CN117932444A

CN117932444A - Soil heavy metal laboratory detection sample point screening method and device

Info

Publication number: CN117932444A
Application number: CN202410310868.5A
Authority: CN
Inventors: 郜允兵; 张芷; 潘瑜春; 马文玉; 马浩森; 顾晓鹤; 李晓岚; 周艳兵; 王柳毅; 刘玉
Original assignee: Research Center of Information Technology of Beijing Academy of Agriculture and Forestry Sciences
Current assignee: Research Center of Information Technology of Beijing Academy of Agriculture and Forestry Sciences
Priority date: 2024-03-19
Filing date: 2024-03-19
Publication date: 2024-04-26

Abstract

The invention provides a method and a device for screening detection sample points of a soil heavy metal laboratory, and relates to the field of pollution detection, wherein the method comprises the following steps: calculating soil investigation comprehensive error indexes of all soil sampling points based on detection results of heavy metal XRF to be detected of all soil sampling points of a research area; the soil investigation comprehensive error index comprises an agricultural land risk area error index, an XRF quantitative limit error index and an XRF detection error index; performing preliminary screening of laboratory detection sample points based on soil investigation comprehensive error indexes of the soil sampling points to obtain a laboratory detection preliminary screening sample point set; and the laboratory detection target sample points are determined from the laboratory detection primary screening sample point set through a multi-target simulated annealing algorithm by using the minimum environmental similarity and the maximum spatial distribution uniformity as optimization targets, so that the number of laboratory detection soil samples is reduced, the regional soil heavy metal prediction precision is improved, and the farmland soil heavy metal monitoring and investigation cost is reduced.

Description

Soil heavy metal laboratory detection sample point screening method and device

Technical Field

The invention relates to the technical field of pollution detection, in particular to a method and a device for screening detection sample points of a soil heavy metal laboratory.

Background

The farmland soil sampling investigation is the basis for carrying out soil pollutant investigation, polluted farmland restoration treatment and regional farmland pollution risk management and control, and is an important way and means for comprehensively knowing the heavy metal content and the spatial variation characteristics of the soil in the region.

At present, soil sampling points with certain density can be distributed in a research area in the process of farmland soil sampling investigation, X-ray fluorescence spectrum (X Ray Fluorescence, XRF) detection and an experimental detection method are combined, a certain number of soil sampling points are selected and distributed in a investigation area with larger XRF error for laboratory detection, and the rest part is detected by a rapid detection instrument.

However, the following disadvantages exist in the soil heavy metal monitoring survey spot layout process under the combination of XRF detection and laboratory detection: 1) The space differences of soil organic matters, soil matrix, soil types, soil textures and the like are large, and the influence of the complexity of the soil matrix on the XRF detection result is not fully considered; 2) When the laboratory is combined with XRF in soil sampling investigation, the selection of laboratory detection sample points has randomness, and the number of laboratory detection samples and the detection cost are easily increased.

Disclosure of Invention

Aiming at the problems existing in the prior art, the embodiment of the invention provides a method and a device for screening detection sample points of a soil heavy metal laboratory.

The invention provides a method for screening detection sample points of a soil heavy metal laboratory, which comprises the following steps:

Calculating soil investigation comprehensive error indexes of all soil sampling points based on detection results of heavy metal XRF to be detected of all soil sampling points of a research area; the soil investigation comprehensive error index comprises an agricultural land risk area error index, an XRF quantitative limit error index and an XRF detection error index;

Performing preliminary screening of laboratory sampling points based on soil investigation comprehensive error indexes of the soil sampling points to obtain a laboratory detection preliminary screening sampling point set;

And determining laboratory detection target sample points from the laboratory detection primary screening sample point set by using a multi-target simulated annealing algorithm according to the optimization targets of the minimization of the environmental similarity and the maximization of the spatial distribution uniformity.

Optionally, calculating the XRF detection error index of each soil sampling point based on the XRF detection result of the heavy metal to be detected of each soil sampling point in the research area includes:

Acquiring priori knowledge and XRF detection data of a research area; the priori knowledge comprises a knowledge base of XRF detection errors under the element concentrations of different heavy metals to be detected, and the XRF detection data of the research area comprise XRF detection results of the heavy metals to be detected at each soil sampling point of the research area;

Based on the prior knowledge and the XRF detection data of the research area, calculating XRF detection error indexes of all soil sampling points of the research area according to a naive Bayesian principle.

Optionally, the calculating, based on the prior knowledge and the XRF detection data of the research area, an XRF detection error index of each soil sampling point of the research area according to a naive bayes principle specifically includes:

Based on the priori knowledge, calculating a priori probability density function of each soil sampling point belonging to different XRF detection error grades in the research area according to a naive Bayes principle;

And for any sample point, based on the prior probability density function, respectively calculating probabilities of different XRF detection error indexes corresponding to the any sample point, and selecting the XRF detection error index with the maximum probability value as the XRF detection error index of the any sample point.

Optionally, calculating the agricultural land risk area misdemarcation error index of each soil sampling point based on the detection result of the heavy metal XRF to be detected of each soil sampling point of the research area includes:

For any sample point, determining an agricultural land risk area misdemarcation error index of the any sample point based on an XRF detection result of heavy metal to be detected of the any sample point, a preset maximum acceptable error, a risk control value and a risk screening value of the heavy metal to be detected under a national standard.

