Contaminated soil treatment method based on big data analysis

Abstract

The present application provides a method for the treatment of contaminated soil based on big data analysis, which relates to the field of data processing technology, including: taking soil color, soil structure and plant growth distribution characteristics as constraints, searching based on big data, and obtaining pollution distribution characteristics of multiple samples; constructing a predicted pollution distribution map based on the characteristic clustering results, and arranging multiple optimized sampling points; obtaining multiple soil pollution detection results; constructing a pollutant distribution simulation map, determining pollution characteristics of multiple local areas; inputting pollution characteristics of multiple local areas into the remediation analysis channel, and outputting an optimized treatment strategy. This application can solve the technical problems that the traditional method cannot accurately predict soil pollution characteristics according to the environmental status, resulting in poor accuracy in the layout of soil sampling points, and cannot accurately set up a suitable remediation plan, resulting in poor soil treatment quality and efficiency. It can improve the accuracy and scientificity of the sampling point layout, and improve the quality and efficiency of contaminated soil treatment.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method for treating contaminated soil based on big data analysis. Background Art

[0002] With the rapid development of industrialization and urbanization, the problem of soil contamination treatment has become increasingly prominent. Contaminated soil not only affects soil fertility and productivity, but may also affect human health through the food chain. Therefore, soil contamination treatment has become an important part of environmental protection and public health.

[0003] At present, traditional methods of contaminated soil treatment are unable to quickly and accurately predict soil pollution characteristics based on environmental conditions, resulting in poor accuracy in the layout of soil sampling points, making the accuracy, comprehensiveness and reliability of soil testing results insufficient. On the other hand, it is impossible to quickly and accurately set up appropriate soil remediation plans based on pollution detection results, ultimately causing technical problems in the quality and efficiency of contaminated soil treatment. Summary of the invention

[0004] The purpose of this application is to provide a contaminated soil treatment method based on big data analysis to solve the problem that traditional contaminated soil treatment methods cannot quickly and accurately predict soil pollution characteristics based on environmental conditions, resulting in poor accuracy in the layout of soil sampling points, making the accuracy, comprehensiveness and reliability of soil detection results insufficient. On the other hand, it is impossible to quickly and accurately set up a suitable soil remediation plan based on the pollution detection results, which ultimately leads to technical problems such as poor quality and efficiency of contaminated soil treatment.

[0005] In view of the above problems, the present application provides a method for contaminated soil treatment based on big data analysis, including: taking aerial photos of the target contaminated area by a drone to obtain a set of contaminated area images, performing feature extraction on the contaminated area image set, and determining a regional feature set, wherein the regional feature set includes soil color distribution characteristics, soil structure distribution characteristics, and plant growth distribution characteristics; taking the soil color distribution characteristics, soil structure distribution characteristics, and plant growth distribution characteristics as constraints, performing positive sample retrieval based on big data to obtain multiple sample pollution distribution characteristics; performing feature clustering on the multiple sample pollution distribution characteristics, constructing a predicted pollution distribution map based on the feature clustering results, and arranging multiple optimized sampling points based on the predicted pollution distribution map; and arranging multiple optimized sampling points based on the multiple optimized sampling points. Soil sampling is performed on the target polluted area, and pollution detection is performed on multiple soil sampling results to obtain multiple soil pollution detection results, wherein the soil pollution detection results include soil property characteristics, soil environmental characteristics, pollutant detection types and pollutant detection concentrations; simulation analysis is performed on the target polluted area based on the multiple soil pollution detection results to determine the predicted pollution distribution trend and the predicted pollution spread path, a pollutant distribution simulation map is constructed, the pollutant distribution simulation map is divided, and pollution characteristics of multiple local areas are determined; the pollution characteristics of the multiple local areas are input into a soil remediation analysis channel, and multiple local remediation schemes are output. An optimized governance strategy is constructed based on the multiple local remediation schemes, and soil remediation of the target polluted area is performed according to the optimized governance strategy.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] By extracting features from the polluted area image set, the regional feature set is determined, and then the regional feature set is used as a constraint condition to retrieve positive samples based on big data to obtain multiple sample pollution distribution characteristics; then the pollution distribution characteristics of multiple samples are clustered, and a predicted pollution distribution map is constructed based on the feature clustering results, and multiple optimized sampling points are targetedly arranged according to the predicted pollution distribution map; further soil sampling is carried out based on multiple optimized sampling points, and pollution detection is carried out to obtain multiple soil pollution detection results; then the target polluted area is simulated and analyzed based on the multiple soil pollution detection results to determine the predicted pollution distribution trend and the predicted pollution spread path, and a pollutant distribution simulation map is constructed; then the pollutant distribution simulation map is divided to determine multiple local area pollution characteristics; finally, the pollution characteristics of multiple local areas are input into the soil remediation analysis channel, and the optimized governance strategy is output to execute soil governance; the accuracy and efficiency of soil pollution feature prediction can be improved, thereby improving the accuracy and scientificity of sampling point layout, and then a soil remediation analysis channel is constructed based on big data technology, and a suitable soil remediation plan is determined according to the soil pollution characteristics through rapid analysis, so as to achieve the technical effect of improving the quality and efficiency of polluted soil governance.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0010] Figure 1 A schematic diagram of a process for the treatment of contaminated soil based on big data analysis for this application;

