of the nearest neighbor.

[0026] The nearest neighbor time index of the explanatory variable Y is calculated from near to far, and the nearest neighbor time index of the explanatory variable X is determined using the nearest neighbor time index of the explanatory variable Y.

[0027] According to the nearest neighbor time index of the explanatory variable Y and the nearest neighbor time index of the explanatory variable X, the CCM model is used to reconstruct the manifold M. x Cross-mapping the explanatory variables Y and reconstructing the manifold M y In the two cross-mapping directions of the cross-mapping explanatory variable X, causal inference is performed on the time series data of the explanatory variable X and the explanatory variable Y to obtain the causal relationship value with causal direction of the explanatory variable X and the explanatory variable Y, and the causal relationship value is compared with the Person correlation coefficient.

[0028] Optionally, according to the spatial data, a GCCM model is used to perform spatial causal reasoning to obtain spatially based causal directions and causal relationship values ​​between every two of the explanatory variables and between the explanatory variables and the cultivated land degradation data, specifically including:

[0029] Matrix conversion processing is performed on the explanatory factor data and the cultivated land degradation data in the spatial data to obtain spatial data after the matrix conversion processing.

[0030] Set the size of the sliding window according to the length and width of the matrix.

[0031] Based on the spatial data after the matrix conversion processing and the size of the sliding window, the spatial lag value is set, and according to the spatial autocorrelation theory, the spatial data in the matrix is ​​screened with a fixed length to obtain the spatial data in the screened matrix format.

[0032] All spatial data in matrix format are processed to eliminate linear trends, including removing the linear trend of cultivated land degradation, to obtain spatial data after removing linear trends.

[0033] According to the spatial data after removing the linear trend, a cross-mapping prediction method is used to predict any one of the explanatory variables according to the reconstructed manifold M from the explanatory variable. x The nearest neighbors identified in are used to predict the value of the explanatory variable y, and the true observed value of the explanatory variable and the corresponding predicted value are obtained.

[0034] According to the actual observed value of the explanatory variable and the corresponding predicted value, the Pearson correlation coefficient between the actual observed value of the explanatory variable and the corresponding predicted value is calculated.

[0035] The results of the cross-mapping predictions are measured based on the Pearson correlation coefficient between the true observed values ​​of the explanatory variables and the corresponding predicted values.

[0036] Optionally, a causal relationship graph is constructed based on the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value as known relationships, specifically including:

[0037] The causal relationship values ​​between each two explanatory variables in the time series data are compared with the first preset threshold value. When the causal relationship values ​​of the two causal directions are both smaller than the first preset threshold value, it is determined that there is no causal relationship between the two explanatory variables. When one or both of the causal relationship values ​​of the two causal directions are greater than the first preset threshold value, the two causal relationship values ​​are compared, and the causal direction with the larger causal relationship value is selected as a causal relationship for establishing a causal relationship diagram.

[0038] The causal relationship values ​​between every two explanatory variables in the spatial data and between the explanatory variables and the cultivated land degradation data are compared with the second preset threshold value. When the causal relationship values ​​in the two causal directions are both less than the second preset threshold value, it is determined that there is no causal relationship between the two explanatory variables; when one or both of the causal relationship values ​​in the two causal directions are greater than the second preset threshold value, the two causal relationship values ​​are compared, and the causal direction with the larger causal relationship value is selected as a causal relationship for establishing a causal relationship diagram, so as to obtain a causal relationship diagram between the explanatory variables and between the explanatory variables and the cultivated land degradation data.

[0039] According to the cause-effect relationship diagram, the cause-effect relationship diagram is used in the SCM model to perform farmland degradation analysis on the target area, and the farmland degradation analysis results are obtained, which specifically include:

[0040] All cultivated land degradation data and explanatory variable data are counted in the same EXCEL table to obtain EXCEL data.

[0041] The EXCEL data and the causal relationship diagram are simultaneously input into the SCM model, a causal relationship network is established using the SCM model, and the causal effect values ​​between causal relationships are output.

[0042] Optionally, the natural factor variables include temperature, precipitation, NDVI, DEM, sand, silt and clay; the human factor variables include GDP, population density and transportation network.

[0043] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-mentioned methods for analyzing driving factors of farmland degradation.

[0044] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned methods for analyzing driving factors of farmland degradation.

[0045] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for analyzing driving factors of farmland degradation.