Optionally, based on the XRF detection result of heavy metal to be detected of each soil sampling point in the research area, calculating the XRF quantitative limit error index of each soil sampling point includes:

and for any sample point, determining the XRF quantitative limit error index of the any sample point based on the XRF detection result of the any sample point on the heavy metal to be detected and the quantitative limit of the XRF detector.

Optionally, the optimizing target is obtained by minimizing the environmental similarity and maximizing the uniformity of the spatial distribution, and the laboratory detection target sample points are determined from the laboratory detection preliminary screening sample point set through a multi-target simulated annealing algorithm, which specifically comprises:

Determining the number of laboratory sample points corresponding to different soil investigation comprehensive error indexes based on the soil investigation comprehensive error indexes of each soil sampling point;

respectively taking the similarity optimization objective function and the spatial distribution uniformity optimization objective function as fitness functions, and respectively setting cooling paths of different fitness functions;

Determining a first laboratory detection sample set based on the laboratory sample number and the preliminary screening laboratory detection sample set, and initializing a simulated annealing algorithm;

disturbing the first laboratory detection sample set to generate a second laboratory detection sample set, and calculating different fitness function values corresponding to the second laboratory detection sample set;

Determining whether to accept the second laboratory detection sample set based on a Metropolis criterion and different fitness function values corresponding to the second laboratory detection sample set, and cooling paths of the different fitness functions according to a set cooling rate; taking the second laboratory test sample set as the updated first laboratory test sample set if the second laboratory test sample set is accepted; maintaining the first set of laboratory test samples unchanged without accepting the second set of laboratory test samples;

repeating the process of disturbing the first laboratory detection sample point set to generate a second laboratory detection sample point set under the condition that the convergence condition is not reached, calculating different fitness function values corresponding to the second laboratory detection sample point set, determining whether to accept the second laboratory detection sample point set, and cooling a cooling path according to a set cooling rate until the convergence condition is reached; in the event that a convergence condition is reached, a laboratory test target sample point is determined based on the current first set of laboratory test sample points.

Optionally, the XRF detection result is obtained by detecting each soil sampling point in a soil monitoring investigation point distribution scheme;

The soil monitoring survey deployment is generated based on historical survey data or auxiliary environmental data.

The invention also provides a soil heavy metal laboratory detection sampling point screening device, which comprises:

The calculation module is used for calculating soil investigation comprehensive error indexes of all soil sampling points based on detection results of heavy metal XRF to be detected of all soil sampling points of the research area; the soil investigation comprehensive error index comprises an agricultural land risk area error index, an XRF quantitative limit error index and an XRF detection error index;

The primary screening module is used for carrying out primary screening on laboratory sampling points based on soil investigation comprehensive error indexes of the soil sampling points to obtain a laboratory detection primary screening sample point set;

And the determining module is used for determining laboratory detection target sample points from the laboratory detection primary screening sample point set through a multi-target simulated annealing algorithm by taking the minimum of the environmental similarity and the maximum of the spatial distribution uniformity as optimization targets.

The invention also provides electronic equipment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the soil heavy metal laboratory detection sampling point screening method according to any one of the above when executing the program.

The invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a soil heavy metal laboratory detection spot screening method as described in any of the above.

According to the soil heavy metal laboratory detection sample point screening method and device, the soil investigation comprehensive error indexes of all soil sampling points are calculated, the laboratory detection sample points are initially screened according to the soil investigation comprehensive error indexes of all soil sampling points, finally, environmental similarity minimization and spatial distribution uniformity maximization are used as optimization targets, and laboratory detection target sample points are determined from the initially screened sample point set through a multi-target simulated annealing algorithm, so that the number of laboratory detection soil samples is reduced, the regional soil heavy metal prediction accuracy is improved, and the farmland soil heavy metal monitoring investigation cost is reduced.

Drawings

In order to more clearly illustrate the invention or the technical solutions of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the invention, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic flow chart of a screening method of a soil heavy metal laboratory detection sample point;

FIG. 2 is a schematic diagram of a method for calculating a maximum acceptable error according to the present invention;

FIG. 3 is a schematic flow chart of a simulated annealing algorithm of the multi-objective optimization function provided by the invention;

FIG. 4 is a flow chart of laboratory test sample layout provided by the present invention;

Fig. 5 is a schematic structural diagram of a screening device for detecting sampling points in a soil heavy metal laboratory;

fig. 6 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Fig. 1 is a schematic flow chart of a screening method of a soil heavy metal laboratory detection sampling point, as shown in fig. 1, the method comprises the following steps:

step 100, calculating soil investigation comprehensive error indexes of all soil sampling points based on detection results of heavy metal XRF to be detected of all soil sampling points of a research area; the soil investigation comprehensive error index comprises an agricultural land risk area error index, an XRF quantitative limit error index and an XRF detection error index.

And 101, performing preliminary screening on laboratory sampling points based on soil investigation comprehensive error indexes of the soil sampling points to obtain a laboratory detection preliminary screening sample point set.

And 102, determining laboratory detection target sample points from a laboratory detection primary screening sample point set by using a multi-target simulated annealing algorithm according to the optimization targets of the minimization of the environmental similarity and the maximization of the spatial distribution uniformity.