[0011] Figure 2 This is a schematic diagram of the process of constructing a soil color feature matrix in a contaminated soil treatment method based on big data analysis in this application. DETAILED DESCRIPTION

[0012] This application provides a method for the treatment of contaminated soil based on big data analysis, which solves the problem that the traditional method for the treatment of contaminated soil cannot quickly and accurately predict the characteristics of soil pollution according to the environmental status, resulting in poor accuracy of the layout of soil sampling points, making the accuracy, comprehensiveness and reliability of soil detection results insufficient. On the other hand, it is impossible to quickly and accurately set up a suitable soil remediation plan based on the pollution detection results, which ultimately leads to technical problems such as poor quality and efficiency of contaminated soil treatment. The accuracy and efficiency of soil pollution characteristic prediction can be improved, thereby improving the accuracy and scientificity of sampling point layout. Then, based on big data technology, a suitable soil remediation plan can be quickly analyzed and determined according to soil pollution characteristics, thereby achieving the technical effect of improving the quality and efficiency of contaminated soil treatment.

[0013] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0014] For examples, please see the attached Figure 1 , this application provides a method for contaminated soil treatment based on big data analysis, which specifically includes the following steps:

[0015] Step 1: Use a drone to take aerial photos of the target polluted area to obtain a polluted area image set, perform feature extraction on the polluted area image set, and determine a regional feature set, wherein the regional feature set includes soil color distribution features, soil structure distribution features, and plant growth distribution features.

[0016] Specifically, drones are used to take aerial photos of the target polluted area according to a predetermined plan. The target polluted area refers to the polluted area to be treated, and a set of polluted area images is obtained; then, features related to soil pollution are extracted from the polluted area image set, such as using image processing technology, image segmentation, feature extraction and other methods to extract features from remote sensing images. Feature recognition models can also be constructed based on convolutional neural networks to extract features, and regional feature sets are obtained. The regional feature set includes soil color distribution features, soil structure distribution features and plant growth distribution features. Soil color can reflect the type and concentration of pollutants in the soil. For example, heavy metal pollution may cause changes in soil color, such as grayish white, brown or purple; soil structure, such as density, porosity, aggregate size, etc., can reflect the migration and accumulation of pollutants; plant growth conditions can reflect the bioavailability of pollutants in the soil, and can indirectly reflect the type and concentration of pollutants. By obtaining a regional feature set, the degree and scope of soil pollution can be more accurately assessed, providing a basis for subsequent prediction of pollutant distribution.

[0017] Step 2: Taking the soil color distribution characteristics, soil structure distribution characteristics and plant growth distribution characteristics as constraints, perform positive sample retrieval based on big data to obtain pollution distribution characteristics of multiple samples.

[0018] Specifically, big data is a series of technologies and methods that enable the value contained in massive data to be mined and displayed, including data planning, data collection, analysis and mining, and other technical means, which can improve the efficiency, accuracy and comprehensiveness of data retrieval. Then, the soil color distribution characteristics, soil structure distribution characteristics and plant growth distribution characteristics are used as retrieval constraints, and positive sample retrieval is performed based on big data, that is, using big data technology to retrieve known polluted samples, quickly find samples that match specific conditions or characteristics from a large amount of data, and obtain multiple sample pollution distribution characteristics, where the sample pollution distribution characteristics include sample pollutant type and sample pollutant concentration. By setting retrieval constraints and performing positive sample retrieval based on big data, known polluted samples similar to the target pollution area can be efficiently found, improving the efficiency and accuracy of sample data acquisition.

[0019] Step three: perform feature clustering on the pollution distribution features of the multiple samples, construct a predicted pollution distribution map based on the feature clustering results, and arrange multiple optimized sampling points according to the predicted pollution distribution map.

[0020] Specifically, the pollution distribution features of the multiple samples are then clustered, where feature clustering refers to extracting multiple high-frequency similar features from the pollution distribution features of the multiple samples, and high-frequency similar features refer to similar features that appear frequently in multiple samples, including high-frequency pollutant types and high-frequency pollutant concentrations, to obtain feature clustering results in multiple regions; then the feature clustering results in multiple regions are combined to generate a predicted pollution distribution map. By extracting high-frequency features from the pollution distribution features of multiple samples and constructing a predicted pollution distribution map, the accuracy and reliability of the predicted pollution distribution map can be improved, and the accuracy and rationality of the subsequent soil sampling point layout can be improved.

[0021] Then, multiple optimized sampling points are arranged according to the predicted pollution distribution map. First, the predicted pollution distribution map is used to identify key areas, determine areas with high pollutant concentrations and sensitive areas where pollutants may diffuse, which are key locations for pollutant migration and accumulation; then, multiple optimized sampling points are arranged with full coverage of high-concentration areas and sensitive areas as constraints to ensure that the sampling points cover areas with high pollutant concentrations and possible pollutant migration paths, while considering the uniformity of the sampling points to reduce sampling errors. By arranging multiple optimized sampling points according to the predicted pollution distribution map, the accuracy and scientificity of the sampling point layout can be improved, providing a basis for subsequent soil sampling and analysis.

[0022] Step 4: Based on the multiple optimized sampling points, soil sampling is performed on the target contaminated area, and pollution detection is performed on the multiple soil sampling results to obtain multiple soil pollution detection results, wherein the soil pollution detection results include soil property characteristics, soil environmental characteristics, pollutant detection types and pollutant detection concentrations.

