Soil humidity spatial characteristic evaluation method and storage medium

Abstract

The invention provides a soil humidity spatial characteristic evaluation method and a storage medium. The method comprises the following steps: acquiring at least three pieces of detection data corresponding to at least three detection products and initial observation data corresponding to at least one initial observation station in a historical time period of a target region; wherein the spatial resolution of the detection product corresponds to the spatial resolution of the detection data; performing fusion processing on the at least three pieces of detection data to obtain initial fusion data; determining target fusion data according to the initial observation data and the initial fusion data; performing analysis processing on the target fusion data based on a preset spatial analysis method to obtain an analysis result of the soil humidity spatial characteristics; the analysis result is used for evaluating the attribution reason of the spatial characteristics of the target region.

Description

Technical Field

[0001] This application relates to the field of ecological environment monitoring technology, specifically to a method for assessing the spatial characteristics of soil moisture and a storage medium. Background Technology

[0002] Soil moisture (especially surface soil moisture, SSM) is a core parameter of eco-hydrological processes in cold regions, and its spatial heterogeneity directly affects regional water balance, energy exchange, and ecosystem stability. In data-scarce areas (e.g., high-altitude permafrost regions and remote mountainous areas), the complex topography, extreme climate, and limited number and spatial coverage of in-situ observation stations pose significant challenges to the accurate quantification of soil moisture spatial heterogeneity. Current analysis of soil moisture spatial heterogeneity mainly relies on single products or limited in-situ data, but this approach suffers from significant technical limitations: single products present a trade-off between spatial resolution and accuracy; high spatial resolution products are susceptible to freeze-thaw interference; medium and coarse spatial resolution products struggle to capture heterogeneity caused by micro-topography; and some products lack data on the freezing period. Summary of the Invention

[0003] This application provides a method and storage medium for assessing the spatial characteristics of soil moisture, aiming to solve the problem of limited accuracy of single products.

[0004] Firstly, this application provides a method for assessing the spatial characteristics of soil moisture, including: Acquire at least three sets of detection data corresponding to at least three detection products and initial observation data corresponding to at least one initial observation station within a historical time period of the target area; wherein, the spatial resolution of the detection products corresponds to the spatial resolution of the detection data; At least three of the aforementioned detection data are fused to obtain initial fused data; Based on the initial observation data and the initial fusion data, the target fusion data is determined; The target fusion data is analyzed and processed based on a preset spatial analysis method to obtain the analysis results of the spatial characteristics of soil moisture; the analysis results are used to evaluate the attribution causes of the spatial characteristics of the target area.

[0005] In some design approaches, the detection data is spatially distributed data with a time series; the spatially distributed data includes data corresponding to at least one location; the fusion processing of detection data corresponding to at least three of the detection products to obtain initial fused data includes: The detection data at different spatial resolutions are resampled to obtain a target detection dataset with a target spatial resolution; the target detection dataset includes a first subset, a second subset, and a third subset. The initial fused data is obtained by fusing the first subset, the second subset, and the third subset.

[0006] In some design embodiments, the detection data includes first data at a first spatial resolution, second data at a second spatial resolution, and third data at a third spatial resolution; the value of the first spatial resolution is less than the value of the second spatial resolution; the value of the second spatial resolution is less than the value of the third spatial resolution; the resampling process performed on the detection data at different spatial resolutions to obtain a target detection dataset at the target spatial resolution includes: The first data is upscaled and resampled based on a preset bilinear resampling method, so that the value of the first spatial resolution is increased to the value of the target spatial resolution, thus obtaining the first subset of data. The second data is downscaled and resampled according to a preset inverse distance weighting method, so that the value of the second spatial resolution is reduced to the value of the target spatial resolution, thus obtaining the second subset data set. The third data is downscaled and resampled according to the inverse distance weighting method, so that the value of the third spatial resolution is reduced to the value of the target spatial resolution, thus obtaining the third subset data.

[0007] In some design approaches, the initial fused data obtained by fusing the first subset of data, the second subset of data, and the third subset of data includes: Based on the preset triple registration method, the first subset, the second subset, and the third subset, a first variance set of the first subset, a second variance set of the second subset, and a third variance set of the third subset are determined; the first variance set includes at least one first variance corresponding to the position; the second variance set includes at least one second variance corresponding to the position; and the third variance set includes at least one third variance corresponding to the position. Based on the first variance set, the second variance set, and the third variance set, a first weight set for the first subset of data, a second weight set for the second subset of data, and a third weight set for the third subset of data are determined; the first weight set includes at least one first weight corresponding to the position; the second weight set includes at least one second weight corresponding to the position; and the third weight set includes at least one third weight corresponding to the position. The initial fused data is determined based on the first subset, the first weight set, the second subset, the second weight set, the third subset, and the third weight set.

[0008] In some of these design approaches, determining the target fusion data based on the initial observation data and the initial fusion data includes: Determine the first standard deviation of the initial observation data; Determine the second standard deviation of the initial fused data based on the initial observation data and the initial fused data; The target fusion data is determined based on the initial fusion data, the first standard deviation, and the second standard deviation.

[0009] In some of these design approaches, the spatial analysis method includes global analysis, local hotspot analysis, and local spatial correlation analysis; the analysis results include global spatial autocorrelation analysis results, local hotspot clustering analysis results, and spatial clustering results; the analysis and processing of the target fused data based on the preset spatial analysis method to obtain the analysis results of the spatial characteristics of soil moisture includes: The global spatial autocorrelation analysis results are determined based on the global analysis method and the target fusion data; the global spatial autocorrelation analysis results are used to characterize the overall spatial autocorrelation features of the soil moisture. Based on the local hotspot analysis method and the target fusion data, the local hotspot clustering analysis results are determined; the local hotspot clustering analysis results are used to characterize the spatial distribution characteristics of the local clustering of soil moisture. Based on the local spatial correlation analysis method and the target fusion data, the spatial clustering result is determined; the spatial clustering result is used to characterize the local spatial correlation features of soil moisture; the spatial clustering result includes at least one spatial correlation type.