[0046] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0047] The present application provides a method, device, medium and product for analyzing driving factors of cultivated land degradation. Time series data in raster format is obtained through cultivated land degradation data and explanatory variable data. Temporal causal reasoning is first performed through the CCM model to obtain the time-based causal direction and causal relationship value between every two explanatory variables. Then, spatial causal reasoning is performed through the GCCM model to obtain the space-based causal direction and causal relationship value between every two explanatory variables and between the explanatory variables and the cultivated land degradation data. Then, a causal relationship diagram is constructed, and cultivated land degradation analysis is performed in the SCM model according to the causal relationship diagram to obtain accurate cultivated land degradation analysis results. This application combines the CCM model and the GCCM model for the scenario of soil degradation analysis. First, the CCM model based on the simplex projection algorithm is used for time series variables to verify the unidirectional causal relationship between the explanatory variables. Then, the GCCM model is used for spatial causal reasoning to reveal the causal relationship between spatial cross-sectional variables, so that the corresponding causal effect can be estimated. Finally, the causal relationship and causal direction based on time and space obtained based on the CCM model and the GCCM model are used as known causal relationships to provide a causal relationship diagram for the SCM model, and further realize the analysis of soil degradation, so as to truly reflect the correlation between the various explanatory variables that affect cultivated land degradation, effectively improve the accuracy of the analysis of the driving factors of cultivated land degradation, and solve the problem of inaccurate and unreliable analysis results of the driving factors of cultivated land degradation in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A schematic flow chart of a method for analyzing driving factors of cultivated land degradation provided in one embodiment of the present application;

[0050] Figure 2 A schematic diagram of convergent cross-mapping results of temperature and precipitation provided in an embodiment of the present application;

[0051] Figure 3 A schematic diagram of the convergence cross-mapping results of temperature and NDVI provided in one embodiment of the present application;

[0052] Figure 4 A schematic diagram of the convergence cross-mapping results of precipitation and NDVI provided in an embodiment of the present application;

[0053] Figure 5A schematic diagram of a spatial center unit and its spatial hysteresis in different grid data orders provided by an embodiment of the present application;

[0054] Figure 6 A schematic diagram of a corresponding manifold for temperature spatial hysteresis reconstruction provided by an embodiment of the present application;

[0055] Figure 7 A schematic diagram of a corresponding manifold of a delayed reconstruction of sand space provided by an embodiment of the present application;

[0056] Figure 8 A schematic diagram of the causal relationship between temperature and soil organic matter degradation provided in one embodiment of the present application;

[0057] Fig. 9 A schematic diagram of the cause-effect relationship between clay and soil organic matter degradation provided in one embodiment of the present application;

[0058] Fig.10 A cause-and-effect relationship diagram generated based on the CCM model and the GCCM model provided in one embodiment of the present application. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0060] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0061] like Figure 1 As shown, this embodiment provides a method for analyzing driving factors of cultivated land degradation, which is applied to the scenario of analyzing cultivated land degradation in the target area, mainly analyzing various factors affecting cultivated land degradation, wherein the driving factors refer to the factors that drive cultivated land degradation, that is, the factors that affect cultivated land degradation, also known as explanatory factors, that is, the factors used to explain why cultivated land degrades. In this embodiment, the method for analyzing driving factors of cultivated land degradation specifically includes the following steps:

[0062] Step S1, obtaining a dataset of driving factors of cultivated land degradation in a target area; the dataset of driving factors of cultivated land degradation includes cultivated land degradation data and corresponding explanatory variable data, the explanatory variable data is data of various explanatory variables that affect cultivated land degradation in the target area, and the explanatory variables include natural factor variables and human factor variables.

[0063] In this embodiment, natural factor variables include temperature, precipitation, NDVI, DEM, sand, silt, clay, etc. Human factor variables include GDP, population density, and transportation network, etc.

[0064] Step S2: rasterizing the cultivated land degradation driving factor analysis dataset to obtain raster data, wherein the raster data includes time series data and spatial data in a raster format.

[0065] In this embodiment, raster data is a data form that divides space into regular grids, each grid is called a grid unit, and a corresponding attribute value is assigned to each grid unit. Step S2 of this embodiment is mainly to obtain time series data in a raster format, which is a raster data in a time dimension and is mainly used in the data processing stage of a CCM (Convergent Cross Mapping) model, while spatial data is a raster data in a spatial dimension, which contains information representing the location, shape, size distribution, etc. of an object and is mainly used in the data processing stage of a GCCM (Geographical Convergent Cross Mapping) model.

[0066] Step S3: Based on the time series data, the CCM model is used to perform temporal causal reasoning to obtain the time-based causal direction and causal relationship value between every two explanatory variables.

[0067] Step S4: Based on the spatial data, the GCCM model is used to perform spatial causal reasoning to obtain the spatially based causal direction and causal relationship value between every two of the explanatory variables and between the explanatory variables and the cultivated land degradation data.

[0068] Step S5: According to the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value, the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value are used as known causal relationships to construct a causal relationship graph.

[0069] Step S6, according to the causal relationship diagram, the causal relationship diagram is used in the SCM model to perform farmland degradation analysis on the target area, and the farmland degradation analysis result is obtained. Among them, the SCM model is a conceptual model that describes the causal mechanism, and the causal relationship between variables can be described by a causal relationship diagram. The causal relationship diagram is a directed graph structure used to represent the causal relationship. For example, a variable can be used as a node. If a variable x is a child node of another variable y, then variable y is the direct cause of variable x; if variable x is a descendant of variable y, then variable y is a potential cause of variable x. In this embodiment, on the basis of obtaining the causal relationship diagram, a causal relationship network is established through the SCM model according to the causal relationship diagram and all the farmland degradation data and explanatory variable data, and the causal effect value between the causal relationships is output.

