Three-dimensional investigation method, device and equipment for obstacle layer of soda saline-alkali soil based on multi-source data fusion

Abstract

The application discloses a soda saline-alkali soil obstacle layer three-dimensional investigation method, device and equipment based on multi-source data fusion, relates to the field of metrology pedology, and comprises the following steps: acquiring ground surface spectrum data by using a unmanned aerial vehicle to analyze the spatial variation scale of a saline-alkali spot, and accordingly, adaptively formulating a multi-frequency electromagnetic induction survey scheme; synchronously acquiring apparent conductivity and spectrum indexes of a survey line; combining typical calibration sections, directly pushing sampling and indoor detection to acquire saturated mud conductivity and alkali degree measured samples; adopting adaptive layered space interpolation to generate a high-resolution section data set; further fusing multi-source data to establish a random forest machine learning prediction model, realizing quantitative prediction of saline-alkali indexes and identification of obstacle layers of the whole line; and finally, generating a plot-scale high-resolution soil obstacle layer three-dimensional distribution map through three-dimensional space reconstruction. The application realizes rapid, nondestructive and high-resolution three-dimensional investigation of obstacle layers of a large range of soda saline-alkali soil by relying on a small amount of drilling calibration models.

Description

Technical Field

[0001] This application relates to the field of quantitative soil science, and in particular to a method, apparatus and equipment for three-dimensional investigation of barrier layers in soda saline-alkali land based on multi-source data fusion. Background Technology

[0002] Soda-saline-alkali soils are mostly formed in semi-arid and semi-humid low-lying closed drainage areas. Shallowly buried soda-type groundwater and heavy clay parent material promote the accumulation and alkalization of salts on the surface, leading to localized differential redistribution of water and salt, and thus forming saline-alkali patches of varying sizes. The salt and alkalized components of this type of soil are mostly concentrated in the surface layer, with surface saline-alkali patches interspersed. Controlled by surface cover, evaporation, and infiltration processes, a differentiated saline-alkali barrier layer forms in the subsurface, with significant vertical differentiation of saline-alkali characteristics in the profile. Sodium soils, saline soils, and salinized sodium soils are spatially interspersed and alternate in this saline-alkali barrier layer. Understanding the spatial pattern of the barrier layer in soda-saline-alkali land can provide support for site management and zonal improvement.

[0003] Traditional surveys of saline-alkali land profile barriers typically employ a combination of field drilling, stratified sampling, and indoor physicochemical testing, with spatial representation achieved through interpolation mapping. While this method offers high data reliability, it is cumbersome, costly in terms of manpower and resources, and only allows for single-point monitoring. Furthermore, limitations in sampling density prevent accurate characterization of barrier profile distribution and spatial heterogeneity, resulting in inefficiencies and low representational precision. Currently, various novel detection technologies have their advantages, but their independent applications have limitations: remote sensing can rapidly acquire surface salinity information over a wide area but cannot probe deeper soil layers; geophysical methods such as electromagnetic induction can non-destructively detect deep soil conductivity but struggle to quantitatively characterize alkali indices; and UAV spectral technology can indirectly reflect the underground salinity status of soda saline-alkali wastelands. Therefore, effective integration of multiple technologies is crucial for the rapid identification of large-scale soda saline-alkali barrier layers. Summary of the Invention

[0004] The purpose of this application is to provide a method, device, and equipment for three-dimensional survey of obstacle layers in soda saline-alkali land based on multi-source data fusion. By combining multi-source data collaboration with machine learning algorithms, the method can improve the efficiency of obstacle identification in soda saline-alkali land and quickly obtain spatial distribution information of soil obstacle types and layers at the field scale.