[0010] In some of these design approaches, the method further includes: The target fusion data is split based on a preset freeze-thaw period to obtain freeze-thaw period data; the freeze-thaw period data includes freezing period data, thawing period data, wetting period data, and initial freezing period data; Based on the global analysis method, the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, determine the global spatial autocorrelation analysis results corresponding to each freeze-thaw period; Based on the local hotspot analysis method, the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, determine the local hotspot cluster analysis results corresponding to each freezing-thaw period; Based on the local spatial correlation analysis method, the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, the spatial clustering results corresponding to each freeze-thaw period are determined.

[0011] In some of these design approaches, the method further includes: Obtain at least one driving factor corresponding to the spatial characteristics of soil moisture; Based on the fusion data of each driving factor and the target, an explanatory power parameter for each driving factor is determined; the explanatory power parameter is used to characterize the explanatory power of the driving factor for the formation and evolution of the spatial characteristics of soil moisture. Based on the magnitude of the explanatory power parameter, target driving factors are determined from the at least one driving factor for the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, respectively; the target driving factor is a factor that influences the formation and evolution of the spatial characteristics of soil moisture.

[0012] In some of these design approaches, the method further includes: Based on the initial observation data and the target fusion data, the verification parameters of the target fusion data are determined; If the verification parameters do not meet the preset verification conditions, the target fusion data is re-determined based on the initial observation data and the initial fusion data.

[0013] Secondly, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, is used to implement the method described above.

[0014] In the embodiments of this application, by acquiring detection data corresponding to at least three detection products and initial observation data corresponding to at least one initial observation station, the detection data is fused to obtain initial fused data. Then, the target fused data is optimized and determined by combining the initial observation data. Finally, the target fused data is analyzed based on a preset spatial analysis method to obtain results that can be used to evaluate the spatial characteristics of the target area. This can achieve an effective evaluation of the accurate analysis of the spatial characteristics of soil moisture in the target area, thereby improving the technical problem in related technologies where the accuracy of single data is limited, thus making it impossible to accurately evaluate the spatial characteristics of the target area. Attached Figure Description

[0015] Figure 1 A schematic flowchart of a method for assessing the spatial characteristics of soil moisture provided in an embodiment of this application; Figure 2 A time-series comparison chart of detection data from multi-source soil moisture products and target fusion data; Figure 3 A schematic diagram of the spatial distribution of detection data from multi-source soil moisture products and target fusion data during the freeze-thaw cycle; Figure 4 A schematic diagram illustrating the time-series variation of the global Moran's I index during different freeze-thaw periods; Figure 5 A schematic diagram showing the spatial distribution of hotspots / colds during different freeze-thaw periods; Figure 6 A schematic diagram of LISA clustering for different freeze-thaw periods; Figure 7 A schematic diagram comparing the q values ​​of driving factors under different freeze-thaw periods; Figure 8 This is a schematic diagram of a hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] In-situ data is scarce in data-deficient areas, making it difficult to effectively verify and calibrate products, resulting in insufficient reliability of heterogeneity analysis results. Related analysis methods often employ single spatial statistical indicators, lacking a multi-level system of "global-local-clustering," making it difficult to comprehensively characterize the complex structure of heterogeneity. The identification of driving factors does not fully consider the phased impact of freeze-thaw cycles, failing to reveal the dynamic driving mechanism of heterogeneity. Furthermore, related methods are not adapted to the characteristics of data sources in data-deficient areas; multi-source data fusion is merely a simple weighting, failing to fully consider the error characteristics of different products, resulting in limited accuracy of the combined product. Therefore, current analysis of soil moisture spatial heterogeneity in data-deficient areas suffers from poor data adaptability, incomplete analytical levels, and unclear driving mechanisms, making it difficult to meet the refined requirements of cold-region eco-hydrological simulation.

[0018] like Figure 1 As shown, Figure 1 This is a flowchart illustrating a method for assessing the spatial characteristics of soil moisture provided in an embodiment of this application. This embodiment provides a method for assessing the spatial characteristics of soil moisture, including: Step 101: Obtain at least three sets of detection data corresponding to at least three detection products and initial observation data corresponding to at least one initial observation station within the historical time period of the target area; wherein, the spatial resolution of the detection products corresponds to the spatial resolution of the detection data.

[0019] The detection product can be a remote sensing product, and the detection data can be remote sensing data. The spatial characteristics of soil moisture can be soil moisture spatial heterogeneity; the assessment method for soil moisture spatial characteristics can be a multi-level analysis method for soil moisture spatial heterogeneity in data-scarce areas, used to address the technical problems of inaccurate heterogeneity quantification caused by insufficient in-situ data, limited accuracy of single products, and unsystematic analysis methods in data-scarce areas. The target area can be the target data-scarce area. The detection product can be a soil moisture product.

[0020] As an example, the detection products may include three products with the same spatial resolution. As another example, the detection products may include a first product with a first spatial resolution, such as a high spatial resolution (500m) SMAP-based product; a second product with a second spatial resolution, such as a medium spatial resolution (9km) ERA5-Land product; and a third product with a third spatial resolution, such as a coarse spatial resolution (25km) ESA CCI product. It should be noted that the three detection products in this application are independent of each other in terms of data sources (e.g., different satellite observation platforms) and soil moisture inversion models, and there are no data or algorithm-level dependencies.

[0021] The initial observation station can be an in-situ observation station. The probe data can be soil moisture data carried by the probe product. The probe data can be raster-type spatial data, and the basic unit of the probe data can be a grid cell. The initial observation data can be in-situ observation data, including measured in-situ soil moisture data and in-situ soil temperature data.

[0022] Areas lacking data are defined as those with fewer than or equal to a preset number of in-situ observation stations, spatial coverage less than or equal to a preset percentage, complex terrain, or extreme climate, including high-altitude permafrost areas and remote mountainous regions. For example, the preset number is 5, the preset percentage is 30%, and spatial coverage can be the proportion of the actual spatial area covered by the target data within the study area to the total area of ​​the region.