[0070] In this embodiment, step S2 performs rasterization processing on the farmland degradation driving factor analysis dataset to obtain time series data in raster format, specifically including:

[0071] Step S21: Selecting cultivated land degradation data and explanatory variable data within the same time range from the cultivated land degradation driving factor analysis data set.

[0072] Step S22: uniformly converting the cultivated land degradation data and the explanatory variable data into raster data of the same spatial resolution, and the pixel positions of the raster data are completely consistent.

[0073] In this embodiment, step S3 uses the CCM model to perform temporal causal reasoning based on the time series data to obtain the time-based causal direction and causal relationship value between every two explanatory variables, which specifically includes the following steps:

[0074] Step S31: group the time series data in pairs according to the types of the explanatory variables to obtain multiple groups of time series data with different explanatory variables.

[0075] Step S32, input multiple groups of time series data of different explanatory variables into the CCM model in sequence, and use the CCM model to perform causal inference on the time series data of each group of different explanatory variables in two cross-mapping directions to obtain the causal relationship value with causal direction between the two explanatory variables in each group of time series data; the CCM model is used to generate attractor reconstruction using time delay embedding on a time series, and use the simplex projection algorithm to estimate the concurrent value of another time series, so as to determine the causal relationship with direction between the two explanatory variables corresponding to each group of time series data.

[0076] In this embodiment, step S32 sequentially inputs multiple groups of time series data of different explanatory variables into the CCM model, and uses the CCM model to perform causal inference on the time series data of each group of different explanatory variables in two cross-mapping directions to obtain a causal relationship value with a causal direction between two explanatory variables in each group of time series data, specifically including the following steps:

[0077] Step S321: normalize each of the time series data respectively, and calculate the Person correlation coefficient between every two of the time series data.

[0078] Step S322: Create index vectors for each of the time series data, and form a lagged coordinate vector based on the time series data as a reconstructed manifold M. x , in the reconstructed manifold M x Based on the variable positioning, the lag coordinate vector is Y'(t)|M x Represents , and determines the time lag E+1 of the nearest neighbor.

[0079] Step S323: Calculate the nearest neighbor time index of each explanatory variable X, and use the nearest neighbor time index of the explanatory variable X to identify the nearest neighbor time index of the explanatory variable Y.

[0080] Step S324: The lagged coordinate vector formed by the time series data corresponding to the explanatory variable Y is used as the reconstructed manifold M y , in the reconstructed manifold M y Based on the variable positioning, the lag coordinate vector is X'(t)|M y Represents , and determines the time lag E+1 of the nearest neighbor.

[0081] Step S325: Calculate the nearest neighbor time index of the explanatory variable Y from near to far, and use the nearest neighbor time index of the explanatory variable Y to identify the nearest neighbor time index of the explanatory variable X.

[0082] Step S326: Based on the nearest neighbor time index of the explanatory variable Y and the nearest neighbor time index of the explanatory variable X, the CCM model is used to reconstruct the manifold M. x Cross-mapping the explanatory variables Y and reconstructing the manifold M y In the two cross-mapping directions of the cross-mapping explanatory variable X, causal inference is performed on the time series data of the explanatory variable X and the explanatory variable Y to obtain the causal relationship value with causal direction of the explanatory variable X and the explanatory variable Y.

[0083] In this embodiment, after the steps of sequentially inputting multiple groups of time series data of different explanatory variables into the CCM model, using the CCM model to perform causal inference on the time series data of each group of different explanatory variables in two cross-mapping directions, and obtaining a causal relationship value with a causal direction between two explanatory variables in each group of time series data, the method for analyzing driving factors of cultivated land degradation further includes the following steps:

[0084] The causal relationship value with causal direction between two explanatory variables in the same set of time series data is compared with the Person correlation coefficient.

[0085] In this embodiment, step S4 uses the GCCM model to perform spatial causal reasoning based on the spatial data to obtain the spatial causal direction and causal relationship value between every two explanatory variables and between the explanatory variables and the cultivated land degradation data, specifically including the following steps:

[0086] Step S41, performing matrix conversion processing on the explanatory factor data and the cultivated land degradation data in the spatial data to obtain the spatial data after the matrix conversion processing.

[0087] Step S42: Set the size of the sliding window according to the length and width of the matrix.

[0088] Step S43: based on the spatial data after the matrix conversion processing and the size of the sliding window, set the spatial lag value, and perform a fixed-length screening on the spatial data in the matrix according to the spatial autocorrelation theory to obtain the screened spatial data in a matrix format.

[0089] Step S44: performing a linear trend elimination process on all the spatial data in matrix format, including removing the linear trend of cultivated land degradation, to obtain spatial data after the linear trend is eliminated.