Specifically, aiming at the prior art, in the current method for combining XRF with laboratory detection, a certain density of XRF soil sampling points are required to be distributed in a research area, a certain number of soil sampling points are selected and distributed in a investigation area with larger XRF errors for laboratory detection, the rest is detected by using a rapid detection instrument, and the relationship modeling of soil environment variables and XRF detection results and the combined solution of the laboratory are fully utilized, so that the accuracy and stability of the XRF detection results can be improved, the requirement of regional soil investigation can be met, the investigation cost of soil heavy metals can be reduced, and an implementation path and a scheme can be further provided for the rapid detection instrument in popularization and application of quantitative detection of soil heavy metals.

In the method, the laboratory detection samples are more in number and the detection cost is increased due to the randomness of the laboratory detection samples in the prior art, so that the embodiment of the invention provides a soil heavy metal laboratory detection sample screening method, and a combined distribution design is developed by adopting high-density XRF soil detection samples and a certain number of laboratory detection samples, so that the laboratory detection samples can be selected according to the high-density XRF detection results.

Firstly, the XRF can be adopted for detecting the content of heavy metal to be detected of each soil sampling point in the soil of the research area, and XRF detection results of each soil sampling point on the heavy metal to be detected are obtained. The XRF detection instrument enables a large-scale soil sampling investigation by virtue of the characteristic of low cost, takes the XRF sampling investigation result of a research area as a preliminary judgment result of soil pollutants, and can provide priori experience and knowledge for the targeted detection and assay analysis of soil samples in subsequent laboratories.

In some embodiments, the XRF detection result is obtained by detecting each soil sampling point in a soil monitoring survey point arrangement; the soil monitoring survey deployment is generated based on historical survey data or auxiliary environmental data.

For the research area of the existing historical survey data, the existing historical survey data can be utilized to explore the spatial characteristics of the heavy metal to be detected, wherein the spatial characteristics comprise spatial autocorrelation, spatial dissimilarity and spatial distribution pattern.

If the heavy metal to be detected has spatial autocorrelation, classical statistical methods (such as a half variance function method) can be directly adopted to arrange XRF sampling points.

If the heavy metal to be detected has space dissimilarity, firstly, after the space of the dissimilarity is partitioned, the space autocorrelation of each subarea is measured, and an XRF sampling point is distributed by adopting a classical statistical method (for example, a half variance function method).

For each space partition, XRF sampling points can be laid by adopting a grid method, the distance between any two sampling points is required to be less than pollutant Range (Range)/2 during point location layout, and the point location falls on the largest land area of cultivated land in the grid as much as possible.

The calculation formula of the half variance is shown as follows:

In the method, in the process of the invention, Is a variable/>Is/>Is here/>Representing existing historical survey data; /(I)And/>Respectively represent the coordinates/>、/>Historical data values at; /(I)Is distance/>Logarithms of test sample points spaced apart.

The steps for laying out XRF sampling points of a specific research area can be as follows:

(1) And (3) measuring the spatial correlation of the soil history sample points by utilizing Moran I in geographic information software such as ArcGIS/super map, if the spatial correlation exists, carrying out spatial interpolation by utilizing ordinary Krige to calculate the spatial correlation distance h of the region, and if the spatial layering distinction exists, clustering by utilizing algorithms such as K-mean or cluster analysis and then partitioning, and measuring the spatial correlation distance h in the sub-region.

(2) And h/2 is used as sampling density, a grid is generated by Fishnet, a rapid XRF check point of soil heavy metal is arranged on the largest land block in the grid, and finally an XRF soil monitoring and investigation point arrangement scheme S _xrf is formed.

The method aims at a research area without historical soil heavy metal investigation data, can be used for carrying out space division on the research area by combining with auxiliary environment data such as the soil type, the topography and the topography of the research area, the distribution of factory enterprises and the like, ensures that each layered partition has obvious regionality, and adopts a grid method to lay investigation sample points in each partition.

The sample point distribution density can be set by combining historical experience and literature data about regional soil heavy metal variation and sampling investigation purposes. For example, according to the basic principle that the soil heavy metal monitoring density of farmland soil remediation and restoration is high, the proper investigation density is selected (generally, the samples are 200m×200m, 500m×500m, 1km×1km, 5km×5km, 10km×10km, and the like). The farmland block where the grid center points are located is a sample point which is preliminarily distributed, so that the XRF detection points are ensured to be uniformly distributed to a research area, and a soil sample point preliminary investigation distribution scheme for comprehensive investigation of the XRF of the research area is formed.

The specific operation steps can refer to the sample point layout step of a research area with historical survey data, firstly, a grid is generated by using Fishnet and the like of ArcGIS/SuperMap and other software, soil heavy metal XRF quick survey points are laid on the largest cultivated land in the grid, and finally an XRF soil monitoring survey point layout scheme S _xrf is formed.

The detection limit and the precision of the portable XRF soil detector are limited, so that the accuracy of soil heavy metal prediction drawing and the accuracy of soil pollution boundaries are effectively improved, and the portable XRF detection is required to be focused on suspected high-error areas.

According to the embodiment of the invention, after the XRF detection results of each soil sampling point of the research area on the heavy metal to be detected are obtained according to the XRF soil monitoring investigation point distribution scheme, the soil investigation comprehensive error index of each soil sampling point can be calculated according to the XRF detection results.