[0023] It can collect in-situ soil moisture data (0-100cm depth, with a focus on extracting 0-5cm surface layer data) and measured in-situ soil temperature data from four or more locations within the target area. The observation frequency is 10 minutes per observation, and the data is processed into a standardized dataset after daily averaging. Quality control is performed on the in-situ observation data. Specifically, sensor calibration (dielectric constant correction based on soil organic matter content) and physical range verification (outliers exceeding 0-0.6 m³ / m³) are performed on the in-situ soil moisture data, and accuracy verification (screening within ±0.1℃ error range) is performed on the in-situ soil temperature data, generating a quality-controlled in-situ dataset.

[0024] Step 102: Perform fusion processing on at least three detection data to obtain initial fused data.

[0025] In some embodiments, the detection data is spatially distributed data with a time series; the spatially distributed data includes data corresponding to at least one location; the detection data corresponding to at least three detection products are fused to obtain initial fused data, including: The detection data at different spatial resolutions are resampled to obtain the target detection dataset with the target spatial resolution; the target detection dataset includes a first subset, a second subset, and a third subset. The initial fused data is obtained by fusing the first, second, and third subsets of the dataset.

[0026] The target detection dataset can be a base dataset with a uniform scale. Resampling processing involves resampling at a uniform spatial scale. The first subset can be derived from the first product; the second subset can be derived from the second product; and the third subset can be derived from the third product. Fusion processing involves merging the subsets using a weighted average method.

[0027] It should be noted that the detection data and / or target detection dataset can be a collection of data containing time-series information and exhibiting spatial distribution characteristics, capable of simultaneously reflecting the spatiotemporal characteristics of the data. At least one location can be at least one spatial location, and the spatial distribution data can be a data format that uses spatial locations as carriers to represent the detection information corresponding to different spatial locations. As an example, the spatial distribution data can be a spatial distribution map or a spatial distribution table. The data corresponding to each spatial location can be the data obtained at that spatial location at the corresponding time node through detection methods.

[0028] Specifically, the probe data and / or target detection dataset is not data from a single time point or a single spatial location, but rather a comprehensive dataset integrating multiple time points and multiple spatial locations. It includes information on the dynamic changes of data over time, as well as the distribution differences of data at different spatial locations at the same time point. The spatial distribution data is indexed by discrete or continuous spatial locations, with the data corresponding to each spatial location being independent yet interconnected. As an example, the time series of the probe data and / or target detection dataset can cover multiple observation periods (e.g., annual, quarterly, monthly, or daily), and the spatial distribution data can correspond to multiple grid cells (i.e., spatial locations) within the target area. The probe data corresponding to each grid cell can be the soil moisture detection value for that grid cell at each time point.

[0029] In some embodiments, the method further includes: determining whether the spatial resolution of the probe data matches the target spatial resolution; if the spatial resolution of the probe data does not match the target spatial resolution, then resampling the probe data at different spatial resolutions to obtain a target probe dataset with the target spatial resolution. Here, matching the spatial resolution of the probe data with the target spatial resolution can mean that the spatial resolution of the probe data is the same as or identical to the target spatial resolution.

[0030] In some embodiments, the detection data includes first data at a first spatial resolution, second data at a second spatial resolution, and third data at a third spatial resolution; the value of the first spatial resolution is less than the value of the second spatial resolution; the value of the second spatial resolution is less than the value of the third spatial resolution; the detection data at different spatial resolutions are resampled to obtain a target detection dataset at the target spatial resolution, including: The first data is upscaled using a pre-defined bilinear resampling method, which increases the value of the first spatial resolution to the value of the target spatial resolution, thus obtaining the first subset of data. The second data is downscaled and resampled according to the preset inverse distance weighting method, so that the value of the second spatial resolution is reduced to the value of the target spatial resolution, thus obtaining the second subset data. The third data is downscaled and resampled using the inverse distance weighting method, reducing the value of the third spatial resolution to the value of the target spatial resolution, thus obtaining the third subset data.

[0031] In this system, the first product corresponds to the first data, the second product corresponds to the second data, and the third product corresponds to the third data. The target spatial resolution can be a pre-defined spatial resolution standard used to unify the spatial resolution of each detection data, with a value between the first and third spatial resolutions. The selection of the resampling method mainly depends on the numerical relationship between the target spatial resolution and the original spatial resolution of each product. Among them, the bilinear resampling method can be used to process high spatial resolution products; the inverse distance weighting (IDW) method can be used to process medium and coarse spatial resolution products.

[0032] In some embodiments, the method further includes: comparing the value of a first spatial resolution with the value of a target spatial resolution; if the value of the first spatial resolution is less than the value of the target spatial resolution, then performing upscaling resampling on the first data based on a preset bilinear resampling method; comparing the value of a second spatial resolution with the value of the target spatial resolution; if the value of the second spatial resolution is greater than the value of the target spatial resolution, then performing downscaling resampling on the second data based on a preset inverse distance weighting method; and comparing the value of a third spatial resolution with the value of the target spatial resolution; if the value of the third spatial resolution is greater than the value of the target spatial resolution, then performing downscaling resampling on the third data based on the inverse distance weighting method.

[0033] In the embodiments of this application, the advantages of different resampling methods in adapting to various spatial resolution data can be fully utilized to achieve accurate and efficient conversion of detection data of different spatial resolutions to the target spatial resolution.

[0034] In some embodiments, initial fused data is obtained by fusing the first subset, the second subset, and the third subset, including: Based on the pre-defined triple registration (TC) method, the first subset, the second subset, and the third subset, determine the first variance set of the first subset, the second variance set of the second subset, and the third variance set of the third subset; the first variance set includes the first variance corresponding to at least one position; the second variance set includes the second variance corresponding to at least one position; and the third variance set includes the third variance corresponding to at least one position. Based on the first variance set, the second variance set, and the third difference set, determine the first weight set of the first subset, the second weight set of the second subset, and the third weight set of the third subset; The initial fused data is determined based on the first subset, the first weight set, the second subset, the second weight set, the third subset, and the third weight set.

[0035] The triple registration method is only applicable to error variance estimation for three independent data sources. When the number of detector products involved in the fusion is greater than three, all detector products can be split into multiple combinations containing three detector products. For each combination, the error variance of each product is calculated using the TC method. For a single detector product, the mean of the error variances calculated by the TC method in all involved combinations is taken as the final error variance of that detector product.

