Soil heavy metal pollution source distinguishing method, device, equipment and medium

Abstract

The invention provides a soil heavy metal pollution source discrimination method. The method comprises the following steps: acquiring natural characteristic factors of a soil sample and an actual test value of soil sample heavy metal; performing pretreatment operation on the natural characteristic factors to obtain main control factors of the soil sample; constructing a prediction model of a soil heavy metal natural background value, and inputting the main control factor into the prediction model to obtain a heavy metal natural background prediction value; and comparing the natural background predicted value with an actual test value, and determining the heavy metal pollution source type of the soil sample. According to the technical scheme of the embodiment, migration of heavy metal pollution in the soil is explained from multiple angles by introducing the main control factor of the soil sample, so that the dimensionality and the information amount of the soil are obtained in multiple dimensions, and then the heavy metal pollution source of the soil is accurately judged.

Description

Technical Field

[0001] The present invention relates to the technical field of soil monitoring, and in particular to a method, device, equipment and medium for identifying the source of heavy metal pollution in soil. Background Art

[0002] The sources of soil heavy metal pollution are diverse, including natural and human sources. However, when collecting information that affects the heavy metal content in the soil, the existing technology cannot effectively obtain the dimension and information of the soil, resulting in the inability to comprehensively analyze the distribution and form of heavy metals in the soil and their relationship with other components in the soil, making it difficult to accurately determine the main sources of soil heavy metal pollution. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a method, device, equipment and medium for identifying the source of heavy metal pollution in soil, which can effectively obtain the dimension and information of soil.

[0004] In a first aspect, an embodiment of the present invention provides a method for identifying the source of heavy metal pollution in soil, comprising:

[0005] Obtain the natural characteristic factors of soil samples and the actual test values ​​of heavy metals in soil samples;

[0006] Performing a preprocessing operation on the natural characteristic factors to obtain a main control factor that is highly correlated with the natural background value of heavy metals in the soil sample;

[0007] Constructing a prediction model for the natural background value of heavy metals in soil, inputting the main control factors into the prediction model, and obtaining the predicted value of the natural background value of heavy metals in soil;

[0008] The predicted natural background value of heavy metals in the soil is compared with the actual test value of heavy metals in the soil to determine the source type of heavy metal pollution in the soil sample.

[0009] In some embodiments of the present invention, the step of obtaining the natural characteristic factors of the soil sample includes:

[0010] Scanning the soil sample to obtain a plurality of two-dimensional scanned images of the soil sample, and performing three-dimensional reconstruction on the two-dimensional scanned images to obtain a first three-dimensional reconstructed image of the soil sample;

[0011] Performing threshold segmentation on the first three-dimensional reconstructed image to obtain a binary segmentation result of the three-dimensional reconstructed image, and distinguishing soil skeletons and soil pores in a plurality of the three-dimensional images according to the binary segmentation result to obtain a second three-dimensional reconstructed image;

[0012] The second three-dimensional reconstructed image is calculated to obtain soil structural parameters of the soil sample.

[0013] In some embodiments of the present invention, after obtaining the multiple two-dimensional scanning images of the soil sample, the method further includes:

[0014] Obtain pixel points of the two-dimensional scanned image and mark them as first pixel points and second pixel points, wherein the first pixel points are all pixel points of the soil sample, and the second pixel points are pixel points of soil pores in the soil sample;

[0015] Acquire a first porosity of the second pixel in the first pixel at a first resolution, and a second porosity of the second pixel in the first pixel at a second resolution, and sum the first porosity and the second porosity to obtain a third porosity;

[0016] Alternatively, a pore distance map of the three-dimensional reconstructed image is obtained, and the soil pores are divided into a plurality of independent pores according to the pore distance map;

[0017] Selecting a mercury injection surface of the independent pore, and injecting mercury into the independent pore according to the mercury injection surface;

[0018] When the radius of the independent pore is zero, obtaining the connected porosity of the three-dimensional reconstructed image;

[0019] Obtaining the diameters of all the independent pores in the three-dimensional reconstructed image, performing statistics on the diameters, and calculating the average value and median value of the soil pore radius in the three-dimensional reconstructed image;

[0020] Calculate the soil permeability coefficient.

[0021] In some embodiments of the present invention, the preprocessing operation on the natural characteristic factor includes:

[0022] Performing standardization processing on the plurality of natural characteristic factors, and calculating a plurality of correlation coefficients of the plurality of natural characteristic factors;

[0023] Eliminate the parts of the plurality of natural characteristic factors whose correlation coefficients with other natural characteristic factors are higher than 0.6, so that the correlation coefficients between all the remaining natural characteristic factors are less than 0.6;

[0024] Recursively eliminate the natural feature factors whose correlation coefficient is less than 0.6, so that all available natural feature factors can train a model, and calculate the weight or coefficient of each feature;

[0025] Sorting the natural characteristic factors according to the weights or coefficients of the natural characteristic factors;

[0026] Deleting one or more of the natural feature factors with the smallest weight or coefficient in the sorting, and retraining the model with the remaining natural feature factors until the number of remaining natural feature factors is unable to continue to eliminate features, and obtaining the contribution rate of each natural feature factor;

[0027] The selected natural characteristic factors are sorted according to their contribution rates, and the natural characteristic factors are selected from high to low according to the contribution rates until the cumulative contribution rate exceeds 99%, and the selected natural characteristic factors are used as the main controlling factors for predicting the natural background value of soil heavy metals, wherein the contribution rate of the main controlling factors is not less than 5%.