The soil investigation comprehensive error index in the embodiment of the invention comprises at least three types, namely an agricultural land risk area error index, an XRF quantitative limit error index and an XRF detection error index. The agricultural land risk area misdemarcation error index is used for expressing errors generated by screening and controlling the easily-misdemarcation area aiming at the soil heavy metal risk; XRF quantitative limit error index indicates errors below the XRF quantitative limit; XRF detection error index indicates detection errors other than the two errors described above.

After the soil investigation comprehensive error indexes of the soil sampling points are obtained through calculation, whether the soil investigation comprehensive error indexes of the soil sampling points are larger or not can be determined, and the sampling points with the larger soil investigation comprehensive error indexes are used as a laboratory detection primary screening sampling point set.

The process of selecting the sampling point as the laboratory test primary screening sample point set according to the embodiment of the present invention is not limited, and for any sample point, there may be no soil investigation integrated error index (it should be understood that there is no soil investigation integrated error index here and is not an absolute zero error, but refers to that various soil investigation integrated error indexes are smaller), or there may be one or more types of soil investigation integrated error indexes (it should be understood that there is a certain type of soil investigation integrated error index here and refers to that the type of soil investigation integrated error index is larger), and the laboratory test primary screening sample point set may be determined according to the actual situation, for example, a sample point where at least one type of soil investigation integrated error index exists is used as the laboratory test primary screening sample point set, or a sample point where three types of soil investigation integrated error indexes exist is used as the laboratory test primary screening sample point set, or the like.

Because of the spatial heterogeneity of soil attributes, the more densely the spatial distribution of soil sample points and the more detailed the attribute information, the more the attribute data is close to the real condition, so that the number of sample points in the obtained laboratory detection primary screening sample point set is still relatively large, and the further optimization of the number of sample points can be considered to reduce the detection cost.

Due to the characteristics of the X-ray, factors influencing XRF detection accuracy mainly include matrix absorption enhancement effect, granularity effect, and the like when actual soil pollutant measurement is performed. The matrix absorption enhancement effect is: when the specific spectral line wavelength of a certain coexisting element is just slightly larger than the absorption limit of the element to be detected, the element to be detected absorbs a great amount of fluorescence spectrum emitted by the coexisting element, such as Fe (26) to Cr (24), ni (28) to Cr (24) and the like in the alloy. The particle size effect is: the size of the granularity can influence the scattering condition of X-ray fluorescence, so that the diffuse reflection of the surface is increased to cause the reduction of the measurement intensity, thereby influencing the XRF measurement result. In the above-mentioned situation, heavy metal elements and other pollutant elements are accompanied in soil investigation, and in addition, the contents (such as Fe, al, mn, ca, etc.) of other metal elements in soil matrix (crust element) are also quite different, that is, XRF detection accuracy can be influenced by the comprehensive effects of soil environments such as soil type, soil matrix, soil particle size, etc.

That is, in the optimal selection of laboratory test spots, the influence of the complex soil matrix needs to be fully considered, so that the laboratory test spots can cover different conditions of the XRF test influence factors, namely, the representativeness of the XRF test influence factor attributes.

Therefore, in the embodiment of the invention, after the laboratory detection primary screening sample point set is obtained, the environmental similarity minimization and the spatial distribution uniformity maximization can be used as optimization targets, and the laboratory detection target sample points are determined from the laboratory detection primary screening sample point set through a multi-target simulated annealing algorithm. The method takes the minimization of the environmental similarity and the maximization of the spatial distribution uniformity as optimization targets, and through optimizing by a multi-target simulated annealing algorithm, representative sampling points can be determined, and abnormal and redundant sampling points in the detection points of an original laboratory are removed, so that the purposes of improving the drawing precision of the heavy metal content of soil or improving the classification accuracy of pollution grades and effectively reducing the detection cost of the heavy metal of the soil are achieved.

According to the soil heavy metal laboratory detection sample point screening method provided by the invention, the soil investigation comprehensive error index of each soil sampling point is calculated, the laboratory detection sample points are initially screened according to the soil investigation comprehensive error index of each soil sampling point, finally, the environmental similarity minimization and the spatial distribution uniformity maximization are used as optimization targets, and the laboratory detection target sample points are determined from the initially screened sample point set through a multi-target simulated annealing algorithm, so that the number of laboratory detection soil samples is reduced, the regional soil heavy metal prediction precision is improved, and the farmland soil heavy metal monitoring investigation cost is reduced.

Specifically, in the embodiment of the present invention, the XRF detection error indexes may be divided into low, medium and high 3 classes, and respectively used=0, 1,2, The calculation formula of which is shown as follows:

In the method, in the process of the invention, For the sampling point/>XRF detection error index,/>For the sampling point/>Is a soil detection error value.Is a sample set in which/>And respectively representing XRF detection results of the sample points on the heavy metal to be detected.

Soil detection error values generally refer to the difference between the XRF detection result and the true value of the soil sample, which in embodiments of the present invention may refer to the difference between the XRF detection result and the laboratory detection result, as quantitatively measured using the following equation (relative error):

Wherein the method comprises the steps of Representing the sample points/>Soil detection error values of (2); /(I)For the sampling point/>XRF detection results for heavy metals to be detected; /(I)For the sampling point/>Is a laboratory test result of (2).