[0036] In this embodiment, the soil moisture time series of the first, second, and third subsets after resampling to the target spatial resolution are denoted as X1, X2, and X3, respectively. Each time series includes a correspondence between at least one time parameter and at least one spatial distribution data. For example, the at least one time parameter can be 365 days, and each time series can include 365 spatial distribution data corresponding to 365 days. The spatial distribution data corresponding to each time parameter (e.g., 1 day) includes data corresponding to at least one location. The first variance set σ1², the second variance set σ2², and the third variance set σ3² are calculated based on the triple registration (TC) method. Each variance set includes variance corresponding to at least one location. Each variance set can be the spatial distribution of variance for each product, and the positions in each variance set correspond one-to-one with the positions in the spatial distribution data. The first variance set, the second variance set, and the third variance set can be the error variance of each subset. The error variance set can be derived from the covariance matrix between subsets, achieving error quantification without ground truth data. Based on the inverse relationship between error variance and weights, the following weight sets are calculated: ω1 = 1 / σ1² ÷ (1 / σ1² + 1 / σ2² + 1 / σ3²), ω2 = 1 / σ2² ÷ (1 / σ1² + 1 / σ2² + 1 / σ3²), and ω3 = 1 / σ3² ÷ (1 / σ1² + 1 / σ2² + 1 / σ3²). Each weight set includes weights corresponding to at least one location. Each weight set can represent the spatial distribution of weights for each product, and the locations within each weight set correspond one-to-one with the locations in the spatial distribution data. Based on the first subset, first weight set, second subset, second weight set, third subset, and third weight set, the initial fused data S0 = ω1X1 + ω2X2 + ω3X3 is generated using a weighted average method.

[0037] It should be noted that the above TC method calculation requires point-by-point operation on each grid cell of the three products after resampling to the target spatial resolution. Finally, a variance spatial distribution (i.e., variance map) consistent with the target spatial resolution can be obtained. At the same time, the weight spatial distribution (i.e., weight map) corresponding to each product can be further calculated based on the variance of each grid.

[0038] In the embodiments of this application, the accuracy advantages of each subset of data can be fully utilized, while the error defects of a single subset of data can be avoided.

[0039] Step 103: Determine the target fusion data based on the initial observation data and the initial fusion data.

[0040] In some embodiments, determining the target fused data based on initial observation data and initial fused data includes: Determine the first standard deviation of the initial observation data; The second standard deviation of the initial fused data is determined based on the initial observation data and the initial fused data. The target fused data is determined based on the initial fused data, the first standard deviation, and the second standard deviation.

[0041] The first standard deviation can be the natural standard deviation of the in-situ soil moisture data, denoted as σ. insitu The second standard deviation can be the empirical standard deviation σ of the residual sequence between the initial fused data and the in-situ soil moisture data. esidual The target fusion data can be a highly applicable combination of soil moisture data.

[0042] In this embodiment, statistical analysis can be performed on the in-situ soil moisture data to calculate the first standard deviation of the in-situ soil moisture data. The initial fused data can be matched with each in-situ soil moisture data point in the in-situ dataset to obtain at least two data points. The difference between each data point is calculated to form a residual sequence, and the second standard deviation is calculated based on this residual sequence. The calculation process for the first and second standard deviations can be determined according to actual conditions and is not limited here. Based on the initial fused data, the first standard deviation, and the second standard deviation, the target fused data S = S0 × (σ) is determined using the variance correction method. insitu / σ esidual ).

[0043] In the embodiments of this application, the true fluctuation characteristics of the initial observation data are determined by the first standard deviation, and the deviation and dispersion of the initial observation data and the initial fused data are quantified based on the second standard deviation, so as to achieve accurate correction and accuracy improvement of the initial fused data.

[0044] In some embodiments, the detection data of the unfused multi-source soil moisture products can be compared with the target fused data to obtain a time series comparison chart of the detection data of the multi-source soil moisture products and the target fused data.

[0045] Step 104: Analyze and process the target fusion data based on the preset spatial analysis method to obtain the analysis results of the spatial characteristics of soil moisture; the analysis results are used to evaluate the attribution causes of the spatial characteristics of the target area.

[0046] In some embodiments, the spatial analysis method includes a global analysis method, a local hotspot analysis method, and a local spatial correlation analysis method; the analysis results include global spatial autocorrelation analysis results, local hotspot clustering analysis results, and spatial clustering results; based on the preset spatial analysis method, the target fused data is analyzed and processed to obtain the analysis results of the spatial characteristics of soil moisture, including: The global spatial autocorrelation analysis results were determined based on the global analysis method and target fusion data; the global spatial autocorrelation analysis results were used to characterize the overall spatial autocorrelation features of soil moisture. Based on the local hotspot analysis method and target fusion data, the results of local hotspot cluster analysis are determined; the results of local hotspot cluster analysis are used to characterize the spatial distribution characteristics of local soil moisture aggregation. Based on local spatial correlation analysis and target fusion data, spatial clustering results are determined; spatial clustering results are used to characterize the local spatial correlation features of soil moisture; spatial clustering results include at least one spatial correlation type.

[0047] Among these methods, a three-tiered analysis system of "global-local-clustering" can be used to quantify the heterogeneity of the target fused data. The global analysis method can be the global Moran's I index analysis method; the local hotspot analysis method can be the Getis-OrdGi hotspot analysis method; and the local spatial association analysis method can be the Local Indicators of Spatial Association (LISA) analysis method.

[0048] In this embodiment, the global Moran's I index analysis method is adopted. The proximity relationships of each grid cell in the target fusion data are calculated using Euclidean distance. An inverse distance spatial weight matrix is ​​set based on these proximity relationships. Row normalization eliminates the scale dependence of each grid cell. The Moran's I significance level is set to p < 0.05, and the Moran's I value ranges from [-1, 1], reflecting the degree of spatial clustering or dispersion. Moran's I > 0 indicates spatial clustering, Moran's I < 0 indicates spatial dispersion, and Moran's I ≈ 0 indicates random distribution. The global spatial autocorrelation analysis results are presented as a time-series variation table of the global Moran's I index.