[0090] Step S45: Based on the spatial data after removing the linear trend, a cross-mapping prediction method is used to predict any one of the explanatory variables according to the reconstructed manifold M from the explanatory variable. x The nearest neighbors identified in are used to predict the value of the explanatory variable y, and the true observed value of the explanatory variable and the corresponding predicted value are obtained.

[0091] Step S46: Calculate the Pearson correlation coefficient between the actual observed value of the explanatory variable and the corresponding predicted value according to the actual observed value of the explanatory variable and the corresponding predicted value.

[0092] Step S47, measuring the result of the cross-mapping prediction according to the Pearson correlation coefficient between the actual observed value of the explanatory variable and the corresponding predicted value.

[0093] In this embodiment, step S5 constructs a causal relationship graph according to the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value, taking the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value as known causal relationships, and specifically includes the following steps:

[0094] Step S51: compare the causal relationship values ​​between every two explanatory variables in the time series data with a first preset threshold value. When the causal relationship values ​​of the two causal directions are both smaller than the first preset threshold value, it is determined that there is no causal relationship between the two explanatory variables. When one or both of the causal relationship values ​​of the two causal directions are greater than the first preset threshold value, compare the two causal relationship values, and select the causal direction with the larger causal relationship value as a causal relationship for establishing a causal relationship diagram.

[0095] Step S52: compare the causal relationship values ​​between every two explanatory variables in the spatial data and between the explanatory variables and the cultivated land degradation data with the second preset threshold value. When the causal relationship values ​​in the two causal directions are both less than the second preset threshold value, it is determined that there is no causal relationship between the two explanatory variables. When one or both of the causal relationship values ​​in the two causal directions are greater than the second preset threshold value, compare the two causal relationship values, and select the causal direction with the larger causal relationship value as a causal relationship for establishing a causal relationship diagram, so as to obtain a causal relationship diagram between the explanatory variables and between the explanatory variables and the cultivated land degradation data.

[0096] In this embodiment, step S6 uses the causal relationship diagram in the SCM model to perform farmland degradation analysis on the target area to obtain farmland degradation analysis results, which specifically includes the following steps:

[0097] Step S61: All cultivated land degradation data and explanatory variable data are counted in the same EXCEL table to obtain EXCEL data.

[0098] Step S62, input the EXCEL data and the causal relationship diagram into the SCM model at the same time, use the SCM model to establish a causal relationship network, and output the causal effect value between the causal relationships. Among them, the causal relationship network established based on the SCM model contains the causal relationship between multiple explanatory variables. The causal effect value is a numerical value that quantifies the specific impact of a certain intervention on the outcome variable. In this embodiment, the causal effect value finally output refers to the effect value between each two explanatory variables with a causal relationship, which characterizes the closeness of the causal relationship between the two explanatory variables.

[0099] In order to make the implementation process of the technical solution of the present application clearer, this embodiment takes Hailun City in Northeast China as the target area, and takes temperature, precipitation and NDVI as examples to illustrate the specific implementation process of the technical solution of the present application in detail, which mainly includes the following contents.

[0100] (1) Collect and process the cultivated land degradation data and explanatory variable data representing the influencing factors in Hailun City, Northeast China, and form a dataset for analyzing the driving factors of cultivated land degradation. The explanatory variables include natural factor variables and human factor variables.

[0101] In this embodiment, the cultivated land degradation data collected is the soil organic matter raster data of Hailun City, Northeast Province between 1980 and 2021, and the raster data obtained by subtraction is used as the regional cultivated land degradation data. Among them, the explanatory variables include natural factor variables such as temperature, precipitation, NDVI, DEM, sand, silt, clay, and human factor variables such as GDP and population density. The data format is raster data. Among them, the temperature, precipitation, and NDVI annual raster data can be obtained from 1982 to 2021. The other raster data have little change, and only the single year of 2000 is considered.

[0102] (2) Time series data correspondence is unified. For natural factor variables, for example, variable data such as temperature, precipitation, and NDVI within a certain time range.

[0103] It should be noted that this embodiment only uses temperature, precipitation and NDVI as examples for illustration. In practice, other explanatory variables may also be used. The analysis methods for other natural factor variables or human factor variables are the same.

[0104] (3) The three time series data of temperature, precipitation and NDVI are input into the CCM model in pairs. The CCM model uses the time delay embedding on one time series to generate the attractor reconstruction, and then applies the simplex projection algorithm to estimate the concurrent value of the other time series, thereby obtaining the influence relationship between temperature→precipitation, precipitation→temperature, temperature→NDVI, NDVI→temperature, precipitation→NDVI, and NDVI→precipitation. The processing process of the CCM model specifically includes the following:

[0105] 1) Normalize the three time series data of temperature, precipitation and NDVI, and calculate the Person correlation coefficient between them.

[0106] 2) Time series data creates an index vector for the time series data, and forms a lagged coordinate vector based on the time series data as the reconstructed manifold M x , in the reconstructed manifold M x Based on the variable positioning, the lag coordinate vector is Y'(t)|Mx Denote , and find the nearest neighbor with time lag E+1.