[0028] In some embodiments of the present invention, constructing a prediction model for the natural background value of heavy metals in soil includes:

[0029] Obtaining a data set of the soil sample and dividing it into K mutually exclusive subsets of equal size, taking the union of K-1 mutually exclusive subsets as a training set, and the remaining mutually exclusive subsets as a test set and a validation set;

[0030] Obtaining K groups of mutually exclusive subsets, the test set, and the validation set, and performing K training times on the K groups of mutually exclusive subsets, the test set, and the validation set to obtain a data set division result of the soil sample;

[0031] With the main controlling factor as input and the natural background value as output, a soil heavy metal natural background value prediction model is established using a random forest algorithm.

[0032] In some embodiments of the present invention, determining the type of heavy metal pollution source of the soil sample includes:

[0033] Setting the upper limit range of the heavy metal pollution source types;

[0034] When the difference between the actual test value and the natural background predicted value is within the upper limit range, the heavy metal pollution source type is natural pollution;

[0035] When the difference between the actual test value and the natural background predicted value is greater than the upper limit interval, the source type of heavy metal pollution is man-made pollution.

[0036] In some embodiments of the present invention, the calculating the soil permeability coefficient includes:

[0037] Performing an initialization operation on the flow field in the three-dimensional reconstructed image so that the distribution function in the flow field is initialized to an equilibrium state distribution;

[0038] Calculating a first distribution function of the flow field according to a preset single relaxation model, and performing a collision on the first pixel point and the second pixel point;

[0039] Determine a plurality of first pore points and second pore points in the three-dimensional reconstructed image, and when a downstream grid of the first pore point is the second pore point, migrate the first pore point, and when a downstream grid of the first pore point is not the second pore point, rebound the first pore point;

[0040] Determine a second distribution function entering the flow field, and assign the second distribution function to a post-collision distribution function value of the flow field flowing out of the flow field on the buffer layer of the single relaxation model;

[0041] Acquiring the average flow rate of the flow field and calculating the pressure difference of the flow field, and calculating the permeability coefficient of the flow field according to the average flow rate and the pressure difference;

[0042] The relative error of the permeability coefficient is calculated according to the permeability coefficient.

[0043] In a second aspect, an embodiment of the present invention provides a device for identifying the source of heavy metal pollution in soil, comprising at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the method for identifying the source of heavy metal pollution in soil as described in the first aspect above.

[0044] In a third aspect, an embodiment of the present invention provides an electronic device, comprising the device for identifying the source of heavy metal pollution in soil as described in the second aspect above.

[0045] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are used to execute the method for identifying the source of heavy metal pollution in soil as described in the first aspect above.

[0046] The method for identifying the source of heavy metal pollution in soil according to the embodiment of the present invention has at least the following beneficial effects:

[0047] Obtain the natural characteristic factors of the soil sample, and obtain the natural background value of the soil sample according to the natural characteristic factors; perform preprocessing operations on the natural characteristic factors to obtain the main control factors of the soil sample; construct a prediction model for the natural background value, input the main control factors into the prediction model, and obtain the natural background prediction value; obtain the actual test value of the soil sample, compare the natural background prediction value with the actual test value, and determine the type of pollution source of the soil sample. According to the technical solution of this embodiment, by introducing the main control factors of the soil sample, the migration of heavy metal pollution in the soil is explained from multiple angles, thereby obtaining the dimensions and information of the soil in multiple dimensions, and then accurately judging the source of soil pollution. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of a method for identifying the source of heavy metal pollution in soil provided by one embodiment of the present invention;

[0049] Figure 2 It is a flow chart of obtaining natural characteristic factors of soil samples provided by the present invention;

[0050] Figure 3 It is a flow chart after obtaining multiple two-dimensional scanning images of soil samples provided by the present invention;

[0051] Figure 4 It is a flow chart of preprocessing operation of natural characteristic factors provided by the present invention;

[0052] Figure 5 It is a flow chart of constructing a prediction model of natural background value provided by the present invention;

[0053] Figure 6 is a flow chart for determining the type of pollution source of a soil sample provided by the present invention;

[0054] Figure 7 is a flow chart for calculating soil permeability coefficient provided by the present invention;

[0055] Figure 8 It is a structural diagram of a soil heavy metal pollution source identification device provided by another embodiment of the present invention. DETAILED DESCRIPTION

[0056] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0057] In the description of the present invention, it should be understood that descriptions involving orientations, such as up, down, front, back, left, right, etc., and orientations or positional relationships indicated are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0058] In the description of the present invention, "several" means one or more, "more" means more than two, "greater than", "less than", "exceed" etc. are understood as not including the number itself, and "above", "below", "within" etc. are understood as including the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing the technical features, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0059] In the description of the present invention, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention in combination with the specific content of the technical solution. For example, heavy metals should be understood in a broad sense, and the types of pollution sources of elements such as arsenic, cadmium, chromium, and mercury can all be distinguished through the description of the present invention.