[0049] In this embodiment, the Getis-OrdGi hotspot analysis method is used. A fixed distance threshold (1000m) is set to ensure that each grid cell in the target fused data has at least one neighboring cell. The target fused data is standardized to obtain the z-value. Hotspots (high-value clustering) and cold spots (low-value clustering) are determined by the z-value. The significance level of the z-value is set to 95% to identify the local extreme value distribution patterns of heterogeneity. The local hotspot clustering analysis results are a hotspot / cold spot spatial distribution map.

[0050] In this embodiment, the LISA analysis method is used to classify spatial association types into high-high clustering (high values ​​are surrounded by high values), low-low clustering (low values ​​are surrounded by low values), high-low outlier clustering (high values ​​are surrounded by low values), and low-high outlier clustering (low values ​​are surrounded by high values). False discovery rate (FDR) correction is used to control multiple test errors and identify spatial clusters and outliers. The spatial clustering results are presented as LISA cluster diagrams.

[0051] In the embodiments of this application, a comprehensive characterization of the spatial characteristics of soil moisture can be achieved, from the whole to the part, and from macroscopic correlation to microscopic clustering.

[0052] In some embodiments, the method further includes: The target fusion data is split based on the preset freeze-thaw period to obtain freeze-thaw period data; the freeze-thaw period data includes freezing period data, thawing period data, wetting period data and initial freezing period data; Based on the global analysis method, freezing period data, thawing period data, wetting period data, and initial freezing period data, the global spatial autocorrelation analysis results corresponding to each freeze-thaw period are determined. Based on the local hotspot analysis method, frozen period data, thawing period data, moist period data, and initial freezing period data, the local hotspot cluster analysis results corresponding to each freeze-thaw period are determined. Based on local spatial correlation analysis, freezing period data, thawing period data, wetting period data, and initial freezing period data, the spatial clustering results corresponding to each freeze-thaw period are determined.

[0053] The preset freeze-thaw period can be a time period pre-divided according to the natural laws of soil freezing and thawing; for example, the freeze-thaw period can be called a freeze-thaw stage. The historical time period is divided into four typical stages of the freeze-thaw period: freezing period (January-March), thawing period (April-June), moist period (July-September), and initial freezing period (October-December), corresponding to different states of the soil freeze-thaw cycle. The global spatial autocorrelation analysis results for each freeze-thaw period can be obtained by substituting the freeze-thaw period data of each stage into the global analysis method, representing the overall spatial correlation characteristics of soil moisture in the corresponding stage. The local hotspot clustering analysis results for each freeze-thaw period can be obtained by substituting the freeze-thaw period data of each stage into the local hotspot analysis method, representing the local spatial distribution characteristics of soil moisture aggregation in the corresponding stage. The spatial clustering results for each freeze-thaw period can be obtained by substituting the freeze-thaw period data of each stage into the local spatial correlation analysis method, representing the local spatial correlation characteristics of soil moisture in the corresponding stage.

[0054] In the embodiments of this application, the differences and evolution patterns of soil moisture spatial characteristics at different freeze-thaw stages can be systematically revealed, making up for the shortcomings of overall analysis in reflecting stage characteristics.

[0055] In some embodiments, the detection data of unfused multi-source soil moisture products and the target fused data can be compared based on the freezing period data, thawing period data, wetting period data and early freezing period data to obtain a spatial distribution diagram of the detection data of multi-source soil moisture products and the target fused data in different freeze-thaw stages.

[0056] In some embodiments, the method further includes: Obtain at least one driving factor corresponding to the spatial characteristics of soil moisture; Based on the fusion data of each driving factor and target, the explanatory power parameter of each driving factor is determined; the explanatory power parameter is used to characterize the explanatory power of the driving factor on the formation and evolution of the spatial characteristics of soil moisture. Based on the magnitude of the explanatory power parameter, the target driving factor corresponding to the freezing period data, thawing period data, wetting period data and initial freezing period data are determined in at least one driving factor; the target driving factor is the factor that influences the formation and evolution of soil moisture spatial characteristics.

[0057] The driving factors can be auxiliary driving factors, including precipitation (PRE), normalized difference vegetation index (NDVI), temperature (T / TEM), land use / cover (LUCC), soil type (ST), digital elevation model (DEM), and aspect (AS) data. PRE and T are downscaled data with a 1km spatial resolution delta, NDVI is monthly composite data from a 250m spatial resolution vegetation index product (e.g., MOD13Q1), and the DEM has a 30m spatial resolution.

[0058] All auxiliary driving factor data can be preprocessed, including data format standardization, projection transformation, and missing value imputation, to generate a standardized driving factor dataset. Projection standardization (WGS84 coordinate system) and missing value imputation (linear interpolation for continuous factors and nearest neighbor interpolation for categorical factors) are performed on all auxiliary driving factors. LUCC is divided into 6 primary types, and ST adopts a 1:1,000,000 soil type classification system. Slope aspect data is extracted from the DEM using software (e.g., ArcGIS). All preprocessed driving factor data needs to be further resampled to the target spatial resolution.

[0059] In this embodiment, the Jenks natural breakpoint method can be used to divide continuous driving factors such as precipitation, temperature, NDVI, and DEM into 9 categories. Each category includes at least one target fusion data to ensure that the variance within the category is minimized and the variance between categories is maximized, thus meeting the analysis requirements of the geographic detector. Categorical factors such as LUCC and ST directly retain their original classification categories.

[0060] Explanatory power parameters can be statistical indicators that quantify the influence of driving factors on the spatial characteristics of soil moisture. A larger value indicates stronger explanatory power for the driving factor. For example, the q-value in the geographic detector method can range from [0,1]. Based on the geographic detector formula, the q-values ​​of each driving factor at different stages are calculated. A larger q-value indicates stronger explanatory power for heterogeneity. The formula is as follows: Where L is the number of driving factor categories, and N is... h Let σ be the number of samples in the target fusion data of class h. h ² represents the variance of the target fusion data for class h, N represents the total number of samples in the target fusion data, and σ² represents the overall variance of the target fusion data.