[0023] Specifically, first, soil sampling is performed on the target contaminated area based on the multiple optimized sampling points to obtain multiple soil samples; then, pollution detection is performed on the multiple soil samples to analyze the pollutants in the soil, and the detection results include soil property characteristics (such as pH value, organic matter content, etc.), soil environmental characteristics (such as soil temperature, soil moisture, etc.), pollutant detection types and pollutant detection concentrations. Through soil sampling and pollution detection based on optimized sampling points, detailed information on the pollution status of the target contaminated area can be obtained, the accuracy, comprehensiveness and reliability of soil detection results can be improved, and a scientific basis can be provided for the subsequent treatment and remediation of contaminated soil.

[0024] Step 5: Perform simulation analysis on the target polluted area based on the multiple soil pollution detection results, determine the predicted pollution distribution trend and the predicted pollution spread path, construct a pollutant distribution simulation map, divide the pollutant distribution simulation map, and determine the pollution characteristics of multiple local areas.

[0025] Specifically, digital twin technology is a virtual, digital entity that corresponds to an entity in the real world. It usually models and simulates physical entities in the real world through data analysis and simulation technology. Based on digital twin technology, numerical simulation methods (such as finite element analysis, fluid mechanics simulation, etc.) or geographic information system (GIS) tools are used to simulate and analyze the target pollution area, simulate the migration and diffusion process of pollutants, and determine the pollution distribution trend and pollution spread path; then, based on the simulation analysis results, a pollutant distribution simulation map is constructed. The simulation map should reflect the distribution pattern and trend of pollutants, as well as the pollution spread path.

[0026] The pollutant distribution simulation map is further divided according to a predetermined grid size, that is, the pollutant distribution simulation map is divided into multiple grid areas; then the multiple grid areas are similarly fused, that is, the areas that meet the predetermined similarity conditions are merged into a local area, wherein in the same local area, the pollution characteristics can be the same by default, and multiple local area pollution characteristics of multiple local areas are obtained. By gridding and fusing the pollutant distribution simulation map, the accuracy of local area division can be improved, and at the same time, the contaminated soil is treated according to the pollution characteristics of the divided local areas, which can improve the precision of the contaminated soil treatment and thus improve the quality of the contaminated soil treatment.

[0027] Step 6: Input the pollution characteristics of the multiple local areas into the soil remediation analysis channel, output multiple local remediation schemes, construct an optimized governance strategy based on the multiple local remediation schemes, and perform soil remediation in the target polluted area according to the optimized governance strategy.

[0028] Specifically, multiple sample remediation schemes are obtained based on big data retrieval, where the sample remediation schemes include sample pollutant types, sample pollutant concentrations, and sample remediation methods; the remediation effects of multiple sample remediation schemes are further evaluated according to the scheme evaluation function, and multiple sample remediation schemes with higher evaluation results are selected as preferred remediation schemes to obtain multiple preferred remediation schemes. Further, multiple preferred remediation schemes are used as training data to iteratively train the random decision forest to construct a soil remediation analysis channel, where the input data of the soil remediation analysis channel is the pollutant type set and the pollutant concentration set, and the output data is the soil remediation method.

[0029] Then, the pollution characteristics of the multiple local areas are respectively input into the soil remediation analysis channel, and multiple local remediation schemes are output; further, a mapping association between the local remediation scheme and the local area is established, and an optimized governance strategy is constructed according to the multiple local areas and the multiple local remediation schemes, and finally, the soil governance of the target polluted area is executed according to the optimized governance strategy. By constructing a soil remediation analysis channel based on big data and random decision forests, the technical goal of quickly analyzing and determining the appropriate soil remediation scheme according to the soil pollution characteristics can be achieved, thereby achieving the technical effect of improving the quality and efficiency of contaminated soil governance.

[0030] By providing a contaminated soil treatment method based on big data analysis, the traditional contaminated soil treatment method can be solved. Since it is impossible to quickly and accurately predict soil pollution characteristics according to environmental conditions, the accuracy of soil sampling point layout is poor, making the accuracy, comprehensiveness and reliability of soil detection results insufficient. On the other hand, it is impossible to quickly and accurately set up a suitable soil remediation plan based on the pollution detection results, which ultimately leads to technical problems such as poor quality and efficiency of contaminated soil treatment. By extracting features from the polluted area image set, the regional feature set is determined, and then the regional feature set is used as a constraint condition to retrieve positive samples based on big data to obtain multiple sample pollution distribution characteristics; then the pollution distribution characteristics of multiple samples are clustered, and a predicted pollution distribution map is constructed based on the feature clustering results, and multiple optimized sampling points are targetedly arranged according to the predicted pollution distribution map; further soil sampling is carried out based on multiple optimized sampling points, and pollution detection is carried out to obtain multiple soil pollution detection results; then the target polluted area is simulated and analyzed based on the multiple soil pollution detection results to determine the predicted pollution distribution trend and the predicted pollution spread path, and a pollutant distribution simulation map is constructed; then the pollutant distribution simulation map is divided to determine multiple local area pollution characteristics; finally, the pollution characteristics of multiple local areas are input into the soil remediation analysis channel, and the optimized governance strategy is output to execute soil governance; the accuracy and efficiency of soil pollution feature prediction can be improved, thereby improving the accuracy and scientificity of sampling point layout, and then a soil remediation analysis channel is constructed based on big data technology, and a suitable soil remediation plan is determined according to the soil pollution characteristics through rapid analysis, so as to achieve the technical effect of improving the quality and efficiency of polluted soil governance.