[0005] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a three-dimensional survey method for the barrier layer of soda-saline-alkali land based on multi-source data fusion, including: UAVs were used to acquire surface spectral data of the plots to be investigated, and the spatial distribution patterns of saline-alkali patches were analyzed. A multi-frequency electromagnetic induction survey scheme was developed based on the spatial distribution pattern of the saline-alkali spots. According to the multi-frequency electromagnetic induction survey scheme, the apparent electrical conductivity of the soil profile along the survey line is obtained. Extract the surface spectral feature data of the corresponding survey line from the surface spectral data; Combining the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, a representative calibration section was planned and borehole sampling was carried out at intervals of 1 / 3 of the spatial autocorrelation range of saline-alkali patches to obtain measured samples of stratified soil salinization index. Using the measured samples of the stratified soil salinization index as the target variable and the apparent electrical conductivity of the soil profile at the corresponding location as the covariate, an adaptive stratified spatial interpolation algorithm is used to generate a high-resolution cross-sectional salinization index dataset. By integrating the high-resolution cross-sectional salinization index dataset, the surface spectral feature data, and the apparent electrical conductivity of the soil profile observed along the entire survey line, a stratified salinization index machine learning prediction model is established. Based on the machine learning prediction model for the stratified salinization index, the prediction results of the salinization index on all survey lines of the target plot are obtained, and the salinity barrier layer is identified accordingly. Based on the results of the identification of the saline-alkali barrier layer, a three-dimensional spatial reconstruction method is used to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested.

[0006] Optionally, the specific process of formulating a multi-frequency electromagnetic induction survey scheme based on the spatial distribution pattern of the saline-alkali spots is as follows: The surface spectral data are image-stitched and orthorectified to extract vegetation index and salinity index, and a spatial distribution map of saline-alkali spots is generated. The spatial autocorrelation range of the saline-alkali patch spatial distribution map was analyzed using a semi-variogram function. One-third or less of the spatial autocorrelation range was taken as the survey line spacing to capture the spatial structure information of the soil to the maximum extent. Based on the target survey depth and the range of soil background conductivity, the working frequency combination of the multi-frequency electromagnetic induction instrument is calculated by using the empirical relationship between detection depth and frequency, thus generating the multi-frequency electromagnetic induction survey scheme.

[0007] Optionally, according to the multi-frequency electromagnetic induction survey scheme, obtaining the apparent electrical conductivity of the soil profile along the survey line specifically includes: Multi-frequency apparent conductivity data were collected along the measurement line using a multi-frequency electromagnetic induction instrument. Based on the layered geodetic medium model, the damped least squares method is used to perform one-dimensional inversion on the multi-frequency apparent conductivity data, converting the apparent conductivity at different frequencies into layered apparent conductivity at continuous depths underground, and obtaining the apparent conductivity of the soil profile of the survey line.

[0008] Optionally, the specific process for planning representative calibration sections is as follows: Based on the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, K-means clustering analysis was performed to divide the surveyed plots into multiple variable units that include at least severely saline-alkali areas, moderately saline-alkali areas and non-saline-alkali background areas. Within each variation unit, a measurement line running through the unit is selected as the representative calibration section to ensure that the representative calibration section covers the entire salinization gradient.

[0009] Optionally, the specific process of using the adaptive hierarchical spatial interpolation algorithm is as follows: Calculate the coefficient of determination for the linear fit between the target variable and the covariate at different standard depth layers; When the determination coefficient is greater than a preset threshold, one-dimensional regression kriging is used for auxiliary interpolation. When the determination coefficient is less than or equal to the preset threshold, the PCHIP piecewise cubic Hermite interpolation method is used to perform interpolation without auxiliary variables, generating continuous saturated mud conductivity and alkalinity data on the survey line profile, which constitutes the high-resolution cross-sectional salinization index dataset.

[0010] Optionally, the specific process for establishing a machine learning prediction model for stratified salinization indicators is as follows: The machine learning prediction model is a random forest regression model; After feature screening of the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, a covariate feature set is formed. The saturated mud conductivity and alkalinity in the high-resolution cross-sectional salinization index dataset were used as target labels. Multiple training subsets are generated by sampling with replacement using Bootstrap. At each node, feature splitting is performed based on the principle of minimizing the Gini coefficient to construct multiple decision trees. The outputs of all decision trees are averaged to establish a mapping relationship between the covariate feature set and the target label, thus obtaining the hierarchical salinization index machine learning prediction model.

[0011] Optionally, the specific rules for identifying the saline-alkali barrier layer are as follows: Each survey line location and its depth stratification unit are used as the judgment object. The stratified salinization index machine learning prediction model is input to obtain the predicted saturated mud conductivity ECe and alkalinity ESP. When ECe ≥ 2 dS / m and ESP < 15%, it is determined to be a saline soil barrier layer; When ECe < 4 dS / m and ESP ≥ 15%, it is determined to be a sodium-rich soil barrier layer; When ECe ≥ 4 dS / m and ESP ≥ 15%, it is determined to be a saline-sodium soil barrier layer; When ECe < 2 dS / m and ESP < 15%, it is determined to be non-saline-alkali soil.