[0060] The embodiment of the present invention provides a method for distinguishing the source of soil arsenic pollution, including obtaining the natural characteristic factors of the soil sample, obtaining the natural background value of arsenic of the soil sample according to the natural characteristic factors; performing pretreatment operations on the natural characteristic factors to obtain the main control factors of the soil sample; constructing a prediction model for the natural background value of arsenic, inputting the main control factors into the prediction model, and obtaining the predicted value of the natural background of arsenic; obtaining the actual test value of arsenic of the soil sample, comparing the predicted value of the natural background with the actual test value, and determining the type of pollution source of arsenic in the soil sample. According to the technical solution of this embodiment, by introducing the main control factors of the soil sample, the migration of arsenic in the soil is explained from multiple angles, thereby obtaining the dimensions and information of the soil in multiple dimensions, and then accurately judging the source type of soil arsenic pollution.

[0061] The control method of the embodiment of the present invention is further described below based on the accompanying drawings.

[0062] Reference Figure 1 , Figure 1 A flow chart of a method for identifying the source of soil arsenic pollution provided by an embodiment of the present invention, the method for identifying the source of soil arsenic pollution includes but is not limited to the following steps:

[0063] Step S11, obtaining the natural characteristic factors of the soil sample and the actual arsenic test value of the soil sample;

[0064] It should be noted that natural original soil is collected to ensure that the sample maintains its original properties and is not disturbed by human intervention, and the natural characteristic factors of the soil samples are obtained through a variety of testing and analysis methods, among which the natural characteristic factors include but are not limited to the structural parameters, mineral elements, physical and chemical properties and environmental indicators of the soil. Furthermore, the structural parameters are obtained by multi-scale scanning through X-ray CT; the mineral elements are obtained through XRD, XRF and MLA; the physical and chemical properties are obtained through physical and chemical analysis methods; and the environmental indicators are recorded synchronously during the soil sample collection process.

[0065] It should be noted that multi-scale scanning refers to CT scanning of natural original structure soil samples of different specifications at different scanning resolutions (i.e., first resolution and second resolution), wherein the whole sample is CT scanned at a resolution of 35 μm, and then a natural lump-shaped micro-sample of about 2 mm is taken from the whole sample and CT scanned at 1.5 μm.

[0066] It should be noted that the natural characteristic factors specifically include the structural parameters, mineral elements, physical and chemical properties and environmental indicators of soil samples. Specifically, the mineral elements of soil samples are obtained by XRD and XRF, and arsenic-containing minerals are obtained by MLA analysis; physical and chemical properties are obtained by physical and chemical analysis methods; and environmental indicators are recorded during the soil sample collection process.

[0067] Step S12, preprocessing the natural background value to obtain the main control factor highly correlated with the natural background value of heavy metals in the soil sample;

[0068] It should be noted that the natural background value of soil arsenic is used as the output, and the natural background value is preprocessed, wherein the processing operations include but are not limited to data standardization, correlation analysis, feature selection and importance analysis. Through the above operations, the natural characteristic factors of the soil are cleaned and reduced in dimension to obtain the main control factors closely related to the natural background value of soil arsenic. Specifically, data standardization refers to the standardization of the natural background value using the Z-score standardization method, and the main control factors after processing conform to the standard normal distribution (mean is 0, standard deviation is 1). Furthermore, importance analysis refers to selecting characteristic factors in order of importance, and finally selecting the part with a cumulative importance of more than 99% as the main control factor closely related to the natural background value of soil arsenic.

[0069] Step S13, constructing a prediction model for the natural background value of soil arsenic, inputting the main control factors into the prediction model, and obtaining the predicted value of the natural background value of soil arsenic;

[0070] It should be noted that the main controlling factors were used as input to establish a soil arsenic natural background value prediction model based on random forests. Before modeling, the sample set needed to be divided into training / testing / validation sets using K-fold cross validation to ensure that the information in the sample could be fully utilized. During the modeling process, it was necessary to analyze the practical relationship between different factors and the soil arsenic natural background value to constrain the model (i.e., mechanism constraint), improve the accuracy of the main controlling factor selection, and the reliability and effectiveness of the prediction model, and solve the adverse effects of too small sample size and insufficient data abundance on machine learning modeling.