[0107] 3) Calculate the nearest neighbor time index of the explanatory variable X from near to far, and use the nearest neighbor time index of the explanatory variable X to identify the nearest neighbor of the explanatory variable Y.

[0108]

[0109] In formula (1), dis[x(t),x(t z )] represents the Euclidean distance between two vectors, t represents a time point of the explanatory variable X, and γ z represents the z-th lagged nearest neighbor time index.

[0110]

[0111] Y'(t)|M x =Σw z Y(t z ), z=1,2,…,E+1 (3)

[0112] In formula (2) and formula (3), w z Represents x(t) and its reconstructed manifold M x The weight of the distance between the zth lag nearest neighbors, Y(t z ) is the value of the explanatory variable Y at the zth lag time point.

[0113] 4) Similarly, the lagged coordinate vector formed by the time series data of the explanatory variable Y is used as the reconstructed manifold M y , in the reconstructed manifold M y Based on the variable positioning, the lag coordinate vector X'(t)|M y Represents , and finds the E+1 nearest neighbors.

[0114] 5) Calculate the nearest neighbor time index of the explanatory variable Y from near to far, and use the nearest neighbor time index of the explanatory variable Y to identify the nearest neighbor of the explanatory variable X.

[0115] (4) According to the CCM model, reconstruct the manifold M x Cross-mapping the explanatory variables Y and reconstructing the manifold M yThe causal inference is performed in two directions of the cross-mapping explanatory variable X, and the causal relationship values ​​with causal directions between the three time series data are explored, and compared with the Person correlation coefficient. In this embodiment, the comparison with the Person correlation coefficient is to highlight that the causal inference results are more robust and more accurate than the correlation analysis. For the data processed by the CCM model and the GCCM model, all the data in the data used first are raster data, and some of the influencing factors are time series data in raster format. The influencing factors in this part of the time series data are first determined by the CCM model to determine the influence relationship between each other, and then for the time series data and non-time series data without influence relationship, the GCCM model is used to mine the influence relationship between each other in space.

[0116] Figure 2 Schematic diagram of the convergent cross-mapping results of temperature and precipitation; Figure 3 It is a schematic diagram of the convergent cross-mapping results of temperature and NDVI; Figure 4 Schematic diagram of the convergent cross-mapping results of precipitation and NDVI. Figure 2 to Figure 4 The relationship between the cross-mapping of each two explanatory variables and the Person correlation coefficient can be intuitively seen.

[0117] (5) Based on the research foundation of (3) and (4), the variables with strong causal relationship in both directions of the two explanatory variables are eliminated to ensure the accuracy of the final analysis process. Specifically, a causal relationship preset threshold can be set. By comparing the causal relationship value with the causal relationship preset threshold, when the causal relationship value in any direction is greater than the causal relationship preset threshold, it indicates that the two explanatory variables have a strong causal relationship in this direction. The child node is retained and the parent node is eliminated; when the causal relationship value is less than or equal to the causal relationship preset threshold, it indicates that the causal relationship in both directions of the two explanatory variables is weak, and it is not eliminated, thereby achieving the screening of explanatory variables. For two variables with strong temporal causal relationship, a causal direction is established, such as the relationship between x1→x2, and only x2 and other variables are placed in the GCCM model later.

[0118] (6) After all the explanatory variable data of natural factor variables such as temperature, NDVI, DEM, sand, silt, clay, and human factor variables such as GDP and population density are unified to the target area, the grid size and position are unified. Among them, unification to the target area means that for the above-mentioned grid data (including the explanatory variable data of natural factor variables and human factor variables and cultivated land degradation data), the pixel positions of the grid data are completely corresponding, the grid size and position are unified, and there is no deviation.

[0119] (7) For the processed raster data, causal inference is performed on the soil organic matter degradation cross-section data based on the GCCM model, including the following:

[0120] 1) Convert the raster data of explanatory variables and soil organic matter degradation into matrix format.

[0121] 2) According to the length and width of the matrix, the sliding window size is set. In this embodiment, the maximum range of the sliding window is smaller than the smaller value of the length and width.

[0122] 3) Set the preset value of spatial lag. In this embodiment, the spatial lag is set to 3, considering the spatial third-order lag (within three layers) grid data. Figure 5 As shown, Figure 5 In the raster data, the black grid represents the spatial center unit, the gray one grid represents the first-order spatial lag grid data, the gray two grid represents the second-order spatial lag grid data, and the gray three grid represents the third-order spatial lag grid data.

[0123] 4) According to the spatial autocorrelation theory, the raster data in the matrix is ​​screened regularly with a fixed length, that is, the raster data in the matrix is ​​screened according to a preset fixed length to meet the fixed length requirement.

[0124] 5) Remove the linear trend of the dependent variable of soil organic matter degradation.