According to the above formula, since the XRF detection error index is divided and the degree of deviation (soil detection error value) between the XRF detection result and the laboratory detection result needs to be calculated, the error classification result cannot be obtained when the laboratory detection data of the unknown sample point is obtained.

Therefore, the embodiment of the invention provides an XRF detection error grading mode based on priori knowledge, namely, according to the relation between the size of the soil detection error and the concentration of the soil to-be-detected pollutant, the XRF detection error index of the sample point of the unknown laboratory detection result is deduced.

In the embodiment of the invention, a Gaussian naive Bayes algorithm can be adopted, and the relation between the concentration of the object to be detected and the soil detection error is constructed by means of priori knowledge so as to judge the XRF detection error index of each soil sampling point.

The cause of errors in soil detection using XRF can be analyzed from the principle of X-ray fluorescence detection. XRF is a process of exciting a sample with X-rays as an excitation source, wherein electrons in the inner layer outside the core undergo transition to generate holes, electrons in the outer layer undergo transition to fill the holes, and in the process, electrons are converted from a high energy level to a low energy level, and a part of energy is emitted in the form of X-rays, so that a characteristic X-ray fluorescence spectrum is formed. However, in actual detection, the soil composition element types and the soil particle size composition are complex, so that the XRF detection result is influenced by the matrix effect, and the detection error exists. Especially when the content of the element to be detected is low, the excitation spectral line intensity is weak, and the interference caused by background interference and the matrix effect is obvious, so that the application effect of the existing XRF matrix correction method is poor, and a large detection error occurs.

Therefore, the soil detection error of the soil detection by using the XRF presents the basic characteristics that the concentration of the low to-be-detected object corresponds to the high detection error and the concentration of the high to-be-detected object corresponds to the low detection error, and the XRF detection error index can be judged based on the Naive Bayes mode (NBC) principle according to the actually measured concentration of the pollutant by the XRF. The naive Bayesian method divides data into two aspects, namely prior knowledge (K _G) which is the data or knowledge used to describe the overall features, in the embodiment of the invention: XRF detection result errors of the research area and element concentration of heavy metal to be detected of the research area; the other part of the data is XRF detection data (K _S): the method characterizes detection data related to a research area, namely XRF detection results of each soil sampling point of the research area on heavy metals to be detected.

According to the embodiment of the invention, the XRF detection error indexes of each soil sampling point of the research area can be obtained through the priori knowledge and XRF detection data according to the naive Bayes principle.

Specifically, in the embodiment of the invention, XRF detection error indexes are divided into low, medium and high 3 classes and respectively used=0, 1,2, Thus allowing XRF detection error index/>Take the value of/>，/>；/>For each soil sample instance,/>(J=1, 2, …, n), n being an integer,/>Representing XRF detection content of the object to be detected; then/>Belonging to class/>From the bayesian formula:

Wherein the method comprises the steps of Is a full probability formula, the same value for all data, here used/>Replacing; /(I)Is a priori probability; /(I)For/>Is a posterior probability of (c).

The detection of each soil sample heavy metal element is an independent event and is not influenced by other soil samples, so that the detection result of each sampling point is independent, and the above formula can be converted into:

naive Bayes is exemplified by computing each class The posterior probability is found to find the class with the maximum posterior probability, and the class is the maximum posterior hypothesis and is marked as/>The calculation formula is as follows:

Calculating a priori probabilities Namely, under the detection content of different XRF actual pollutants in prior knowledge, the probability of each XRF detection error is calculated as follows:

Wherein the method comprises the steps of Representing the total number of training samples,/>Representing category values as/>Is a number of samples of (a).

For the followingSince the XRF detection sample content is a continuous value, it is assumed that the continuous variable is Gaussian distribution, and therefore/>Features/>Expressed as mean is/>Sum of variances/>Is shown in the following formula:

The XRF detection data (namely the XRF detection results of each soil sampling point of the research area on the heavy metal to be detected) can be calculated respectively through the above process Time,/>Is a probability of (2).

Wherein the maximum probability value corresponds toThe value of (2) is the XRF detection error index of the sample to be classified, and finally the XRF detection error index of all the samples is obtained.

Specifically, for any sample point, the agricultural land risk region error index can be determined by the XRF detection result of the sample point on the heavy metal to be detected, a preset maximum acceptable error, the risk control value and the risk screening value of the heavy metal to be detected.

According to the embodiment of the invention, when the XRF detection result is closer to the soil heavy metal management and control threshold value, the management and control area to which the sampling point belongs is easier to be misplaced, so that any sampling point is taken as an example according to GB/T15618 soil environmental quality agricultural soil pollution risk management and control Standard, and the soil heavy metal cadmium Cd survey is taken as an exampleXRF detection result/>, for heavy metal to be detectedAccording to/>< Risk screening value B1, risk screening value B1. Ltoreq./>Risk control value B2 and/>And classifying the risk control value B2 to realize quantitative calculation of the agricultural land risk zone misdividing error index. The specific division basis is shown in table 1.

TABLE 1 soil pollution risk value for agricultural land (unit: mg/kg)

When the XRF detection result is within the threshold range of the risk control value and the threshold range of the risk screening value, the heavy metal content of the soil sample point is easily divided into other intervals in a staggered manner, so in some embodiments, the sample point can be defined by the following formulaIs a management region error value/>：

In the method, in the process of the invention,For the sampling point/>XRF detection results for heavy metals to be detected; /(I)For the preset maximum acceptable error, fig. 2 is a schematic diagram of a calculation method of the maximum acceptable error provided by the present invention, and as shown in fig. 2, in the embodiment of the present invention, calculation may be performed according to 15% of the threshold { B1, B2} (allowable error of soil heavy metal laboratory detection < 15%).