[0071] It should be noted that the prediction model for the natural background value of soil arsenic is constructed with the main control factor as input and the natural background value as output, and the random forest algorithm is used for modeling. Furthermore, the mechanism constraint specifically means that the selected main control factor must have a significant correlation (including positive and negative correlation) with the migration and transformation mechanism of arsenic in the soil, and if there is a relevant empirical calculation formula, it needs to be taken into account in the prediction model to improve the accuracy of the selection of the main control factor and the reliability and effectiveness of the prediction model, and solve the adverse effects of too small sample size and insufficient data abundance on machine learning modeling.

[0072] Step S14, comparing the predicted value of soil arsenic natural background with the actual test value of soil arsenic to determine the type of arsenic pollution source of the soil sample.

[0073] It should be noted that after the prediction model is built, the prediction interval estimation method is used to draw the discrimination line for the source of anthropogenic arsenic pollution in soil based on the performance of the prediction model on the validation set. The predicted value of the natural background of arsenic in the target soil is compared with the actual test value to determine the source type of soil arsenic pollution (natural / anthropogenic).

[0074] It should be noted that this embodiment conducts correlation analysis, feature selection and importance analysis on the natural characteristic factors of soil samples to find the main controlling factors affecting the natural background value of soil arsenic, builds a soil arsenic natural background value prediction model and delineates the discrimination line of the source of artificial pollution of soil arsenic, and judges the type of pollution source of arsenic in the target soil. By obtaining the structural parameters of the soil, the problem of insufficient dimensions and information in the traditional method when collecting soil structural parameters is solved; further, this embodiment accurately identifies and quantifies the main controlling factors of the natural background value of arsenic in the target soil through a machine learning method, realizes the limitations and pain points of the traditional method that cannot accurately calculate the natural background value of soil arsenic, and discriminates the source of arsenic pollution in the soil sample, providing strong technical support for the scientific construction of the soil arsenic pollution risk management system. Further, the feature selection method is recursive feature elimination combined with cross validation (RFECV), which can determine the optimal number of characteristic factors while retaining the actual significance of the main controlling factors, and obtain the importance of each characteristic factor.

[0075] In addition, in one embodiment, referring to Figure 2 ,exist Figure 1 Step S11 of the illustrated embodiment also includes but is not limited to the following steps:

[0076] Step S21, scanning the soil sample to obtain a plurality of two-dimensional scanned images of the soil sample, and performing three-dimensional reconstruction on the two-dimensional scanned images to obtain a first three-dimensional reconstructed image of the soil sample;

[0077] Step S22, performing threshold segmentation on the first 3D reconstructed image to obtain a binary segmentation result of the 3D reconstructed image, and distinguishing soil skeletons and soil pores in the multiple 3D images according to the binary segmentation result to obtain a second 3D reconstructed image;

[0078] Step S23: Calculate the second three-dimensional reconstructed image to obtain soil structure parameters of the soil sample.

[0079] It should be noted that the soil samples are scanned using X-ray CT to perform multi-scale digital scanning of the soil samples with natural original structure, and the multiple two-dimensional scanned images obtained are reconstructed in three dimensions to obtain a first three-dimensional reconstructed image, and then the first three-dimensional reconstructed image is intelligently identified to distinguish between soil skeleton and soil pores. Specifically, the image intelligent recognition is based on a variety of smoothing and denoising algorithms and multi-order derivative processing to construct a neural network classification model to perform threshold segmentation on the first three-dimensional reconstructed image to obtain the optimal image binarization segmentation result, so as to distinguish between soil skeletons and soil pores in multiple three-dimensional images. Finally, the three-dimensional reconstructed image with the soil skeleton and soil pores identified is numerically simulated to obtain the soil structure parameters of the soil sample. The soil structure parameters specifically include but are not limited to total porosity, connected porosity, pore radius and permeability coefficient.

[0080] In addition, in one embodiment, referring to Figure 3 ,exist Figure 2 After step S23 in the illustrated embodiment, the following steps are also included but not limited to:

[0081] Step S31, marking the pixel points of the second three-dimensional reconstructed image as first pixel points and second pixel points, wherein the first pixel points are all pixel points of the soil sample, and the second pixel points are pixel points of the soil pores in the soil sample;

[0082] Step S32, obtaining a first porosity of the second pixel in the first pixel at the first resolution, and a second porosity of the second pixel in the first pixel at the second resolution, and summing the first porosity and the second porosity to obtain a third porosity;

[0083] Step S33, alternatively, obtaining a pore distance map of the second three-dimensional reconstructed image, and dividing the soil pores into a plurality of independent pores according to the pore distance map;

[0084] Step S34, selecting a mercury injection surface of the independent pore, and injecting mercury into the independent pore according to the mercury injection surface;

[0085] Step S35, when the radius of the independent pore is zero, obtaining the connected porosity of the three-dimensional reconstructed image;

[0086] Step S36, obtaining the diameters of all independent pores in the three-dimensional reconstructed image, performing statistics on the diameters, and calculating the average and median values ​​of the soil pore radius in the three-dimensional reconstructed image;

[0087] Step S37, calculating the soil permeability coefficient.