[0012] Optionally, the specific process of using the three-dimensional spatial reconstruction method is as follows: Using the results of the saline-alkali barrier layer identification as three-dimensional spatial discrete point data, the three-dimensional kriging spatial interpolation method is adopted to consider the spatial anisotropy in the horizontal and vertical directions, and the soil barrier layer of the entire plot is spatially reconstructed to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested.

[0013] Secondly, this application provides a three-dimensional survey device for the barrier layer of soda saline-alkali land based on multi-source data fusion, comprising: The module for analyzing the spatial distribution pattern of saline-alkali patches is used to acquire surface spectral data of the plot under investigation using UAVs and analyze the spatial distribution pattern of saline-alkali patches. A multi-frequency electromagnetic induction survey scheme formulation module is used to formulate a multi-frequency electromagnetic induction survey scheme based on the spatial distribution law of the saline-alkali spots. The apparent conductivity acquisition module for the soil profile of the survey line is used to acquire the apparent conductivity of the soil profile of the survey line according to the multi-frequency electromagnetic induction survey scheme. The surface spectral feature data extraction module is used to extract the surface spectral feature data corresponding to the survey line from the surface spectral data. The layered soil salinization index measured sample acquisition module is used to combine the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, plan representative calibration sections and perform borehole sampling with an interval of 1 / 3 of the spatial autocorrelation range of saline-alkali patches to obtain layered soil salinization index measured samples. The high-resolution cross-sectional salinity index dataset generation module is used to generate a high-resolution cross-sectional salinity index dataset by using the measured samples of the stratified soil salinity index as the target variable, the apparent electrical conductivity of the soil profile at the corresponding location as the covariate, and an adaptive stratified spatial interpolation algorithm. A stratified salinization index machine learning prediction model construction module is used to integrate the high-resolution cross-sectional salinization index dataset, the surface spectral feature data, and the apparent electrical conductivity of the soil profile observed along the entire survey line to establish a stratified salinization index machine learning prediction model. The saline-alkali barrier layer identification module is used to obtain the predicted results of salinization indicators on all survey lines of the target plot based on the layered salinization index machine learning prediction model, and to complete the identification of the saline-alkali barrier layer accordingly. The three-dimensional distribution map generation module is used to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested by using a three-dimensional spatial reconstruction method based on the results of the identification of the saline-alkali barrier layer.

[0014] Thirdly, this application provides a computer device, including: 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 steps of the three-dimensional survey method for the barrier layer of soda saline-alkali land based on multi-source data fusion as described above.

[0015] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides a method, apparatus, and equipment for three-dimensional investigation of the barrier layer in soda saline-alkali land based on multi-source data fusion, with the following advantages: Space-ground collaboration and adaptive schemes: By utilizing the spatial variation scale of UAV spectral analysis, electromagnetic survey schemes are adaptively formulated, avoiding the waste of resources caused by blindly deploying survey lines and ensuring the matching of the detection scale with the spatial characteristics of saline-alkali spots.

[0016] Deep fusion of multi-source data: By combining adaptive hierarchical spatial interpolation with random forest machine learning models, electromagnetic induction data (reflecting deep salinity) and spectral data (reflecting surface stress) are cleverly used as covariates, breaking through the technical bottleneck that a single geophysical method cannot quantitatively characterize alkalinity (ESP).

[0017] Precise 3D characterization: The final generated 3D distribution map can intuitively show the spatial distribution pattern of barrier layers such as saline soil and sodium soil, providing direct data support for subsequent precise salt drainage and variable application of soil conditioners. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a three-dimensional survey method for the barrier layer of soda saline-alkali land based on multi-source data fusion, provided as an embodiment of this application; Figure 2 A schematic diagram of the salinization index and barrier layer identification of the survey line provided in an embodiment of this application; Figure 3 A schematic diagram of the results of a site obstacle layer survey provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0020] 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.

[0021] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] In one exemplary embodiment, such as Figure 1 As shown, a three-dimensional survey method for the barrier layer of soda saline-alkali land based on multi-source data fusion is provided. Specifically, it is a detailed survey method executed by computer equipment, which can be executed by a computer device such as a terminal or server alone, or by a terminal and server together. In the embodiments of this application, the method includes the following steps: Step 101: Use drones to acquire surface spectral data of the plots to be investigated and analyze the spatial distribution pattern of saline-alkali spots.