[0031] Further, to obtain the pollution distribution characteristics of multiple samples, step 2 of this application includes:

[0032] The soil color distribution characteristics, soil structure distribution characteristics and plant growth distribution characteristics are respectively subjected to principal component analysis to construct a soil color feature matrix, a soil structure feature matrix and a plant growth feature matrix; taking the soil color feature matrix, the soil structure feature matrix and the plant growth feature matrix as a benchmark, sample retrieval is performed based on big data, and sample data whose soil color similarity, soil structure similarity and plant growth similarity are all greater than a predetermined similarity threshold are selected to obtain the pollution distribution characteristics of the multiple samples.

[0033] Specifically, principal component analysis is a statistical method used to convert multiple variables into a few principal components, which can reflect most of the information of the original variables. In the treatment of contaminated soil, principal component analysis of soil color distribution characteristics, soil structure distribution characteristics and plant growth distribution characteristics can help us extract the main components of these characteristics for subsequent analysis and processing. Then, principal component analysis is performed on the soil color distribution characteristics, soil structure distribution characteristics and plant growth distribution characteristics respectively, that is, principal component analysis technology is applied to extract features with high correlation to reduce the number of features; further, based on the results of principal component analysis, soil color feature matrix, soil structure feature matrix and plant growth feature matrix are constructed, each of which will contain the principal components of the corresponding features. Through principal component analysis, we can extract the principal components that best represent these features from the original features, thereby simplifying the amount of data processing and improving the efficiency and accuracy of data processing.

[0034] Then, taking the soil color feature matrix, soil structure feature matrix and plant growth feature matrix as benchmarks, sample data retrieval is performed based on big data technology, that is, the soil color feature matrix, soil structure feature matrix and plant growth feature matrix are respectively compared with the sample soil color feature matrix, sample soil structure feature matrix and sample plant growth feature matrix of the retrieval data for similarity. Commonly used similarity comparison methods include Euclidean distance, cosine similarity, etc., and an adapted similarity comparison algorithm can be selected according to actual needs to obtain soil color similarity, soil structure similarity and plant growth similarity; further, the soil color similarity, soil structure similarity and plant growth similarity are judged respectively according to a predetermined similarity threshold, and the retrieval data whose soil color similarity is greater than the predetermined similarity threshold, soil structure similarity is greater than the predetermined similarity threshold and plant growth similarity is greater than the predetermined similarity threshold are selected as sample data to obtain a plurality of sample pollution distribution characteristics that meet a predetermined number.

[0035] Further, the soil color feature matrix is ​​constructed as shown in the attached Figure 2 As shown, the application also includes the following steps:

[0036] The soil color distribution characteristics are divided into regions to obtain multiple regional color distribution characteristics, and a first color distribution characteristic of a first region is randomly selected; a first feature data set of the first color distribution characteristics is obtained, and the first feature data set is decentralized to obtain a second feature data set; based on the second feature data set, a first covariance matrix of the first feature data set is calculated, and according to the first covariance matrix, a first eigenvalue and a first eigenvector are calculated to construct a first soil color characteristic; and the soil color characteristic matrix is ​​constructed based on the first region and the first soil color characteristic.

[0037] Specifically, first, the soil color distribution characteristics are divided into regions, wherein the regions can be divided according to predetermined sizes and specifications, and multiple regions and multiple regional color distribution characteristics corresponding to the multiple regions are obtained; a region is randomly selected from the multiple regions and set as the first region, and the first color distribution characteristics of the first region are obtained. Then, a first feature data set of the first color distribution characteristics is obtained, and the first feature data set is decentralized, that is, each feature is subtracted from its mean to eliminate the central trend of the data. The decentralized data set is called the second feature data set, and the second feature data set is obtained. Then, based on the second feature data set, the first covariance matrix of the first feature data set is calculated. The covariance matrix is ​​a symmetric matrix, and the elements on the diagonal are the variance of each feature, and the elements on the non-diagonal are the covariance between two features, and the first covariance matrix is ​​obtained; further, according to the first covariance matrix, the first eigenvalue and the first eigenvector are calculated, that is, the covariance matrix is ​​eigendecomposed to obtain the eigenvalue and eigenvector, the eigenvalue is the eigenvalue of the covariance matrix, and the eigenvector is the eigenvector corresponding to each eigenvalue. Finally, according to the first eigenvalue and the first eigenvector, the first soil color feature is constructed. The first soil color feature is a matrix composed of eigenvectors corresponding to the first k largest eigenvalues. These eigenvectors can explain most of the variance of the original data.

[0038] The same method is used to calculate multiple soil color features of multiple regions, and the multiple soil color features are fused based on multiple regions to generate a soil color feature matrix. The construction method of the soil structure feature matrix and the plant growth feature matrix is ​​the same as that of the soil color feature matrix, which is not described in detail here. Those skilled in the art can refer to the construction method of the soil color feature matrix.