[0088] It should be noted that the second three-dimensional reconstructed image is numerically simulated to calculate the porosity, average value and median of the soil sample. Specifically, the proportion of the second pixel point marked as soil pore to the first pixel point is obtained, that is, the first porosity and the second porosity corresponding to the first resolution and the second resolution (the first porosity corresponds to the submillimeter porosity, and the second porosity corresponds to the micrometer porosity). Finally, the first porosity and the second porosity are weighted and summed to obtain the third porosity (total porosity), that is, total porosity = submillimeter porosity + (1-submillimeter porosity) × micrometer porosity; the connected porosity is calculated using the mercury injection curve, which is the relationship curve between the mercury injection volume V and the void radius r, and its mechanical principle is Young-Lap l ace equation, the maximum sphere method and distance method are used to carry out mercury injection simulation (simulating the intrusion range of mercury). According to the pore distance map, the pore space is divided into a series of individual pores or throats (inner diameter size), and a mercury injection surface is selected. Mercury first invades the largest pore space, that is, the maximum sphere radius. The corresponding capillary pressure is the minimum mercury injection pressure. Then mercury enters the smaller connected pores according to the injection direction. The corresponding capillary radius is the individual pore radius (pore throat radius). The pore volume connected by the individual pores is the mercury injection volume. When the radius of the individual pore (throat pore) is 0, the simulation process stops, and the process obtains the connected porosity; for the average pore radius and the median pore radius, the sizes of all pores in the second three-dimensional reconstructed image are statistically analyzed, and finally the average and median values ​​of the pores are calculated.

[0089] In addition, in one embodiment, referring to Figure 4 ,exist Figure 1 Step S12 of the illustrated embodiment also includes but is not limited to the following steps:

[0090] Step S41, performing standardization processing on a plurality of natural characteristic factors, and calculating a plurality of correlation coefficients of the plurality of natural characteristic factors;

[0091] Step S42, removing the parts of the multiple natural characteristic factors whose correlation coefficients with other natural characteristic factors are higher than 0.6, so that the correlation coefficients between all the remaining natural characteristic factors are less than 0.6;

[0092] Step S43, recursively eliminating the natural feature factors with correlation coefficients less than 0.6, so that all available natural feature factors can train a model, and calculate the weight or coefficient of each feature;

[0093] Step S44, sorting the natural characteristic factors according to their weights or coefficients;

[0094] Step S45, deleting one or more of the natural characteristic factors with the smallest weight or the smallest coefficient in the sorting, and retraining the model with the remaining natural characteristic factors until the number of remaining natural characteristic factors cannot continue to eliminate features, and obtaining the contribution rate of each natural characteristic factor;

[0095] Step S46, sorting the selected natural characteristic factors according to contribution rate, selecting natural characteristic factors from high to low according to contribution rate until the cumulative contribution rate exceeds 99%, and using the selected natural characteristic factors as the main controlling factors for predicting the natural background value of soil heavy metals, wherein the contribution rate of the main controlling factors is not less than 5%.

[0096] It should be noted that the correlation analysis method in this embodiment is Pearson analysis, which calculates the correlation coefficient |r| between all natural characteristic factors, first selects a pair of first characteristic factors with |r| greater than 0.6, and simultaneously calculates the mean |r| of the first pair of first characteristic factors for the second characteristic factor, and eliminates the characteristic factors with larger means, until the |r| between all characteristic factors is less than 0.6, and the remaining characteristic factors are retained.

[0097] It should be noted that the natural characteristic factors are sorted according to their weights or coefficients. Usually, the natural characteristic factors with the smallest weights or coefficients are considered to be the least important. One or more natural characteristic factors with the smallest weights or coefficients in the sorting are deleted, and the model is trained with the remaining natural characteristic factors until the required number of natural characteristic factors is reached or no further characteristics can be proposed. At this time, the contribution rate of each natural characteristic factor is obtained, and the selected natural characteristic factors are sorted according to the contribution rate. Natural characteristic factors are selected from high to low according to the contribution rate until the cumulative contribution rate exceeds 99%. The selected natural characteristic factors are used as the main controlling factors for predicting the natural background value of soil arsenic. Furthermore, the main controlling factor should ensure that its contribution rate is not less than 5%. If the cumulative contribution rate of all natural characteristic factors with a contribution rate greater than 5% does not exceed 99%, appropriate choices should be made based on actual conditions.

[0098] In addition, in one embodiment, referring to Figure 5 ,exist Figure 1 Step S13 of the illustrated embodiment also includes but is not limited to the following steps:

[0099] Step S51, obtaining a data set of soil samples and dividing it into K mutually exclusive subsets of the same size, taking the union of K-1 mutually exclusive subsets as a training set, and the remaining mutually exclusive subsets as a test set and a validation set;

[0100] Step S52, obtaining K groups of mutually exclusive subsets, test sets and validation sets, and performing K training times on the K groups of mutually exclusive subsets, test sets and validation sets to obtain a data set division result of the soil samples;

[0101] Step S53, taking the main control factor as input and the arsenic natural background value as output, a soil arsenic natural background value prediction model is established using a random forest algorithm.