[0023] Step 102: Develop a multi-frequency electromagnetic induction survey scheme based on the spatial distribution pattern of the saline-alkali spots.

[0024] UAVs were used to collect surface spectral data of the site in July, as this data is more conducive to analyzing the spatial distribution patterns of saline-alkali patches. Multispectral images of the study area were acquired using a UAV platform, and high-resolution spatial data were generated through image stitching, geometric correction, and orthorectification. Variables such as original spectral bands, vegetation indices, and brightness were further extracted. A multi-frequency electromagnetic induction instrument (GEM-2) survey plan was developed based on the analysis of the spatial variation scale characteristics of saline-alkali patches and the survey depth of interest. The survey line spacing was 5 m, and the operating frequencies of the GEM-2 were set to 20025 Hz, 30025 Hz, 50025 Hz, 70025 Hz, and 90025 Hz.

[0025] Step 103: Obtain the apparent electrical conductivity of the soil profile along the survey line according to the multi-frequency electromagnetic induction survey scheme.

[0026] Following the survey route determined in step 102, a drone equipped with a multi-frequency electromagnetic induction device was used to collect land parcel data. After data processing, the apparent electrical conductivity of the soil profiles along all survey lines was obtained. At the same time, the surface spectral characteristic data of the corresponding survey lines were extracted from step 101, including spectral indices such as NDVI, GNDVI, NDRE, PSRI, and SAVI, as well as data and brightness of each band.

[0027] Step 104: Extract the surface spectral feature data corresponding to the survey line from the surface spectral data.

[0028] Step 105: Combining the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, plan representative calibration sections and conduct borehole sampling at intervals of 1 / 3 of the spatial autocorrelation range of saline-alkali patches to obtain measured samples of stratified soil salinization indicators.

[0029] Based on the obtained surface spectral characteristics and subsurface apparent electrical conductivity distribution characteristics, 2-3 representative calibration sections were planned along the survey line. These sections were selected to cover both typical saline-alkali patches and non-saline-alkali patches. A series of boreholes were drilled at predetermined intervals along each section. Uncircular soil core samples at a depth of 1.2 meters were collected using a direct-push sampling device, with one soil sample taken every 10 cm, for a total of 17 boreholes. The soil samples were then subjected to laboratory analysis to determine two core salinization indicators: saturated mud conductivity (ECe) and alkalinity (ESP).

[0030] Step 106: Using the measured samples of the stratified soil salinization index as the target variable and the apparent electrical conductivity of the soil profile at the corresponding location as the covariate, an adaptive stratified spatial interpolation algorithm is used to generate a high-resolution cross-sectional salinization index dataset.

[0031] Measured samples of saturated mud conductivity and alkalinity from boreholes along the cross-section were selected, along with covariate data of apparent conductivity measured by GEM-2 electromagnetic induction at corresponding locations. Then, based on the correlation between apparent conductivity and soil properties, an adaptive interpolation method was selected. When the coefficient of determination of the linear fit between the target variable and apparent conductivity exceeded a specified threshold, one-dimensional regression kriging was used for interpolation; otherwise, pchip interpolation was employed. This completed the adaptive stratified spatial interpolation of the borehole data from the calibration cross-section, outputting a high-resolution continuous salinity index dataset for the cross-section.

[0032] Step 107: Integrate the high-resolution cross-sectional salinization index dataset, the surface spectral feature data, and the apparent electrical conductivity of the soil profile observed along the entire survey line to establish a stratified salinization index machine learning prediction model.

[0033] Step 108: Based on the machine learning prediction model of the stratified salinization index, obtain the prediction results of salinization index on all survey lines of the target plot, and complete the identification of the salinity barrier layer accordingly.

[0034] The cross-sectional salinization index dataset obtained in step 106, along with the surface spectral index and soil apparent electrical conductivity data acquired in steps 103 and 104, are integrated. Feature filtering is performed on these covariate sets, and a predetermined number or conditions of covariates are selected based on the feature filtering results to establish a stratified random forest machine learning prediction model for salinization indices. Using this model in conjunction with full-line observation data, prediction results for salinization indices along all survey lines in the target area are obtained.