[0061] The target driving factor can be the one with the largest explanatory power parameter and the most significant effect on the formation and evolution of soil moisture spatial characteristics during a specific freeze-thaw stage. By longitudinally comparing the q-values ​​of different driving factors within the same freeze-thaw period (freezing period, thawing period, wetting period, and freezing period), the driving factor with the largest q-value in each stage is selected as the target driving factor. For example, the q-value of the temperature factor increases by 62-104% during the freezing period, and the q-value of the precipitation factor during the thawing period is 0.343-0.481. Furthermore, by horizontally comparing the changes in the q-value of the same driving factor across different freeze-thaw periods (freezing period, thawing period, wetting period, and freezing period), and combining this with the differences in core factors during each freeze-thaw period, the reshaping effect of the freeze-thaw cycle on the driving mechanism can be revealed.

[0062] In the embodiments of this application, the factors that play a dominant role in the spatial characteristics of soil moisture at different freeze-thaw stages can be accurately identified, and the stage-specific differences in the driving factors can be clarified.

[0063] In some embodiments, the method further includes: Based on the initial observation data and the target fusion data, determine the validation parameters for the target fusion data; If the verification parameters do not meet the preset verification conditions, the target fusion data is re-determined based on the initial observation data and the initial fusion data.

[0064] The validation parameters can be statistical indicators used to measure the degree of fit between the target fused data and the actual soil moisture conditions, including the correlation coefficient (R) and the unbiased root mean square error (ubRMSE). The preset validation conditions can be that the correlation coefficient R is greater than or equal to 0.75 and the unbiased root mean square error (ubRMSE) is less than or equal to 0.055 m³ / m³.

[0065] In this embodiment, based on the initial observation data and the target fusion data, the verification parameters of the target fusion data are determined through statistical calculations. For example, the correlation coefficient between the two is calculated to measure the degree of linear correlation, and the unbiased root mean square error is calculated to measure the magnitude of the deviation. The obtained verification parameters are compared with the preset verification conditions to determine whether the verification parameters have reached the preset threshold. If the verification parameters do not meet the preset verification conditions, it means that the accuracy of the target fusion data is not up to standard. It is necessary to re-correct and optimize the initial fusion data by combining it with the initial observation data, adjusting the relevant calculation parameters of variance correction or the fusion weight allocation method, and re-determining the target fusion data until the verification parameters meet the preset verification conditions.

[0066] In some embodiments, a spatial heterogeneity analysis report of soil moisture in data-scarce areas can be generated. This report includes a global Moran's I index time-series variation map, a hotspot / colds spatial distribution map, a LISA clustering map, a radar map of driving factor q values, and a map of heterogeneity evolution patterns at different freeze-thaw stages, providing data support and methodological reference for eco-hydrological simulation.

[0067] In the embodiments of this application, the accuracy verification and iterative optimization of the target fusion data can be realized to ensure that the target fusion data can accurately reflect the real soil moisture conditions.

[0068] The following describes the method for assessing the spatial characteristics of soil moisture provided in this application. This embodiment uses the Yangtze River source region (a typical high-altitude permafrost region with limited data, consisting of four in-situ observation stations) as the study area, and the study period is 20xx-20xx. The implementation process of the method is described in detail below: Step 1: Data Acquisition.

[0069] Multi-source soil moisture products: SMAP-based products (500m, 20xx-20xx) are from the Tibetan Plateau Multi-Source Fusion Dataset Platform, ERA5-Land products (9km, 20xx-20xx) are from the Copernicus Climate Data Storage Center, and ESACCI products (25km, 20xx-20xx) are from the European Space Agency Climate Change Programme website.

[0070] In-situ observation data: In-situ soil moisture and temperature data from 20xx to 20xx were obtained from four in-situ observation stations in the Yangtze River source area. The sensors used were CS650 (soil moisture, accuracy ±2%) and 109 series probes (soil temperature, accuracy ±0.1℃). Daily average soil moisture data in the 0-5cm surface layer were extracted.

[0071] Auxiliary driving factor data: PRE and T data (1km spatial resolution) are from the National Tibetan Plateau Scientific Data Center, NDVI data (250m, 16-day composite) are from the MOD13Q1 product, LUCC and ST data are from the Resource and Environmental Science Data Center of the Chinese Academy of Sciences, DEM (30m) is from ASTERGDEM data, and AS data is extracted from DEM using ArcGIS 10.5.

[0072] Step 2: Data preprocessing.

[0073] Resampling: SMAP-based products were bilinearly resampled to 1km, and ERA5-Land and ESA CCI products were IDW resampled to 1km, with a unified projection of WGS84 / UTMZone47N.

[0074] In-situ observation data standardization: Dielectric constant correction was performed on soil moisture data (based on soil organic matter content at each station), and outliers greater than 0.6 or less than 0 were removed; Soil temperature data underwent precision screening, retaining data with errors within ±0.1℃, ultimately forming a daily-scale dataset for 20xx-20xx from four in-situ observation stations.

[0075] Driving factor preprocessing: All driving factor data were unified to a spatial resolution of 1km, and a small number of missing values ​​were filled by linear interpolation; LUCC was divided into 6 categories: cultivated land, forest land, grassland, water body, construction land, and unused land; ST adopted the 1:1,000,000 Chinese soil classification system, with a total of 12 main types.

[0076] Step 3: Generate combined soil moisture products.

[0077] Error variance estimation: Based on the daily time series of three resampled products from 20xx to 20xx, the triple registration (TC) method was implemented in Python to calculate the error variance distribution of SMAP-based, ERA5-Land and ESA CCI products.

[0078] Weighted fusion: Calculate the weight distribution and use the weighted average method to generate the initial fused data S0.

[0079] In-situ observation data correction: Calculate the residual sequence between the initial fused data and the in-situ observation data of the four in-situ observation stations, and generate the target fused data S using the variance correction formula.

[0080] like Figure 2 As shown, Figure 2This is a time series comparison chart of detection data from multiple soil moisture products and target fusion data. It shows the temporal variation trend of soil moisture from three products: SMAP-based, ERA5-Land, and ESA CCI, as well as in-situ observation data and combined detection data. Specifically, it includes (a) the watershed average, (b) the grid of station 1, (c) the grid of station 2, (d) the grid of station 3, and (e) the grid of station 4. The vertical axis represents the soil moisture value (0-0.6), and the horizontal axis represents time. The soil moisture time series of the three single products and the combination are compared to verify the stability and accuracy of the combined product in terms of time series. The verification results are then labeled (R=0.82, ubRMSE=0.049m³ / m³).