[0102] It should be noted that the dataset is divided into K mutually exclusive subsets of similar size, the union of K-1 subsets is used as the training set, and the remaining subsets are used as test sets and validation sets. K groups of training sets, test sets and validation sets can be obtained. After K training times, the dataset partition result under the optimal loss function evaluation condition can be obtained. The value of K depends on the sample size. For 200-500 samples, K=5, and for more than 500 samples, K=10.

[0103] In addition, in one embodiment, referring to Figure 6 ,exist Figure 1 Step S14 of the illustrated embodiment also includes but is not limited to the following steps:

[0104] Step S61, setting the upper limit range of soil arsenic pollution source types;

[0105] Step S62, when the difference between the actual test value and the natural background prediction value is within the upper limit range, the source type of soil arsenic pollution is natural pollution;

[0106] Step S63: When the difference between the actual test value and the natural background prediction value is greater than the upper limit interval, the source type of soil arsenic pollution is human pollution.

[0107] It should be noted that the prediction upper limit interval is used as the discrimination line for the source of soil arsenic pollution. In actual application, the main control factors of the target soil sample are input into the model to obtain the predicted value of the natural background of arsenic in the soil sample, and the predicted value of the natural background of arsenic is compared with the actual measured value of arsenic. If the difference between the actual test value and the predicted value of the natural background is within the prediction upper limit interval, it indicates that the arsenic in the soil is of natural origin; on the contrary, if the difference between the actual test value and the natural background value exceeds the prediction upper limit interval, it indicates that the arsenic in the soil is of human origin, and it is recommended to check the historical production activities of surrounding enterprises.

[0108] It should be noted that the upper limit interval is the prediction interval calculation method of the normal distribution. The prediction upper limit interval (predicted value + σ) is defined based on the error standard deviation σ of the model on the validation set.

[0109] In addition, in one embodiment, referring to Figure 7 ,exist Figure 3 Step S37 of the illustrated embodiment also includes but is not limited to the following steps:

[0110] Step S71, performing an initialization operation on the flow field in the three-dimensional reconstructed image so that the distribution function in the flow field is initialized to an equilibrium distribution;

[0111] Step S72, calculating a first distribution function of the flow field according to a preset single relaxation model, and performing a collision on the first pixel point and the second pixel point;

[0112] Step S73, determining a plurality of first pore points and second pore points in the three-dimensional reconstructed image, and when the downstream grid of the first pore point is the second pore point, migrating the first pore point, and when the downstream grid of the first pore point is not the second pore point, rebounding the first pore point;

[0113] Step S74, determining a second distribution function of the incoming flow field, and assigning the second distribution function to a post-collision distribution function value of the outgoing flow field on the buffer layer of the single relaxation model;

[0114] Step S75, obtaining the average flow rate of the flow field and calculating the pressure difference of the flow field, and calculating the permeability coefficient of the flow field according to the average flow rate and the pressure difference;

[0115] Step S76, calculating the relative error of the permeability coefficient according to the permeability coefficient.

[0116] It should be noted that the specific calculation method of soil permeability coefficient is as follows: ① Select the mercury injection surface as reservoir, set the radius of a single pore r = 0, search for the mercury injection area domain(0) connected to reservoir in the entire area, and calculate the voxel number v(0) of domain(0); ② Search for the pre-processed area pre-domain(r+1,1) containing a sphere with a radius of r+1 in domain(r), search for the pre-processed area pre-domain(r+1,2) connected to reservoir in the pre-processed area pre-domain(r+1,1), use pre-domain(r+1,2) as the seed and r+1 as the radius, and obtain the mercury injection area domain(r+1) through expansion search; ③ Update the radius r = r+1, calculate the voxel number v(0) of domain(r) l number v(r), if v(r)>0, go to step ②, otherwise the calculation is completed; for the permeability coefficient, the lattice Boltzmann method (LBM) will be used for numerical simulation calculation, and the calculation steps are as follows: ① Initialize the flow field: let the density ρ=1, the velocity U=(0,0,0), and initialize the distribution function fi to the equilibrium distribution according to the initial flow field; ② Collision: Use the single relaxation model for calculation, first calculate the equilibrium distribution function gi containing the volume force, and then execute the collision fi*; ③ Migration-rebound: If the downstream grid point of the current pore point (x, y, z) along the i direction is a pore point, then execute migration, otherwise execute rebound; ④ Assign boundary conditions: Execute periodic boundary conditions, that is, the unknown distribution function entering the flow field on one side is assigned to the post-collision distribution function value of the outflow flow field on the buffer layer on the other side of the corresponding position; ⑤ Calculate the permeability coefficient: First, calculate the average flow rate Q, then calculate the pressure difference ΔP, and finally calculate the permeability coefficient k. ⑥ Calculate the relative error of permeability coefficient: let ε=|k-k0| / k0. If ε<10-6, stop the calculation and take the k value at this time as the permeability coefficient; otherwise let k0=k and go to step ② to recalculate.