[0035] Based on the ECe and alkalinity values ​​in the classification standard for soda saline-alkali soil barriers, a comprehensive judgment was made to identify the saline-alkali barrier layers (saline soil, sodium soil, and salinized sodium soil) on the survey line profile. Specifically, when ECe ≥ 2 dS / m and ESP < 15%, it was identified as saline soil; when ECe < 4 dS / m and ESP ≥ 15%, it was identified as sodium soil; when ECe ≥ 4 dS / m and ESP ≥ 15%, it was identified as salinized-sodium soil; and when ECe < 2 dS / m and ESP < 15%, it could be identified as non-saline-alkali soil. For details, please refer to [reference needed]. Figure 2 .

[0036] Step 109: Based on the results of the identification of the saline-alkali barrier layer, a three-dimensional spatial reconstruction method is used to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested.

[0037] The specific process is as follows: Using the results of the saline-alkali barrier layer identification as three-dimensional spatial discrete point data, a three-dimensional kriging spatial interpolation method is employed to reconstruct the soil barrier layer across the entire plot, considering the spatial anisotropy in both the horizontal and vertical directions. This generates a high-resolution three-dimensional distribution map of the soil barrier layer in the plot under test. See details... Figure 3 .

[0038] Based on the same inventive concept, this application also provides a device for three-dimensional survey of soda-saline-alkali land obstacle layers based on multi-source data fusion, used to implement the aforementioned method for three-dimensional survey of soda-saline-alkali land obstacle layers based on multi-source data fusion. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the device for three-dimensional survey of soda-saline-alkali land obstacle layers based on multi-source data fusion provided below can be found in the limitations of the method for three-dimensional survey of soda-saline-alkali land obstacle layers based on multi-source data fusion described above, and will not be repeated here.

[0039] In one exemplary embodiment, a three-dimensional survey device for the barrier layer of soda saline-alkali land based on multi-source data fusion is provided, comprising: The module for analyzing the spatial distribution pattern of saline-alkali patches is used to acquire surface spectral data of the plot under investigation using UAVs and analyze the spatial distribution pattern of saline-alkali patches. A multi-frequency electromagnetic induction survey scheme formulation module is used to formulate a multi-frequency electromagnetic induction survey scheme based on the spatial distribution law of the saline-alkali spots. The apparent conductivity acquisition module for the soil profile of the survey line is used to acquire the apparent conductivity of the soil profile of the survey line according to the multi-frequency electromagnetic induction survey scheme. The surface spectral feature data extraction module is used to extract the surface spectral feature data corresponding to the survey line from the surface spectral data. The layered soil salinization index measured sample acquisition module is used to combine the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, plan representative calibration sections and perform borehole sampling with an interval of 1 / 3 of the spatial autocorrelation range of saline-alkali patches to obtain layered soil salinization index measured samples. The high-resolution cross-sectional salinity index dataset generation module is used to generate a high-resolution cross-sectional salinity index dataset by using the measured samples of the stratified soil salinity index as the target variable, the apparent electrical conductivity of the soil profile at the corresponding location as the covariate, and an adaptive stratified spatial interpolation algorithm. A stratified salinization index machine learning prediction model construction module is used to integrate the high-resolution cross-sectional salinization index dataset, the surface spectral feature data, and the apparent electrical conductivity of the soil profile observed along the entire survey line to establish a stratified salinization index machine learning prediction model. The saline-alkali barrier layer identification module is used to obtain the predicted results of salinization indicators on all survey lines of the target plot based on the layered salinization index machine learning prediction model, and to complete the identification of the saline-alkali barrier layer accordingly. The three-dimensional distribution map generation module is used to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested by using a three-dimensional spatial reconstruction method based on the results of the identification of the saline-alkali barrier layer.

[0040] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 4As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores three-dimensional survey data of the soda-saline-alkali land barrier layer based on multi-source data fusion. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a three-dimensional survey method for the soda-saline-alkali land barrier layer based on multi-source data fusion.

[0041] Those skilled in the art will understand that Figure 4 The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0042] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0043] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0044] 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, data stored, data displayed, 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 the relevant data must comply with relevant regulations.

[0045] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0046] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0047] 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.