[0039] Further, to construct a predicted pollution distribution map, step three of this application includes:

[0040] Perform regional registration on the pollution distribution characteristics of the multiple samples, divide the pollution distribution characteristics of the multiple samples into regions based on the regional registration results, and determine multiple subsets of pollution distribution characteristics of the samples, wherein the pollution distribution characteristics of the samples include pollutant types and pollutant concentrations; randomly select a first subset of pollution distribution characteristics of the sample, count the frequency of occurrence of pollutant types in the first subset of pollution distribution characteristics of the sample, select pollutant types greater than a predetermined frequency threshold as first pollutant types, and obtain multiple first pollutant types; extract pollutant concentrations in the first subset of pollution distribution characteristics of the sample with the multiple first pollutant types as constraints, determine multiple pollutant concentration sets, and calculate multiple pollutant concentration means; perform deviation calculation on the multiple pollutant concentration sets based on the multiple pollutant concentration means, select pollutant concentrations with concentration deviations less than a predetermined deviation threshold as standard pollutant concentrations, determine multiple standard pollutant concentration sets, perform mean calculation on the multiple standard pollutant concentration sets, and obtain multiple first pollutant concentrations; construct a first regional predicted pollution distribution based on the multiple first pollutant types and the multiple first pollutant concentrations, and generate the predicted pollution distribution map according to the first regional predicted pollution distribution.

[0041] Specifically, first, the pollution distribution characteristics of the multiple samples are regionally registered, for example, by applying image processing technology (such as image registration algorithm, etc.), to perform regional registration on the pollution distribution characteristics of the multiple samples, ensuring that the pollution distribution characteristics of each sample are aligned with the reference sample to facilitate subsequent comparison and analysis; then, based on the regional registration results, the pollution distribution characteristics of the multiple samples are divided into regions, where each region can represent a subset of the pollution distribution, which is helpful to more carefully analyze the pollution characteristics and determine multiple sample pollution distribution feature subsets, where the sample pollution distribution characteristics include pollutant type and pollutant concentration.

[0042] Then, any one of the multiple sample pollution distribution feature subsets is randomly selected and set as the first sample pollution distribution feature subset; then, the occurrence frequency of the same pollutant type in the first sample pollution distribution feature subset is counted, and the pollutant type greater than a predetermined frequency threshold is selected and set as the first pollutant type, wherein the predetermined frequency threshold can be set according to actual conditions, that is, the pollutant type with a higher occurrence frequency is set as the first pollutant type, and multiple first pollutant types are obtained.

[0043] Then, based on the multiple first pollutant types, the first sample pollution distribution feature subset is subjected to pollutant concentration extraction, multiple pollutant concentration sets of multiple first pollutant types are determined, and the multiple pollutant concentration sets are subjected to mean calculation to obtain multiple pollutant concentration means. Further, the multiple pollutant concentration sets are subjected to deviation calculation based on the multiple pollutant concentration means, and the pollutant concentration with a concentration deviation less than a predetermined deviation threshold is selected as the standard pollutant concentration to obtain multiple standard pollutant concentration sets; finally, the multiple standard pollutant concentration sets are subjected to mean calculation, and the mean calculation result is set as the first pollutant concentration to obtain multiple first pollutant concentrations.

[0044] Finally, a first regional predicted pollution distribution is constructed based on the multiple first pollutant types and the multiple first pollutant concentrations, and multiple regional predicted pollution distributions are obtained in sequence using the same method as that for obtaining the first regional predicted pollution distribution; and the predicted pollution distribution map is constructed based on the multiple regional predicted pollution distributions.

[0045] Further, to construct a pollutant distribution simulation map, step five of this application includes:

[0046] Based on the digital twin technology, simulation is performed according to the multiple soil pollution detection results to generate multiple pollution simulation distributions, and data fusion and deviation correction are performed on the multiple pollution simulation distributions according to the multiple optimized sampling point location coordinates to construct a target pollution distribution map; through the soil remediation expert system, pollution trend analysis and spread path prediction are performed on the target pollution distribution map to determine the predicted pollution distribution trend and the predicted pollution spread path; the target pollution distribution map is marked according to the predicted pollution distribution trend and the predicted pollution spread path to obtain the pollutant distribution simulation map.

[0047] Specifically, digital twin technology is used to simulate the multiple soil pollution detection results and create a virtual model. For example, numerical simulation methods (such as finite element analysis, fluid mechanics simulation, etc.) or geographic information system (GIS) tools are used to simulate and analyze the target pollution area to generate multiple pollution simulation distributions. Then, according to the coordinates of multiple optimized sampling points, the multiple pollution simulation distributions are subjected to data fusion and deviation correction, that is, according to the sampling point coordinates, the data of multiple pollution simulation distributions are associated with the sampling point positions, such as applying data fusion technology, such as spatial interpolation, Kriging interpolation, etc., to correspond the simulation data with the sampling point coordinates; then, based on the correlation relationship of the sampling point coordinates, the pollution simulation distribution is corrected for deviations, that is, the erroneous data is corrected to construct the target pollution distribution map.