[0125] 6) Using the nearest neighbor-based prediction is called cross-mapping prediction. For any explanatory variable, we can use the reconstructed manifold M x The nearest neighbor identified in predicts the value of y, the expression is:

[0126]

[0127] In formula (4), s represents the spatial unit, w si Y si |M x It represents the product of the weight of the i-th point of the s spatial unit and the y value of the corresponding position during the prediction process, L represents the order of spatial lag, and Y' s |M x Represents the reconstructed manifold M x The predicted result of y in . Figure 6 The corresponding manifold for the temperature spatial lag reconstruction is shown, Figure 7 The corresponding manifold for the spatially lagged reconstruction of the sand is shown.

[0128]

[0129]

[0130] In formula (5) and formula (6), w si |M x Represents the reconstructed manifold M of the i-th point of the s-space unit in the prediction processx The weight on represents the weight function between two states in the reconstructed manifold, represents the distance function between two states in the reconstructed manifold.

[0131]

[0132] In formula (7), The distance function between two states in the reconstructed manifold is defined as abs[h si(k) (x),h s(k) (x)] represents the distance function of two vectors, h si (x) is The observation function corresponding to the spatial unit in h s(k) (x) represents the observation function of the k-order lag unit in the state space s, h s (x) represents the observation function in the state space s.

[0133]

[0134] In formula (8), μ si(k,d) (x) represents the k-th spatial lag of the spatial unit si in the D direction, where D is the number of k-th spatial units.

[0135] 7) The Pearson correlation coefficient between the true observations and the corresponding predicted values ​​is used to measure the results of the cross-mapping predictions.

[0136]

[0137] In formula (9), Y represents the true observation value, represents the corresponding predicted value, express Covariance, Var(Y) represents the variance of Y, and ρ represents the Pearson correlation coefficient.

[0138] (8) In terms of both time and space, the causal inference of the CCM model and the GCCM model was used to obtain the relationship between the various explanatory variables and their impact on soil organic matter degradation. Figure 8 The causal relationship between temperature and soil organic matter degradation is shown. Fig. 9 The causal relationship between clay and soil organic matter degradation is shown. Then, a causal relationship diagram is constructed based on the magnitude and direction of the causal relationship values ​​between the explanatory variables. The resulting causal relationship diagram is shown in Fig.10 shown.

[0139] In this embodiment, when the bidirectional causal relationship values ​​between two explanatory variables are both less than 0.4, it is determined that there is no direct causal relationship between the two explanatory variables. When any one or both of the causal relationship values ​​in the two directions are greater than 0.4, it is determined that there is a direct causal relationship between the two explanatory variables, and the direction of the causal relationship with the larger causal relationship value is selected as the known causal relationship, thereby obtaining a causal relationship diagram.

[0140] (9) All explanatory variables and soil organic matter degradation grid data in 2000 are counted in the same EXCEL table, and input into the SCM model together with the causal relationship diagram to realize soil degradation analysis and obtain the analysis results. Among them, the SCM model inputs the statistical EXCEL data (explanatory variables and farmland degradation grid data) into the SCM model, and also inputs the causal relationship diagram into the SCM model, so that the SCM model can establish a causal relationship network based on the statistical EXCEL data and the causal relationship diagram, and output the causal effect value between the causal relationships, thereby realizing the farmland degradation analysis of the target area. The farmland degradation analysis results are mainly the causal effect values ​​between the causal relationships.

[0141] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0142] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0143] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0144] This embodiment mainly aims at the demand of mining the intricate relationship between driving factors in the analysis of multivariable driving factors in complex dynamic systems, and proposes a method that can realize directional causal inference between dependent variables and multiple driving factors and between multiple driving factors in complex nonlinear correlation systems from the perspectives of time and space. The time series data in raster format is obtained through the cultivated land degradation data and the explanatory variable data. The time causal reasoning is first performed through the CCM model to obtain the time-based causal direction and causal relationship value between each two of the explanatory variables, and then the spatial causal reasoning is performed through the GCCM model to obtain the space-based causal direction and causal relationship value between each two of the explanatory variables, and then the causal relationship diagram is constructed. The cultivated land degradation analysis is performed according to the causal relationship diagram, and accurate cultivated land degradation analysis results can be obtained. This application combines the CCM model and the GCCM model for the scenario of soil degradation analysis. First, the CCM model based on the simplex projection algorithm is used for time series variables to verify the unidirectional causal relationship between the explanatory variables. Then, the GCCM model is used for spatial causal reasoning to reveal the causal relationship between spatial cross-sectional variables, so that the corresponding causal effect can be estimated. Finally, based on the causal relationship and causal direction obtained by the CCM model and the GCCM model, a causal relationship diagram is provided for the SCM model to further realize the analysis of soil degradation, so as to truly reflect the correlation between the various explanatory variables that affect cultivated land degradation, effectively improve the accuracy of the analysis of the driving factors of cultivated land degradation, and solve the problem of inaccurate and unreliable analysis results of the driving factors of cultivated land degradation in the prior art.