[0048] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A three-dimensional survey method for the barrier layer of soda-saline-alkali land based on multi-source data fusion, characterized in that, The three-dimensional survey method for the barrier layer of soda saline-alkali land based on multi-source data fusion includes: UAVs were used to acquire surface spectral data of the plots to be investigated, and the spatial distribution patterns of saline-alkali patches were analyzed. A multi-frequency electromagnetic induction survey scheme was developed based on the spatial distribution pattern of the saline-alkali spots. According to the multi-frequency electromagnetic induction survey scheme, the apparent electrical conductivity of the soil profile along the survey line is obtained. Extract the surface spectral feature data of the corresponding survey line from the surface spectral data; Combining the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, a representative calibration section was planned and borehole sampling was carried out at intervals of 1 / 3 of the spatial autocorrelation range of saline-alkali patches to obtain measured samples of stratified soil salinization index. Using the measured samples of the stratified soil salinization index as the target variable and the apparent electrical conductivity of the soil profile at the corresponding location as the covariate, an adaptive stratified spatial interpolation algorithm is used to generate a high-resolution cross-sectional salinization index dataset. By integrating the high-resolution cross-sectional salinization index dataset, the surface spectral feature data, and the apparent electrical conductivity of the soil profile observed along the entire survey line, a stratified salinization index machine learning prediction model is established. Based on the machine learning prediction model for the stratified salinization index, the prediction results of the salinization index on all survey lines of the target plot are obtained, and the salinity barrier layer is identified accordingly. Based on the results of the saline-alkali barrier layer identification, a three-dimensional spatial reconstruction method is used to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested. The specific process for planning representative calibration sections is as follows: Based on the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, K-means clustering analysis was performed to divide the surveyed plots into multiple variable units that include at least severely saline-alkali areas, moderately saline-alkali areas and non-saline-alkali background areas. Within each variation unit, a measurement line traversing the unit is selected as the representative calibration section to ensure that the representative calibration section covers the entire salinization gradient. The specific process of using the adaptive hierarchical spatial interpolation algorithm is as follows: Calculate the coefficient of determination for the linear fit between the target variable and the covariate at different standard depth layers; When the determination coefficient is greater than a preset threshold, one-dimensional regression kriging is used for auxiliary interpolation. When the determination coefficient is less than or equal to the preset threshold, the PCHIP piecewise cubic Hermite interpolation method is used to perform interpolation without auxiliary variables, generating continuous saturated mud conductivity and alkalinity data on the survey line profile, which constitutes the high-resolution cross-sectional salinization index dataset.

2. The method for three-dimensional investigation of barrier layers in soda-saline-alkali land based on multi-source data fusion according to claim 1, characterized in that, The specific process for formulating a multi-frequency electromagnetic induction survey scheme based on the aforementioned spatial distribution pattern of saline-alkali spots is as follows: The surface spectral data are image-stitched and orthorectified to extract vegetation index and salinity index, and a spatial distribution map of saline-alkali spots is generated. The spatial autocorrelation range of the saline-alkali patch spatial distribution map was analyzed using a semi-variogram function. One-third or less of the spatial autocorrelation range was taken as the survey line spacing to capture the spatial structure information of the soil to the maximum extent. Based on the target survey depth and the range of soil background conductivity, the working frequency combination of the multi-frequency electromagnetic induction instrument is calculated by using the empirical relationship between detection depth and frequency, thus generating the multi-frequency electromagnetic induction survey scheme.

3. The method for three-dimensional investigation of barrier layers in soda-saline-alkali land based on multi-source data fusion according to claim 1, characterized in that, According to the aforementioned multi-frequency electromagnetic induction survey scheme, obtaining the apparent electrical conductivity of the soil profile along the survey line specifically includes: Multi-frequency apparent conductivity data were collected along the measurement line using a multi-frequency electromagnetic induction instrument. Based on the layered geodetic medium model, the damped least squares method is used to perform one-dimensional inversion on the multi-frequency apparent conductivity data, converting the apparent conductivity at different frequencies into layered apparent conductivity at continuous depths underground, and obtaining the apparent conductivity of the soil profile of the survey line.