[0081] like Figure 3 As shown, Figure 3 This is a spatial distribution diagram of the detection data and target fusion data of multi-source soil moisture products during the freeze-thaw stages. It shows the spatial distribution of soil moisture by detection data, in-situ observation data and combined soil moisture data of three products, namely SMAP-based, ERA5-Land and ESA CCI, according to the time period, freezing period (January-March) and thawing period (April-June). The color depth represents the moisture level (0-0.7m³ / m³) to reflect the spatial differences in moisture at different stages.

[0082] Step 4: Multilevel spatial heterogeneity analysis.

[0083] like Figure 4 As shown, Figure 4 This diagram illustrates the temporal variation of Moran's I index across different freeze-thaw periods. It shows the Moran's index, expected index, variance, z-value, and p-value. The expected index is the theoretically random distribution of Moran's I (close to 0); the variance is the statistical variance of Moran's I, used to calculate the z-value. The results show that Moran's I = 0.87 for the combined product year, I = 0.62 for the frozen period (January-March), and I = 0.91 for the wet period (July-September), indicating a strong clustering distribution of soil moisture overall, with more significant heterogeneity during the wet period. Figure 5 As shown, Figure 5 This diagram illustrates the spatial distribution of hotspots and cold spots during different freeze-thaw periods. The results show that hotspot areas are mainly concentrated in the southeastern river valleys (particularly noticeable during the thawing period), while cold spots are concentrated in the high-altitude permafrost areas of the northwest, consistent with the topography and freeze-thaw conditions. Figure 6 As shown, Figure 6This is a schematic diagram of LISA clustering at different freeze-thaw periods. The results show that high-high clusters account for 23.5% (southern valleys), low-low clusters account for 31.2% (northwestern high-altitude areas), and high-low outliers account for 4.8% (transition zone between valleys and plateaus), clearly revealing the fine structure of heterogeneity.

[0084] Step 5: Identification of driving factors and analysis of mechanisms.

[0085] Freeze-thaw phase division: The years 20xx-20xx are divided into four phases: January-March (freezing period), April-June (thawing period), July-September (wetting period), and October-December (early freezing period).

[0086] Factor discretization: The Jenks natural break method was used to divide PRE, T, NDVI, and DEM into 9 categories, while LUCC (6 categories) and ST (12 categories) retained their original classifications.

[0087] The result of q-value calculation. For example... Figure 7 As shown, Figure 7 This diagram illustrates the comparison of q-values ​​of driving factors during different freeze-thaw periods. Statistical analysis of the q-values ​​of different driving factors on the spatial pattern yielded the average surface soil moisture (SSM, unit m³ / m³) in the XX source region from 20XX to 20XX. The results show that temperature dominates during the freezing period, precipitation dominates during the thawing period, NDVI dominates during the wetting period, and temperature dominates at the initial freezing stage, revealing the phased reshaping effect of the freeze-thaw cycle on the driving mechanism.

[0088] This application is adaptable to the characteristics of data sources in data-scarce regions: by fusing multi-source products (high-medium-coarse spatial resolution) and combining them with a small amount of in-situ data for correction, it solves the technical bottlenecks of insufficient in-situ data and limited accuracy of single products. The combined product can effectively capture soil moisture characteristics during the freeze-thaw cycle, filling the data gaps of the freeze-thaw period in some products. A multi-level analysis system is constructed: innovatively integrating global spatial autocorrelation, local hotspot clustering, and spatial outlier identification methods, it achieves comprehensive quantification of heterogeneity from the overall to the local, and from the macro to the micro, overcoming the limitations of single-index analysis. A phased driving mechanism is revealed: considering the special characteristics of the freeze-thaw cycle, four typical stages are divided for driving factor identification, clarifying the dominant roles of temperature (freezing period) and precipitation (thawing period), and quantifying the reshaping effect of freeze-thaw phase change on the heterogeneity driving mechanism. The results are highly reliable: through multi-stage quality control including triple registration error estimation, variance correction, FDR correction, and in-situ data verification, the accuracy of the combined product and the reliability of the heterogeneity analysis results are ensured, making it suitable for typical data-scarce regions such as high-altitude permafrost areas. This application provides a complete and feasible technical solution for analyzing the spatial heterogeneity of soil moisture in data-scarce areas. It can provide key technical support for eco-hydrological simulation, water resource management, and climate change impact assessment in cold regions, and has broad practicality and application prospects.

[0089] To implement the method of the embodiments of this application, Figure 8 A schematic diagram of a hardware structure of an electronic device provided in an embodiment of this application, such as... Figure 8 As shown in the illustration, this application embodiment also provides an electronic device 80 that may include: a memory 801 for storing a computer program; and a processor 802 for implementing the method described above when executing the computer program. The processor 802 can implement the steps of any of the methods described above, which will not be elaborated further here.

[0090] Of course, in practical applications, such as Figure 8 As shown, the electronic device 80 may further include at least one network interface 803. Various components in the electronic device are coupled together via a bus system 804. It is understood that the bus system 804 is used to implement communication between these components. In addition to a data bus, the bus system 804 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 8Various buses are labeled as bus systems 804. The number of processors 802 can be at least one. A network interface 803 is used for wired or wireless communication between electronic devices and other devices. The memory 801 in this embodiment is used to store various types of data to support the operation of the electronic device. The methods disclosed in the above embodiments can be applied to or implemented by the processor 802. The processor 802 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuit of the hardware in the processor 802 or by instructions in software form. The processor 802 can be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 802 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the methods disclosed in the embodiments of this application can be directly reflected in the combined execution of hardware and software modules in a microcontroller. The software module can reside in a storage medium located in memory 801. Processor 802 reads information from memory 801 and, in conjunction with its hardware, completes the steps of the aforementioned method. In an exemplary embodiment, electronic device 80 can be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components to execute the aforementioned method.