When (when)The soil sample detection value/>, can be consideredNot falling within the interval of detection value + -detection value x 15%, i.e. not within the agricultural land risk area misclassification error index generation area (/ >)) Similarly, let/>When the soil sample points are considered to be in the agricultural land risk zone error index generation region (/ >)) The following formula is shown:

In the method, in the process of the invention, For the sampling point/>Error index of agricultural land risk zone,/>For the sampling point/>Is used for controlling the region error value.

Specifically, for any sample point, the XRF quantitative limit error index of the sample point can be determined by the XRF detection result of the sample point on the heavy metal to be detected and the instrument quantitative limit of the XRF on the heavy metal to be detected.

In some embodiments, the sample isXRF quantitative limit error value/>The calculation can be performed by the following formula:

In the method, in the process of the invention, For the sampling point/>XRF detection results for heavy metals to be detected; c is the limit of instrument quantification. Taking cadmium Cd as an example, the quantitative limit of the instrument is 0.2-0.3mg/kg, and the value of C is 0.3mg/kg.

When the sample is/><0 It can be considered that the accuracy of the measurement results of the instrument is not high, and the sample points/>The XRF detection result of the (B) is in a high uncertainty area of the detection result, so that the sample points of the type can be listed as laboratory detection candidate points, laboratory detection can be properly carried out, and errors caused by XRF detection limits on regional soil heavy metal estimation are reduced.

When the XRF detection instrument detects low-concentration soil pollutants, a detection result is given to soil sample points below a quantitative limit through a built-in algorithm, and the result is only a theoretical value and does not have the judgment of whether the soil sample points are in the XRF quantitative limit region.

Therefore, the embodiment of the invention can be usedRepresenting the sample points/>Whether the soil sample is in an XRF quantitative limit area or not, and dividing each soil sample into soil samples below the quantitative limit (/ >) based on data obtained by XRF detection=0) And two regions above the quantitative limit=1), As shown in the following formula:

In the method, in the process of the invention, For the sampling point/>XRF quantitative limit error index of/(For the sampling point/>XRF quantitative limit error value of (c).

Disturbing the first laboratory detection sample point set to generate a second laboratory detection sample point set, and calculating different fitness function values corresponding to the second laboratory detection sample point set;

Determining whether to accept the second laboratory detection sample set based on the Metropolis criterion and different fitness function values corresponding to the second laboratory detection sample set, and cooling paths of different fitness functions according to the set cooling rate; taking the second laboratory test sample set as the updated first laboratory test sample set under the condition of receiving the second laboratory test sample set; maintaining the first set of laboratory test samples unchanged without accepting the second set of laboratory test samples;

Under the condition that the convergence condition is not reached, repeatedly disturbing the first laboratory detection sample set to generate a second laboratory detection sample set, calculating different fitness function values corresponding to the second laboratory detection sample set, determining whether to accept the second laboratory detection sample set, and cooling the cooling path according to the set cooling rate until the convergence condition is reached; in the event that a convergence condition is reached, a laboratory test target sample point is determined based on the current first set of laboratory test sample points.

Specifically, in the optimizing process using the simulated annealing algorithm, the number of laboratory sampling points corresponding to each soil investigation comprehensive error index type needs to be determined based on the soil investigation comprehensive error index type of each soil sampling point.

The process needs to fully consider the magnitude of the integrated error and allow for sampling errors (or accuracy levels) and confidence intervals, and determine a reasonable number of samples according to classical statistics.

The soil pollutant variation conditions (including variation coefficient, variance and the like) of the same comprehensive error classification sampling points can be calculated, and when the variation conditions are large, the soil pollutant contents are greatly different under different comprehensive error categories; conversely, a small difference in the spatial concentration distribution of soil contaminants is demonstrated. Therefore, the reasonable sampling point number determined according to classical statistics is the minimum sampling number [ ]) I represents the sample point of the soil investigation integrated error index (namely, agricultural land risk area error index+XRF quantitative limit error index+XRF detection error index), and the calculation process of the soil investigation integrated error index is shown in the following formula:

In the method, in the process of the invention, For the sampling point/>Soil investigation comprehensive error index of/(For the sampling point/>Error index of agricultural land risk zone,/>For the sampling point/>XRF quantitative limit error index of/(For the sampling point/>An XRF detection error index of (c).

However, when the actual laboratory point location is laid, the soil investigation comprehensive error index is [ ]) The larger samples should be screened as many as possible as laboratory test samples, so the least number of samples are weighted for e=1, 2, 3, respectively, as calculated by the following formula.