4. The method for three-dimensional investigation of barrier layers in soda-saline-alkali land based on multi-source data fusion according to claim 1, characterized in that, The specific process of establishing a machine learning prediction model for stratified salinization indicators is as follows: The machine learning prediction model is a random forest regression model; After feature screening of the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, a covariate feature set is formed. The saturated mud conductivity and alkalinity in the high-resolution cross-sectional salinization index dataset were used as target labels. Multiple training subsets are generated by sampling with replacement using Bootstrap. At each node, feature splitting is performed based on the principle of minimizing the Gini coefficient to construct multiple decision trees. The outputs of all decision trees are averaged to establish a mapping relationship between the covariate feature set and the target label, thus obtaining the hierarchical salinization index machine learning prediction model.

5. The method for three-dimensional investigation of barrier layers in soda-saline-alkali land based on multi-source data fusion according to claim 1, characterized in that, The specific rules for completing the identification of the saline-alkali barrier layer are as follows: Each survey line location and its depth stratification unit are used as the judgment object. The stratified salinization index machine learning prediction model is input to obtain the predicted saturated mud conductivity ECe and alkalinity ESP. When ECe ≥ 2 dS / m and ESP < 15%, it is determined to be a saline soil barrier layer; When ECe < 4 dS / m and ESP ≥ 15%, it is determined to be a sodium-rich soil barrier layer; When ECe ≥ 4 dS / m and ESP ≥ 15%, it is determined to be a saline-sodium soil barrier layer; When ECe < 2 dS / m and ESP < 15%, it is determined to be non-saline-alkali soil.

6. The method for three-dimensional investigation of barrier layers in soda-saline-alkali land based on multi-source data fusion according to claim 1, characterized in that, The specific process of using the three-dimensional spatial reconstruction method is as follows: Using the results of the saline-alkali barrier layer identification as three-dimensional spatial discrete point data, the three-dimensional kriging spatial interpolation method is adopted to consider the spatial anisotropy in the horizontal and vertical directions, and the soil barrier layer of the entire plot is spatially reconstructed to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested.

7. A three-dimensional survey device for the barrier layer of soda-saline-alkali land based on multi-source data fusion, characterized in that, The three-dimensional survey device for the barrier layer of soda saline-alkali land based on multi-source data fusion is applied to the three-dimensional survey method for the barrier layer of soda saline-alkali land based on multi-source data fusion as described in any one of claims 1-6. The three-dimensional survey device for the barrier layer of soda saline-alkali land based on multi-source data fusion includes: The module for analyzing the spatial distribution pattern of saline-alkali patches is used to acquire surface spectral data of the plot under investigation using UAVs and analyze the spatial distribution pattern of saline-alkali patches. A multi-frequency electromagnetic induction survey scheme formulation module is used to formulate a multi-frequency electromagnetic induction survey scheme based on the spatial distribution law of the saline-alkali spots. The apparent conductivity acquisition module for the soil profile of the survey line is used to acquire the apparent conductivity of the soil profile of the survey line according to the multi-frequency electromagnetic induction survey scheme. The surface spectral feature data extraction module is used to extract the surface spectral feature data corresponding to the survey line from the surface spectral data. The layered soil salinization index measured sample acquisition module is used to combine the surface spectral characteristic data and the apparent electrical conductivity of the soil profile of the survey line, plan representative calibration sections and perform borehole sampling with an interval of 1 / 3 of the spatial autocorrelation range of saline-alkali patches to obtain layered soil salinization index measured samples. The high-resolution cross-sectional salinity index dataset generation module is used to generate a high-resolution cross-sectional salinity index dataset by using the measured samples of the stratified soil salinity index as the target variable, the apparent electrical conductivity of the soil profile at the corresponding location as the covariate, and an adaptive stratified spatial interpolation algorithm. A stratified salinization index machine learning prediction model construction module is used to integrate the high-resolution cross-sectional salinization index dataset, the surface spectral feature data, and the apparent electrical conductivity of the soil profile observed along the entire survey line to establish a stratified salinization index machine learning prediction model. The saline-alkali barrier layer identification module is used to obtain the predicted results of salinization indicators on all survey lines of the target plot based on the layered salinization index machine learning prediction model, and to complete the identification of the saline-alkali barrier layer accordingly. The three-dimensional distribution map generation module is used to generate a high-resolution three-dimensional distribution map of the soil barrier layer of the plot to be tested by using a three-dimensional spatial reconstruction method based on the results of the identification of the saline-alkali barrier layer.

8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the three-dimensional survey method for the barrier layer of soda saline-alkali land based on multi-source data fusion as described in any one of claims 1-6.

Images