[0117] Furthermore, in this embodiment, the three-dimensional reconstruction of the two-dimensional scanned image is performed in the process of performing CT scanning on the soil sample, and the soil sample is rotated 360°. During the rotation of 360°, the soil sample is scanned and photographed in sequence every time the soil sample rotates 0.5°, and finally 720 two-dimensional projection images are obtained. The 720 two-dimensional images are then reconstructed in three dimensions to obtain the first three-dimensional reconstructed image of the sample.

[0118] The specific implementation process is as follows:

[0119] 210 natural original structure soil samples were collected, and the structural parameters of the soil samples were obtained through X-ray CT combined with numerical simulation calculations; the mineral phase composition and element content of the soil samples were obtained through XRD and XRF, and the type and content of arsenic-containing minerals were obtained through MLA analysis; the physical and chemical properties such as pH and moisture content were obtained through physical and chemical analysis methods; finally, environmental indicators such as altitude and precipitation were obtained through field records and data collection.

[0120] Data standardization, correlation analysis, feature selection and importance analysis were performed on the natural characteristic factors of soil, and the main controlling factors were determined, which were connected porosity, arsenic-containing limonite, iron, lead, sulfur, manganese, and pH value. The 210 samples were subjected to K (K = 5) fold cross validation to divide the data set. The main controlling factors were used as input and the natural background value of soil arsenic was used as output. A soil arsenic natural background value prediction model based on random forest was built. The model was optimized by means of mechanism constraints, and the final model had an R of 1.0 on the validation set. 2 It reached 0.7909. Subsequently, the prediction interval estimation method was used to establish the discrimination line of the source type of soil arsenic pollution. In this embodiment, the standard deviation of the error of the model on the validation set is 41.69. Therefore, when the established model is used to calculate the natural background value of arsenic in the target soil, if the difference between the measured value and the predicted value does not exceed 41.69, it indicates that the arsenic in the soil is of natural origin; if the measured value is higher than the predicted value by 41.69 or more, it indicates that the arsenic in the soil is of human origin.

[0121] like Figure 8 As shown, Figure 8 The present invention also provides a soil heavy metal pollution source identification device, including:

[0122] The processor 801 may be implemented by a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0123] The memory 802 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 802 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 802, and the processor 801 calls and executes the soil heavy metal pollution source identification method of the embodiment of this application;

[0124] Input / output interface 803, used to implement information input and output;

[0125] The communication interface 804 is used to realize the communication interaction between the present apparatus and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WI FI, Bluetooth, etc.);

[0126] A bus 805 that transmits information between the various components of the device (e.g., the processor 801, the memory 802, the input / output interface 803, and the communication interface 804);

[0127] The processor 801 , the memory 802 , the input / output interface 803 and the communication interface 804 are connected to each other in communication within the device via a bus 805 .

[0128] An embodiment of the present application also provides an electronic device, including the device for identifying the source of heavy metal pollution in soil as described above.

[0129] An embodiment of the present application also provides a storage medium, which is a computer-readable storage medium. The storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method for identifying the source of heavy metal pollution in soil is implemented.

[0130] As a non-transient computer-readable storage medium, the memory can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof. The device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and are implemented to be located in one place, or may also be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0131] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media generally include computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0132] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the above-mentioned implementation mode. Technical personnel familiar with the field can also make various equivalent deformations or substitutions under the shared conditions without violating the spirit of the present invention. These equivalent deformations or substitutions are all included in the scope defined by the claims of the present invention.

Claims

1. A method for identifying the source of heavy metal pollution in soil, characterized in that: include: Obtain the natural characteristic factors of soil samples and the actual test values ​​of heavy metals in soil samples; Performing a preprocessing operation on the natural characteristic factors to obtain a main control factor that is highly correlated with the natural background value of heavy metals in the soil sample; Constructing a prediction model for the natural background value of heavy metals in soil, inputting the main control factors into the prediction model, and obtaining the predicted value of the natural background value of heavy metals in soil; The predicted natural background value of heavy metals in the soil is compared with the actual test value of heavy metals in the soil to determine the type of heavy metal pollution source of the soil sample.

2. The method for identifying the source of heavy metal pollution in soil according to claim 1, characterized in that: The natural characteristic factors of the soil sample obtained include: Scanning the soil sample to obtain a plurality of two-dimensional scanned images of the soil sample, and performing three-dimensional reconstruction on the two-dimensional scanned images to obtain a first three-dimensional reconstructed image of the soil sample; Performing threshold segmentation on the first three-dimensional reconstructed image to obtain a binary segmentation result of the three-dimensional reconstructed image, and distinguishing soil skeletons and soil pores in a plurality of the three-dimensional images according to the binary segmentation result to obtain a second three-dimensional reconstructed image; The second three-dimensional reconstructed image is calculated to obtain soil structural parameters of the soil sample.