[0145] This embodiment obtains the temporal lag E and spatial lag L through the simplex projection algorithm, and mines the temporal causal changes of cultivated land degradation based on the CCM model, and clarifies the causes of spatial changes based on the GCCM model, ensuring that the coupling relationship between the driving factors of cultivated land degradation is fully analyzed regardless of the lack of long time series data or time series. Compared with the existing spatiotemporal causal relationship model analysis method, the fusion method of convergent cross mapping (CCM) and causal inference of earth system cross-sectional data based on geographic convergent cross mapping (GCCM) can directly overcome the false correlation in linear correlation, consider the influence of time lag, and fully consider spatial correlation and spatial heterogeneity, and obtain the directional causal effect between the two variables. At the same time, it breaks through the SCM model in the original statistical field, realizes causal mining in complex dynamic systems, and can be applied to the causal relationship model measurement of nonlinear spatiotemporal data. In the case of weak correlation, weak and moderate causal relationships can still be detected; for strong correlation, the causal direction can be identified.

[0146] This embodiment proposes a method for causal analysis of driving factors in cultivated land quality degradation from the perspective of complex dynamic systems for temporal and spatial cross-sectional data. First, in the temporal dimension, the CCM model is used to determine the direction and causal effect value of each influencing factor, thereby filtering out data with large mutual influence. Then, the remaining data is used in the spatial dimension, and the GCCM model is used to determine the causal relationship and direction between each influencing factor and between the influencing factor and the dependent variable. The above results are used as known causal relationships to establish a causal relationship diagram, which is used as the basis of the SCM model for causal inference. The fusion method of convergent cross mapping (CCM) and causal inference of earth system cross-sectional data based on geographic convergent cross mapping (GCCM) is used to mine the causal relationship between the driving factors of cultivated land degradation as known causal relationships, construct a causal relationship diagram, provide a causal network for the SCM model, and realize the analysis of cultivated land degradation factors in complex dynamic systems. This method develops a method for causal relationship diagrams of driving factors of cultivated land degradation, combines it with the SCM model, and develops a software system to support the attribution of cultivated land degradation.

[0147] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0148] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0149] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for analyzing driving factors of cultivated land degradation, characterized in that: The method for analyzing the driving factors of cultivated land degradation includes: Acquire a data set of analysis of driving factors of cultivated land degradation in a target area; the data set of analysis of driving factors of cultivated land degradation includes cultivated land degradation data and corresponding explanatory variable data, the explanatory variable data is data of various explanatory variables affecting cultivated land degradation in the target area, and the explanatory variables include natural factor variables and human factor variables; rasterizing the cultivated land degradation driving factor analysis dataset to obtain raster data, wherein the raster data includes time series data and spatial data in a raster format; According to the time series data, a CCM model is used to perform temporal causal reasoning to obtain a time-based causal direction and a causal relationship value between each two explanatory variables; According to the spatial data, a GCCM model is used to perform spatial causal reasoning to obtain a spatially based causal direction and a causal relationship value between every two of the explanatory variables and between the explanatory variables and the cultivated land degradation data; Constructing a causal relationship graph according to the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value as known relationships; According to the cause-effect relationship diagram, the cause-effect relationship diagram is used in the SCM model to perform farmland degradation analysis on the target area to obtain a farmland degradation analysis result.

2. The method for analyzing driving factors of cultivated land degradation according to claim 1, characterized in that: The cultivated land degradation driving factor analysis dataset is subjected to rasterization processing to obtain raster data, which specifically includes: Select the cultivated land degradation data and explanatory variable data within the same time range from the cultivated land degradation driving factor analysis data set; The cultivated land degradation data and all explanatory variable data are uniformly converted into raster data with the same spatial resolution, and the pixel positions of the raster data are completely consistent.

3. The method for analyzing driving factors of cultivated land degradation according to claim 1, characterized in that: According to the time series data, the CCM model is used to perform temporal causal reasoning to obtain the time-based causal direction and causal relationship value between each two explanatory variables, specifically including: According to the types of the explanatory variables, the time series data are grouped in pairs to obtain multiple groups of time series data with different explanatory variables; Multiple groups of time series data of different explanatory variables are sequentially input into the CCM model, and the CCM model is used to perform causal inference on the time series data of each group of different explanatory variables in two cross-mapping directions to obtain the causal relationship value with causal direction between the two explanatory variables in each group of time series data; the CCM model is used to generate attractor reconstruction by time delay embedding on a time series, and the simplex projection algorithm is used to estimate the concurrent value of another time series to determine the causal relationship with direction between the two explanatory variables corresponding to each group of time series data.