Wherein i represents soil investigation integrated error indexes E=1, 2 and 3 types of sample points; For/> The minimum sampling point number under the soil investigation comprehensive error index; /(I)For/>The number of laboratory sampling points under the soil investigation comprehensive error index; /(I)For the weight coefficient under the comprehensive error index of different soil investigation, the weight coefficient is determined according to the comprehensive error index of the soil investigation of a research area, so that the number of laboratory samples with E=3 is ensured to be more than E=2, the number of laboratory samples with E=1 is ensured to be more than E=2, and the preferred/>The numerical values are 1.5, 1.2 and 1.0 from high to low in sequence; /(I)For the corresponding standard normal deviation under a certain confidence level, the method can be inquired by a t distribution table,/>To a significant level, 90%, 95% are typically set; /(I)Is a degree of freedom; /(I)For the sample coefficient of variation,/>Is error,/>Can be obtained by prior data.

In order to facilitate quantitative screening of laboratory detection sampling points, the invention introduces similarity to characterize the similarity degree of environmental factors where two soil sampling points are located.

The laboratory sampling points are assumed to be located in a space with each environmental factor as an axis, i.e., a seven-dimensional space of soil type, matrix type, organic content, PH, powder content, clay content, and loam content (soil texture is typically characterized by powder content, clay content, loam content). Order theRepresents randomly selected/>Point set (Source Point set) of soil Point detected by each laboratory,/>Screening procedure Point set (candidate Point set-Source Point set) representing M samples,/>For the whole candidate point set, the number of the sample points is Q (Q=M+N),Attribute set representing two sample sets,/>. To avoid the impact of different sets of numbers on the computation, the continuous variables need to be normalized.

Similarity degreeThe similarity degree of the environmental attribute between the input point and the known point set is represented, and the similarity degree is higher when the attribute is similar; the larger the attribute gap is, the lower the similarity is, and the calculation formula is as follows:

In the method, in the process of the invention, The kth attribute value representing the ith point in the X set,/>The kth attribute value of the jth point in the Y set is represented, U is the total sample point set, t represents the quantitative attribute in the environmental factors in similarity calculation,/>Representing qualitative attributes in environmental factors in similarity calculation,/>The binary variable is a value of 1 when the attribute values are not equal, and a value of 0 when the condition is not satisfied. /(I)

In order to ensure the spatial uniformity of the sample points, the invention adopts Nearest Neighbor distance analysis (Nearest Neighbor DISTANCE ANALYSIS) criterion as a geospatial optimization objective function, wherein the criterion minimizes the distance from any point to the Nearest point, and the formula is as follows:

In the middle of To optimize the objective function, N is the number of monitoring points,/>Is at any point/>For monitoring points/>Nearest monitoring point,/>Is the euclidean distance of any point from its nearest point.

The multi-objective optimization is to improve the statistical inference precision of each index under the specific sampling purpose when the soil sampling point is selected, so as to form a more comprehensive and wide multi-objective space sampling optimization problem, and the invention adopts the similarity based on the simulated annealing algorithm) Spatial uniformity (/ >)) And two indexes are used for realizing screening of laboratory detection sample points.

The simulated annealing algorithm is a probability algorithm, and is derived from the simulation of the annealing process in thermodynamics, and the optimal solution of propositions is found in a large search space through slowly lowering the temperature parameters. The multi-objective simulated annealing algorithm optimizes the functions of multiple objectivesEach optimization target is used as a fitness function respectively, and a respective cooling path is set for each fitness function. The algorithm can be decomposed into four steps of solution space, objective function, initial solution, generation and acceptance of new solution, fig. 3 is a schematic flow chart of a simulated annealing algorithm of the multi-objective optimization function provided by the invention, and as shown in fig. 3, the steps are as follows:

(1) Initial sample set selection and optimization objective function setting: respectively optimizing all sample points with E not equal to 0 by laboratory initial sample point selection, namely randomly selecting N in a candidate set U ) X point sets of individual samples. Setting an optimization target and expressing in the form of a sample point function, namely/>. Optimization objective is to minimize/>The value of the function.

(2) Initializing an algorithm: setting respective initial temperature, end temperature and cooling rate for each optimization target，/>And other iteration termination conditions are set. Setting corresponding temperature variable/>, of each targetTaking the sample set in the step (1) as an initial sample s1 for the initial temperature, enabling the target sample set s=s1, and calculating/>Is a value of (2).

(3) Disturbance s, creating a new solution: adopting sample points in the point set Y, randomly selecting a sample point position in s, replacing the original sample point with a new sample point to generate a new sample set s2, and calculating an objective function corresponding to the s2Is a value of (2).

(4) Determining whether to accept the new solution according to the Metropolis criterion: the Metropolis criterion is that from the current state i to a new state j, if j new state energy is smaller than i current state energy, the new state j is used as a new current state, otherwise, the new state is accepted with a certain probability P, and the acceptance probability of each target to s2 is calculated by the following formula in turn:

generating a random number rand between 0 and 1, if rand < Target/>If not, judging that the judgment result is not acceptable. In addition, if s2 is accepted, let s=s2, and whether accepted or not, the temperature is lowered according to the rule in the following formula:

In the method, in the process of the invention, For different paths of initial temperature,/>Is the cooling rate of different paths.

(5) Judging convergence: selecting whether iteration is finished or not according to the fire lowering temperature and the sample point rejection times, and if the convergence condition is not met, continuing to execute the step (3); otherwise, ending the iteration and outputting s.