3. The method for identifying the source of heavy metal pollution in soil according to claim 2, characterized in that: After obtaining the multiple two-dimensional scanning images of the soil sample, the method further includes: Obtain pixel points of the two-dimensional scanned image and mark them as first pixel points and second pixel points, wherein the first pixel points are all pixel points of the soil sample, and the second pixel points are pixel points of soil pores in the soil sample; Acquire a first porosity of the second pixel in the first pixel at a first resolution, and a second porosity of the second pixel in the first pixel at a second resolution, and sum the first porosity and the second porosity to obtain a third porosity; Alternatively, a pore distance map of the three-dimensional reconstructed image is obtained, and the soil pores are divided into a plurality of independent pores according to the pore distance map; Selecting a mercury injection surface of the independent pore, and injecting mercury into the independent pore according to the mercury injection surface; When the radius of the independent pore is zero, obtaining the connected porosity of the three-dimensional reconstructed image; Obtaining the diameters of all the independent pores in the three-dimensional reconstructed image, performing statistics on the diameters, and calculating the average value and median value of the soil pore radius in the three-dimensional reconstructed image; Calculate the soil permeability coefficient.

4. The method for identifying the source of heavy metal pollution in soil according to claim 1, characterized in that: The natural characteristic factors are preprocessed, including: Performing standardization processing on the plurality of natural characteristic factors, and calculating a plurality of correlation coefficients of the plurality of natural characteristic factors; Eliminate the parts of the plurality of natural characteristic factors whose correlation coefficients with other natural characteristic factors are higher than 0.6, so that the correlation coefficients between all the remaining natural characteristic factors are less than 0.6; Recursively eliminate the natural feature factors whose correlation coefficient is less than 0.6, so that all available natural feature factors can train a model, and calculate the weight or coefficient of each feature; Sorting the natural characteristic factors according to the weights or coefficients of the natural characteristic factors; Deleting one or more of the natural feature factors with the smallest weight or coefficient in the sorting, and retraining the model with the remaining natural feature factors until the number of remaining natural feature factors is unable to continue to eliminate features, and obtaining the contribution rate of each natural feature factor; The selected natural characteristic factors are sorted according to their contribution rates, and the natural characteristic factors are selected from high to low according to the contribution rates until the cumulative contribution rate exceeds 99%, and the selected natural characteristic factors are used as the main controlling factors for predicting the natural background value of soil heavy metals, wherein the contribution rate of the main controlling factors is not less than 5%.

5. The method for identifying the source of heavy metal pollution in soil according to claim 1, characterized in that: The step of constructing a prediction model for natural background values ​​of heavy metals in soil comprises: Obtaining a data set of the soil sample and dividing it into K mutually exclusive subsets of equal size, taking the union of K-1 mutually exclusive subsets as a training set, and the remaining mutually exclusive subsets as a test set and a validation set; Obtaining K groups of mutually exclusive subsets, the test set, and the validation set, and performing K training times on the K groups of mutually exclusive subsets, the test set, and the validation set to obtain a data set division result of the soil sample; With the main controlling factor as input and the natural background value as output, a soil heavy metal natural background value prediction model is established using a random forest algorithm.

6. The method for identifying the source of heavy metal pollution in soil according to claim 1, characterized in that: Determining the type of heavy metal pollution source of the soil sample includes: Setting the upper limit range of the heavy metal pollution source types; When the difference between the actual test value and the natural background predicted value is within the upper limit range, the heavy metal pollution source type is natural pollution; When the difference between the actual test value and the natural background predicted value is greater than the upper limit interval, the source type of heavy metal pollution is man-made pollution.

7. The method for identifying the source of heavy metal pollution in soil according to claim 3, characterized in that: The calculating of the soil permeability coefficient comprises: Performing an initialization operation on the flow field in the three-dimensional reconstructed image so that the distribution function in the flow field is initialized to an equilibrium state distribution; Calculating a first distribution function of the flow field according to a preset single relaxation model, and performing a collision on the first pixel point and the second pixel point; Determine a plurality of first pore points and second pore points in the three-dimensional reconstructed image, and when a downstream grid of the first pore point is the second pore point, migrate the first pore point, and when a downstream grid of the first pore point is not the second pore point, rebound the first pore point; Determine a second distribution function entering the flow field, and assign the second distribution function to a post-collision distribution function value of the flow field flowing out of the flow field on the buffer layer of the single relaxation model; Acquiring the average flow rate of the flow field and calculating the pressure difference of the flow field, and calculating the permeability coefficient of the flow field according to the average flow rate and the pressure difference; The relative error of the permeability coefficient is calculated according to the permeability coefficient.

8. A device for identifying the source of heavy metal pollution in soil, characterized in that: It includes at least one control processor and a memory for communicating with the at least one control processor; the memory stores instructions that can be executed by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute the method for identifying the source of heavy metal pollution in soil as described in any one of claims 1 to 7.

9. An electronic device, characterized in that: It includes the soil heavy metal pollution source identification device as described in claim 8.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the method for identifying the source of heavy metal pollution in soil as described in any one of claims 1 to 7.

Images