4. The method for analyzing driving factors of cultivated land degradation according to claim 3, characterized in that: Multiple groups of time series data with different explanatory variables are sequentially input into the CCM model, and the CCM model is used to perform causal inference on the time series data of each group of different explanatory variables in two cross-mapping directions to obtain the causal relationship value with causal direction between the two explanatory variables in each group of time series data, specifically including: Each of the time series data is normalized respectively, and the Person correlation coefficient between every two of the time series data is calculated; Create an index vector for each of the time series data, and form a lagged coordinate vector based on the time series data as a reconstructed manifold M x , in the reconstructed manifold M x Based on the variable positioning, the lag coordinate vector is Y'(t)|M x Represents, and determines the time lag E+1 of the nearest neighbor; Calculate the nearest neighbor time index of each explanatory variable X, and use the nearest neighbor time index of the explanatory variable X to determine the nearest neighbor time index of the explanatory variable Y; The lagged coordinate vector formed by the time series data corresponding to the explanatory variable Y is used as the reconstructed manifold M y , in the reconstructed manifold M y Based on the variable positioning, the lag coordinate vector is used to realize the synchronous positioning of the variable. y Represents, and determines the time lag E+1 of the nearest neighbor; Calculate the nearest neighbor time index of the explanatory variable Y from near to far, and use the nearest neighbor time index of the explanatory variable Y to determine the nearest neighbor time index of the explanatory variable X; According to the nearest neighbor time index of the explanatory variable Y and the nearest neighbor time index of the explanatory variable X, the CCM model is used to reconstruct the manifold M. x Cross-mapping the explanatory variables Y and reconstructing the manifold M y In the two cross-mapping directions of the cross-mapping explanatory variable X, causal inference is performed on the time series data of the explanatory variable X and the explanatory variable Y to obtain the causal relationship value with causal direction of the explanatory variable X and the explanatory variable Y, and the causal relationship value is compared with the Person correlation coefficient.

5. The method for analyzing driving factors of cultivated land degradation according to claim 4, characterized in that: According to the spatial data, the GCCM model is used to perform spatial causal reasoning to obtain the spatial causal direction and causal relationship value between each two explanatory variables and between the explanatory variables and the cultivated land degradation data, specifically including: Performing matrix conversion processing on the explanatory factor data and the cultivated land degradation data in the spatial data to obtain spatial data after the matrix conversion processing; Set the size of the sliding window according to the length and width of the matrix; Based on the spatial data after the matrix conversion processing and the size of the sliding window, a spatial lag value is set, and according to the spatial autocorrelation theory, the spatial data in the matrix is ​​screened for a fixed length to obtain the spatial data in the screened matrix format; Eliminate the linear trend of all matrix-format spatial data, including the linear trend of cultivated land degradation, to obtain spatial data after the linear trend is removed; According to the spatial data after removing the linear trend, a cross-mapping prediction method is used to predict any one of the explanatory variables according to the reconstructed manifold M from the explanatory variable. x The nearest neighbors identified in are used to predict the value of the explanatory variable y, and the true observed value of the explanatory variable and the corresponding predicted value are obtained; According to the actual observed value of the explanatory variable and the corresponding predicted value, the Pearson correlation coefficient between the actual observed value of the explanatory variable and the corresponding predicted value is calculated; The results of the cross-mapping predictions are measured based on the Pearson correlation coefficient between the true observed values ​​of the explanatory variables and the corresponding predicted values.

6. The method for analyzing driving factors of cultivated land degradation according to claim 5, characterized in that: Based on the time-based causal direction and causal relationship value and the space-based causal direction and causal relationship value as known relationships, a causal relationship graph is constructed, specifically including: The mutual causal relationship values ​​between each two explanatory variables in the time series data are compared with the first preset threshold value. When the causal relationship values ​​of the two causal directions are both less than the first preset threshold value, it is determined that there is no causal relationship between the two explanatory variables; when one or both of the causal relationship values ​​of the two causal directions are greater than the first preset threshold value, the two causal relationship values ​​are compared, and the causal direction with the larger causal relationship value is selected as a causal relationship for establishing a causal relationship diagram; The causal relationship values ​​between each two explanatory variables in the spatial data and between the explanatory variables and the cultivated land degradation data are compared with the second preset threshold value. When the causal relationship values ​​of the two causal directions are both less than the second preset threshold value, it is determined that there is no causal relationship between the two explanatory variables; when one or both of the causal relationship values ​​of the two causal directions are greater than the second preset threshold value, the two causal relationship values ​​are compared, and the causal direction with the larger causal relationship value is selected as a causal relationship for establishing a causal relationship diagram, so as to obtain a causal relationship diagram between the explanatory variables and between the explanatory variables and the cultivated land degradation data; According to the cause-effect relationship diagram, the cause-effect relationship diagram is used in the SCM model to perform farmland degradation analysis on the target area, and the farmland degradation analysis results are obtained, which specifically include: All cultivated land degradation data and explanatory variable data are counted in the same EXCEL table to obtain EXCEL data; The EXCEL data and the causal relationship diagram are simultaneously input into the SCM model, a causal relationship network is established using the SCM model, and the causal effect values ​​between causal relationships are output.

7. The method for analyzing driving factors of cultivated land degradation according to claim 1, characterized in that: The natural factor variables include temperature, precipitation, NDVI, DEM, sand, silt and clay; the human factor variables include GDP, population density and transportation network.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for analyzing driving factors of farmland degradation as described in any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for analyzing the driving factors of cultivated land degradation described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for analyzing the driving factors of cultivated land degradation described in any one of claims 1 to 7 is implemented.

Images