[0048] Then, a soil remediation expert system is constructed based on neural networks and big data, that is, a soil remediation expert system is constructed by combining the neural network model and expert knowledge in the field of soil remediation. The expert system should be able to make decisions and judgments based on the input data and imitate the judgment thinking of human soil experts. Then, through the soil remediation expert system, the target pollution distribution map is subjected to pollution trend analysis and spread path prediction, that is, the possible diffusion path and impact range of pollutants are predicted according to the target pollution distribution map, and the predicted pollution distribution trend and predicted pollution spread path are determined. Finally, the target pollution distribution map is marked according to the predicted pollution distribution trend and predicted pollution spread path, such as using color coding, icons, lines and other visual elements to indicate the concentration change trend and spread path of pollutants, and combining the prediction results and the target pollution distribution map to construct a pollutant distribution simulation map. The simulation map should clearly show the distribution pattern and trend of pollutants, as well as the path of pollution spread.

[0049] Further, to determine the pollution characteristics of multiple local areas, step five of this application includes:

[0050] The pollutant distribution simulation map is grid-divided according to a predetermined size to obtain a plurality of pollutant grid distribution maps; a first grid distribution map and a second grid distribution map are randomly selected from the plurality of pollutant grid distribution maps, wherein the first grid distribution map and the second grid distribution map are any two adjacent grid distribution maps in the plurality of pollutant grid distribution maps; if the pollutant types in the first grid distribution map and the second grid distribution map are the same and the pollution concentration deviations are both less than the concentration deviation threshold, the first grid distribution map and the second grid distribution map are merged to obtain a first updated grid distribution map, wherein the pollutant concentration of the first updated grid distribution map is the average of the pollutant concentrations of the merged two grid distribution maps; iterative merging is continuously performed until there is no more grid distribution map that can be merged, then the merging is stopped, and the plurality of local area pollution characteristics are output, wherein the local area pollution characteristics include the local area pollutant type and the local area pollutant concentration.

[0051] Specifically, first, the pollutant distribution simulation map is grid-divided according to a predetermined size, and the predetermined size can be set according to actual conditions to obtain multiple pollutant grid distribution maps; then, the first grid distribution map and the second grid distribution map are randomly selected from the multiple pollutant grid distribution maps, wherein the first grid distribution map and the second grid distribution map are any two adjacent grid distribution maps in the multiple pollutant grid distribution maps; further identify the pollutant type of the first grid distribution map and the second grid distribution map, if the pollutant type in the first grid distribution map and the second grid distribution map is the same; then calculate the concentration deviation of multiple pollutant types respectively to obtain multiple pollution concentration deviations; if the multiple pollution concentration deviations are all less than the concentration deviation threshold, then merge the first grid distribution map and the second grid distribution map to obtain a first updated grid distribution map, wherein the pollutant concentration of the first updated grid distribution map is the average of the pollutant concentrations of the two grid distribution maps after merging. The same method is used to continuously iterate and merge until there is no grid distribution map that can be merged, then the merging is stopped, and the multiple local area pollution features are output, wherein the local area pollution features include the local area pollutant type and the local area pollutant concentration. By iteratively merging grid areas and outputting local area pollution characteristics, the distribution characteristics of contaminated soil can be analyzed and understood more accurately, providing a basis for the subsequent refined treatment of contaminated soil.

[0052] Further, to construct a soil remediation analysis channel, step six of this application includes:

[0053] Based on big data, soil remediation scheme mining is carried out to obtain multiple sample soil remediation schemes, wherein the sample soil remediation schemes include multiple sample pollutant types, multiple sample pollutant concentrations, sample remediation methods and sample remediation results, and the sample remediation results include remediation cycle, resource consumption and remediation quality coefficient; the multiple sample soil remediation schemes are evaluated according to a predetermined scheme evaluation function, and multiple preferred remediation schemes are determined based on the evaluation results; an initial soil remediation analysis channel is constructed based on a random decision forest, and the initial soil remediation analysis channel is iteratively trained with the multiple preferred remediation schemes as training data to obtain a soil remediation analysis channel that meets convergence constraints.

[0054] Specifically, firstly, soil remediation schemes are mined based on big data to obtain multiple sample soil remediation schemes, wherein the sample soil remediation schemes include multiple sample pollutant types, multiple sample pollutant concentrations, sample remediation methods and sample remediation results, and the sample remediation results include remediation cycle, resource consumption and remediation quality coefficient. The remediation cycle refers to the time period required for soil remediation, resource consumption refers to the amount of resources used for soil remediation, and the remediation quality coefficient refers to the soil remediation effect. The larger the quality coefficient, the better the remediation effect. Then, the multiple sample soil remediation schemes are evaluated according to the predetermined scheme evaluation function to obtain multiple evaluation fitness of the multiple sample soil remediation schemes, and the multiple sample soil remediation schemes are arranged from large to small according to the fitness to generate a sample soil remediation scheme sequence. Finally, the first M sample soil remediation schemes in the sample soil remediation scheme sequence are selected as the preferred remediation schemes to obtain multiple preferred remediation schemes, wherein the specific value of M can be set according to actual needs.