[0091] Specifically, embodiments of this application provide a computer-readable storage medium storing a computer program thereon, such as a memory 801 storing the computer program, which can be executed by a processor 802 to complete the aforementioned method steps. The computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM.

[0092] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

Claims

1. A method for assessing the spatial characteristics of soil moisture, characterized in that, include: Acquire at least three sets of detection data corresponding to at least three detection products and initial observation data corresponding to at least one initial observation station within a historical time period of the target area; wherein, the spatial resolution of the detection products corresponds to the spatial resolution of the detection data; At least three of the aforementioned detection data are fused to obtain initial fused data; Based on the initial observation data and the initial fusion data, the target fusion data is determined; The target fusion data is analyzed and processed based on a preset spatial analysis method to obtain the analysis results of the spatial characteristics of soil moisture; the analysis results are used to evaluate the attribution causes of the spatial characteristics of the target area.

2. The method according to claim 1, characterized in that, The detection data is spatially distributed data with a time series; the spatially distributed data includes data corresponding to at least one location; The process of fusing detection data corresponding to at least three of the detection products to obtain initial fused data includes: The detection data at different spatial resolutions are resampled to obtain a target detection dataset with a target spatial resolution; the target detection dataset includes a first subset, a second subset, and a third subset. The initial fused data is obtained by fusing the first subset, the second subset, and the third subset.

3. The method according to claim 2, characterized in that, The detection data includes first data at a first spatial resolution, second data at a second spatial resolution, and third data at a third spatial resolution; the value of the first spatial resolution is smaller than the value of the second spatial resolution. The value of the second spatial resolution is smaller than the value of the third spatial resolution; The resampling process performed on the detection data at different spatial resolutions to obtain a target detection dataset with target spatial resolution includes: The first data is upscaled and resampled based on a preset bilinear resampling method, so that the value of the first spatial resolution is increased to the value of the target spatial resolution, thus obtaining the first subset of data. The second data is downscaled and resampled according to a preset inverse distance weighting method, so that the value of the second spatial resolution is reduced to the value of the target spatial resolution, thus obtaining the second subset data set. The third data is downscaled and resampled according to the inverse distance weighting method, so that the value of the third spatial resolution is reduced to the value of the target spatial resolution, thus obtaining the third subset data.

4. The method according to claim 2, characterized in that, The initial fused data is obtained by fusing the first subset of data, the second subset of data, and the third subset of data, including: Based on the preset triple registration method, the first subset, the second subset, and the third subset, a first variance set of the first subset, a second variance set of the second subset, and a third variance set of the third subset are determined; the first variance set includes at least one first variance corresponding to the position; the second variance set includes at least one second variance corresponding to the position; and the third variance set includes at least one third variance corresponding to the position. Based on the first variance set, the second variance set, and the third variance set, a first weight set for the first subset of data, a second weight set for the second subset of data, and a third weight set for the third subset of data are determined; the first weight set includes at least one first weight corresponding to the position; the second weight set includes at least one second weight corresponding to the position; and the third weight set includes at least one third weight corresponding to the position. The initial fused data is determined based on the first subset, the first weight set, the second subset, the second weight set, the third subset, and the third weight set.

5. The method according to claim 1, characterized in that, The step of determining the target fused data based on the initial observation data and the initial fused data includes: Determine the first standard deviation of the initial observation data; Determine the second standard deviation of the initial fused data based on the initial observation data and the initial fused data; The target fusion data is determined based on the initial fusion data, the first standard deviation, and the second standard deviation.

6. The method according to claim 1, characterized in that, The spatial analysis method includes global analysis, local hotspot analysis, and local spatial correlation analysis; the analysis results include global spatial autocorrelation analysis results, local hotspot clustering analysis results, and spatial clustering results; the analysis and processing of the target fused data based on the preset spatial analysis method to obtain the analysis results of the spatial characteristics of soil moisture includes: The global spatial autocorrelation analysis results are determined based on the global analysis method and the target fusion data; the global spatial autocorrelation analysis results are used to characterize the overall spatial autocorrelation features of the soil moisture. Based on the local hotspot analysis method and the target fusion data, the local hotspot clustering analysis results are determined; the local hotspot clustering analysis results are used to characterize the spatial distribution characteristics of the local clustering of soil moisture. Based on the local spatial correlation analysis method and the target fusion data, the spatial clustering result is determined; the spatial clustering result is used to characterize the local spatial correlation features of soil moisture; the spatial clustering result includes at least one spatial correlation type.

7. The method according to claim 6, characterized in that, The method further includes: The target fusion data is split based on a preset freeze-thaw period to obtain freeze-thaw period data; the freeze-thaw period data includes freezing period data, thawing period data, wetting period data, and initial freezing period data; Based on the global analysis method, the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, determine the global spatial autocorrelation analysis results corresponding to each freeze-thaw period; Based on the local hotspot analysis method, the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, determine the local hotspot cluster analysis results corresponding to each freezing-thaw period; Based on the local spatial correlation analysis method, the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, the spatial clustering results corresponding to each freeze-thaw period are determined.

8. The method according to claim 7, characterized in that, The method further includes: Obtain at least one driving factor corresponding to the spatial characteristics of soil moisture; Based on the fusion data of each driving factor and the target, an explanatory power parameter for each driving factor is determined; the explanatory power parameter is used to characterize the explanatory power of the driving factor for the formation and evolution of the spatial characteristics of soil moisture. Based on the magnitude of the explanatory power parameter, target driving factors are determined from the at least one driving factor for the freezing period data, the thawing period data, the wetting period data, and the initial freezing period data, respectively; the target driving factor is a factor that influences the formation and evolution of the spatial characteristics of soil moisture.

9. The method according to any one of claims 1-8, characterized in that, The method further includes: Based on the initial observation data and the target fusion data, the verification parameters of the target fusion data are determined; If the verification parameters do not meet the preset verification conditions, the target fusion data is re-determined based on the initial observation data and the initial fusion data.

10. A computer-readable storage medium, characterized in that, The computer-readable medium stores a computer program that, when executed by a processor, is used to implement the method according to any one of claims 1 to 9.

Images