[0055] Random decision forest is a classifier or regressor based on decision trees. It predicts results by building multiple decision trees and uses the majority vote of these decision trees to obtain the final result. For classification problems, each decision tree gives a prediction result, and the final result of the random decision forest is the majority vote of all decision tree prediction results; for regression problems, each decision tree gives a prediction value, and the final result of the random decision forest is the average of all decision tree prediction values. Then, a random decision forest algorithm is selected as the model for soil remediation analysis, and the parameters of the random decision forest, such as the number of trees, random feature selection for node splitting, etc., are determined to construct an initial soil remediation analysis channel; further, the multiple preferred remediation schemes are used as training data to iteratively train the initial soil remediation analysis channel, such as randomly selecting a part of the training data as the training set of the current iteration, using the random decision forest algorithm to train the current training set, and evaluating the performance of the current channel, such as prediction accuracy, convergence constraints, etc.; then, according to the evaluation results, the parameters of the random decision forest, such as the number of trees, random feature selection for node splitting, etc., are adjusted; the above steps are repeated until the convergence constraints are met or the preset number of iterations is reached, and a soil remediation analysis channel that meets the convergence constraints is obtained.

[0056] By building a soil remediation analysis channel based on big data and random decision forests, we can provide intelligent and precise decision support for contaminated soil management and improve the accuracy and efficiency of soil remediation settings.

[0057] Further, to construct a predetermined scheme evaluation function, the present application also includes the following steps:

[0058] The expression of the predetermined scheme evaluation function is: ; Where P is the fitness of the scheme evaluation, To repair the cycle weight, is the resource consumption weight, To fix the quality weight, To schedule the repair cycle, For the solution repair cycle, To pre-determine resource consumption, For the resource consumption of the solution, is the solution repair quality coefficient, is the predetermined repair quality factor.

[0059] Specifically, in the predetermined scheme evaluation function, P is the scheme evaluation fitness, and the higher the evaluation fitness, the better the comprehensive governance effect of the representation scheme; To repair the cycle weight, is the resource consumption weight, is the repair quality weight, where , and The sum is 1. The influence on the overall fitness can be set according to the indicator type. The higher the influence, the greater the corresponding weight. To schedule the repair cycle, For the solution repair cycle, To pre-determine resource consumption, For the resource consumption of the solution, is the solution repair quality coefficient, is the predetermined remediation quality coefficient. By constructing a predetermined scheme evaluation function, the efficiency and accuracy of the sample soil remediation scheme evaluation can be improved, providing support for scheme screening.

[0060] In summary, the contaminated soil treatment method based on big data analysis provided by this application has the following technical effects:

[0061] By extracting features from the polluted area image set, the regional feature set is determined, and then the regional feature set is used as a constraint condition to retrieve positive samples based on big data to obtain multiple sample pollution distribution characteristics; then the pollution distribution characteristics of multiple samples are clustered, and a predicted pollution distribution map is constructed based on the feature clustering results, and multiple optimized sampling points are targetedly arranged according to the predicted pollution distribution map; further soil sampling is carried out based on multiple optimized sampling points, and pollution detection is carried out to obtain multiple soil pollution detection results; then the target polluted area is simulated and analyzed based on the multiple soil pollution detection results to determine the predicted pollution distribution trend and the predicted pollution spread path, and a pollutant distribution simulation map is constructed; then the pollutant distribution simulation map is divided to determine multiple local area pollution characteristics; finally, the pollution characteristics of multiple local areas are input into the soil remediation analysis channel, and the optimized governance strategy is output to execute soil governance; the accuracy and efficiency of soil pollution feature prediction can be improved, thereby improving the accuracy and scientificity of sampling point layout, and then a soil remediation analysis channel is constructed based on big data technology, and a suitable soil remediation plan is determined according to the soil pollution characteristics through rapid analysis, so as to achieve the technical effect of improving the quality and efficiency of polluted soil governance.

[0062] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0063] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A method for treating contaminated soil based on big data analysis, characterized in that: include: Taking aerial photos of the target polluted area by a drone to obtain an image set of the polluted area, extracting features from the image set of the polluted area, and determining a regional feature set, wherein the regional feature set includes soil color distribution features, soil structure distribution features, and plant growth distribution features; Taking the soil color distribution characteristics, soil structure distribution characteristics and plant growth distribution characteristics as constraints, positive sample retrieval is performed based on big data to obtain pollution distribution characteristics of multiple samples; Performing feature clustering on the pollution distribution features of the multiple samples, constructing a predicted pollution distribution map based on the feature clustering results, and arranging multiple optimized sampling points according to the predicted pollution distribution map, including identifying key areas of the predicted pollution distribution map, determining areas with high pollutant concentrations and sensitive areas where pollutant diffusion may occur, and arranging multiple optimized sampling points with full coverage of high-concentration areas and sensitive areas as constraints; Based on the multiple optimized sampling points, soil sampling is performed on the target contaminated area, and pollution detection is performed on the multiple soil sampling results to obtain multiple soil pollution detection results, wherein the soil pollution detection results include soil property characteristics, soil environmental characteristics, pollutant detection types and pollutant detection concentrations; Performing simulation analysis on the target polluted area according to the multiple soil pollution detection results, determining the predicted pollution distribution trend and the predicted pollution spread path, constructing a pollutant distribution simulation map, dividing the pollutant distribution simulation map, and determining pollution characteristics of multiple local areas; Inputting the pollution characteristics of the multiple local areas into a soil remediation analysis channel, outputting multiple local remediation schemes, constructing an optimized governance strategy based on the multiple local remediation schemes, and executing soil remediation in the target polluted area according to the optimized governance strategy; Among them, constructing a predicted pollution distribution map includes: Performing regional registration on the multiple sample pollution distribution features, and dividing the multiple sample pollution distribution features into regions based on the regional registration results to determine multiple sample pollution distribution feature subsets, wherein the sample pollution distribution features include pollutant types and pollutant concentrations; Randomly selecting a first sample pollution distribution feature subset, counting the frequency of occurrence of pollutant types in the first sample pollution distribution feature subset, selecting a pollutant type greater than a predetermined frequency threshold as a first pollutant type, and obtaining a plurality of first pollutant types; Taking the multiple first pollutant types as constraints, extracting pollutant concentrations in the first sample pollution distribution feature subset, determining multiple pollutant concentration sets, and calculating and obtaining multiple pollutant concentration means; Based on the mean values ​​of the multiple pollutant concentrations, deviation calculation is performed on the multiple pollutant concentration sets, a pollutant concentration whose concentration deviation is less than a predetermined deviation threshold is selected as a standard pollutant concentration, multiple standard pollutant concentration sets are determined, and mean calculation is performed on the multiple standard pollutant concentration sets to obtain multiple first pollutant concentrations; constructing a first regional predicted pollution distribution based on the multiple first pollutant types and the multiple first pollutant concentrations, and generating the predicted pollution distribution map according to the first regional predicted pollution distribution; Among them, multiple local area pollution characteristics are determined, including: Dividing the pollutant distribution simulation map into grids according to a predetermined size to obtain a plurality of pollutant grid distribution maps; Randomly selecting a first grid distribution map and a second grid distribution map from the multiple pollutant grid distribution maps, wherein the first grid distribution map and the second grid distribution map are any two adjacent grid distribution maps from the multiple pollutant grid distribution maps; If the pollutant types in the first grid distribution map and the second grid distribution map are the same and the pollution concentration deviations are both less than the concentration deviation threshold, then the first grid distribution map and the second grid distribution map are merged to obtain a first updated grid distribution map, wherein the pollutant concentration of the first updated grid distribution map is the average of the pollutant concentrations of the two merged grid distribution maps; Continuously iterate the merging until there is no grid distribution map that can be merged, then stop merging, and output the multiple local area pollution characteristics, wherein the local area pollution characteristics include the local area pollutant type and the local area pollutant concentration; Construct soil remediation analysis channel, including: Mining soil remediation solutions based on big data to obtain multiple sample soil remediation solutions, where the sample soil remediation solutions include multiple sample pollutant types, multiple sample pollutant concentrations, sample remediation methods, and sample remediation results. The sample remediation results include remediation cycle, resource consumption, and remediation quality coefficient; Evaluate the multiple sample soil remediation schemes according to a predetermined scheme evaluation function, and determine multiple preferred remediation schemes based on the evaluation results; An initial soil remediation analysis channel is constructed based on random decision forests, and the initial soil remediation analysis channel is iteratively trained using the multiple preferred remediation schemes as training data to obtain a soil remediation analysis channel that satisfies convergence constraints.

2. The method for contaminated soil remediation based on big data analysis according to claim 1 is characterized in that: Obtain multiple sample contamination distribution characteristics, including: Performing principal component analysis on the soil color distribution characteristics, soil structure distribution characteristics and plant growth distribution characteristics respectively, and constructing a soil color feature matrix, a soil structure feature matrix and a plant growth feature matrix; Based on the soil color feature matrix, soil structure feature matrix and plant growth feature matrix, sample retrieval is performed based on big data, and sample data whose soil color similarity, soil structure similarity and plant growth similarity are all greater than a predetermined similarity threshold are selected to obtain the pollution distribution characteristics of the multiple samples.

3. The method for contaminated soil remediation based on big data analysis according to claim 2 is characterized in that: Construct soil color feature matrix, including: Dividing the soil color distribution characteristics into regions to obtain a plurality of regional color distribution characteristics, and randomly selecting a first color distribution characteristic of a first region; Acquire a first feature data set of the first color distribution feature, and perform decentralization processing on the first feature data set to obtain a second feature data set; Based on the second feature data set, a first covariance matrix of the first feature data set is calculated, and according to the first covariance matrix, a first eigenvalue and a first eigenvector are calculated to construct a first soil color feature; The soil color feature matrix is ​​constructed based on the first region and the first soil color feature.

4. The method for contaminated soil remediation based on big data analysis according to claim 1, characterized in that: Construct a pollutant distribution simulation map, including: Based on the digital twin technology, simulation is performed according to the multiple soil pollution detection results to generate multiple pollution simulation distributions, and data fusion and deviation correction are performed on the multiple pollution simulation distributions according to the multiple optimized sampling point position coordinates to construct a target pollution distribution map; Through the soil treatment expert system, the pollution trend analysis and spread path prediction of the target pollution distribution map are performed to determine the predicted pollution distribution trend and the predicted pollution spread path; The target pollution distribution map is marked according to the predicted pollution distribution trend and the predicted pollution spreading path to obtain the pollutant distribution simulation map.

5. The method for contaminated soil remediation based on big data analysis according to claim 1, characterized in that: Construct a predetermined scheme evaluation function, including: The expression of the predetermined scheme evaluation function is: ； Among them, P is the fitness of the scheme evaluation, To repair the cycle weight, is the resource consumption weight, To fix the quality weight, To schedule the repair cycle, For the solution repair cycle, To pre-determine resource consumption, For the resource consumption of the solution, is the solution repair quality coefficient, is the predetermined repair quality factor.

Images