It should be noted that, each path of the simulated annealing algorithm of the multiple objective functions is independently selected to accept the new solution, and the probability of accepting the new solution depends on the improvement degree of the new solution to the objective functions and the temperature of the current path. If the new solution improves the objective function to a greater extent or the temperature is higher, a greater probability is accepted. The lower the temperature, the lower the probability of acceptance of the non-improvement solution, i.e. the less willing to sacrifice its own target improvement, to match the improvement of other targets under different objective function paths. Therefore, according to the cooling rule, when a new solution is accepted, each path is cooled synchronously, and when the new solution is rejected, only the path accepting the solution is selected for cooling, so that the temperature of the two paths is different along with the iteration. When a certain path receives a non-improved solution, the target improvement is not cooled to a certain extent, and if the path receives a new non-improved solution, the path is still searching for an optimal solution, and the cooling optimization needs to be continued.

In the embodiment of the invention, two targets of the environmental similarity and the spatial uniformity can be considered to have the same importance, so in some embodiments, the initial temperature can be set to 1, the cooling rate is set to 0.90, cooling judgment is performed in each iteration, the minimum temperature of the ending condition is set to 1×10 ^-15, the continuous minimum rejection number is set to 500, and in practical application, the setting of the ending condition can be performed through practical situations.

The method for screening the soil heavy metal laboratory detection sample points is further described through the embodiment in the specific application scene.

Fig. 4 is a flow chart of laboratory detection sample layout provided by the present invention, as shown in fig. 4, the flow mainly includes the following points:

(1) And (3) considering related factors such as polluted enterprises, soil types, matrix and mother rock in the range of the research area, and arranging XRF rapid detection sample points to obtain initial investigation data of the research area.

(2) Quantification describes XRF detection errors: and classifying the low, medium and high detection errors according to the influence of the content of the object to be detected on the XRF detection errors. Quantitatively describing threshold errors: and classifying the threshold errors according to the threshold range of the risk screening value and the risk tube control value of the heavy metal pollution index of the soil. Quantitatively describing quantitative limit errors: the quantification limit error classification is performed according to the XRF quantification limit.

(3) And comprehensively describing the three errors to form a comprehensive error, calculating reasonable laboratory detection sample points according to the regional concentration space variability of the object to be detected, the detection cost and the like, primarily selecting XRF+laboratory combined detection sample points, and forming a laboratory detection candidate set.

(4) And fully considering the environmental similarity and the spatial distribution uniformity of the soil sample points, and adopting a simulated annealing algorithm of a multi-objective optimization function to select the combined detection sample points from the laboratory detection candidate set so as to form an XRF+laboratory combined detection soil sample point investigation scheme.

The embodiment provides a laboratory detection sampling point optimizing and screening method based on soil investigation comprehensive error indexes, namely a laboratory detection sampling point layout method. By establishing a soil investigation comprehensive error index quantification model, selecting a certain number of sampling points, optimizing the sampling points according to the environmental influence factor conditions among the soil sampling points and the spatial distribution conditions of the sampling points, eliminating redundant sampling points, realizing the selection of laboratory detection sampling points, and achieving the purposes of reducing the soil sampling cost and improving the prediction accuracy of the heavy metal content of the soil.

Compared with the existing point distribution scheme, the embodiment fully considers the problem of laboratory detection cost, and provides a sample point distribution method for XRF combined laboratory detection. And (3) fully excavating the agricultural land risk region error, the quantitative limit error and the XRF detection error caused by XRF detection, and removing redundant points through a multi-path simulated annealing algorithm, so that laboratory detection of high sample points of the soil investigation comprehensive error index is realized. The soil sampling point layout scheme integrates the advantages of the two detection modes, provides technical support for efficient soil heavy metal pollutant detection sample point layout, also better solves the problem of high cost of soil general investigation under a large range of various points, and provides a feasible technical scheme for the wide application of realizing low-cost soil heavy metal detection and XRF rapid instrument quantitative detection in the future. The method is beneficial to popularization and application of the XRF soil detector in farmland soil heavy metal monitoring investigation and soil heavy metal restoration treatment investigation. Meanwhile, the detection point layout selection method can also be used in the practice of layout selection of the sampling points of monitoring and investigation of various ecological environments and natural resources such as other organic pollutants, soil nutrients, agriculture and forestry.

The soil heavy metal laboratory detection sampling point screening device provided by the invention is described below, and the soil heavy metal laboratory detection sampling point screening device described below and the soil heavy metal laboratory detection sampling point screening method described above can be correspondingly referred to each other.

Fig. 5 is a schematic structural diagram of a screening device for detecting sample points in a soil heavy metal laboratory, as shown in fig. 5, the device comprises:

The calculation module 500 is configured to calculate a soil investigation integrated error index of each soil sampling point based on the detection result of the heavy metal XRF to be detected of each soil sampling point in the research area; the soil investigation comprehensive error index comprises an agricultural land risk area error index, an XRF quantitative limit error index and an XRF detection error index;

the primary screening module 510 is configured to perform primary screening of laboratory samples based on the soil investigation comprehensive error indexes of the soil sampling points, so as to obtain a laboratory detection primary screening sample set;

The determining module 520 is configured to determine laboratory detection target sample points from the laboratory detection preliminary screening sample point set by using the multi-target simulated annealing algorithm with the environmental similarity minimized and the spatial distribution uniformity maximized as optimization targets.

Optionally, the XRF detection result is obtained by detecting each soil sampling point in the soil monitoring survey point arrangement scheme;

Soil monitoring survey point placement schemes are generated based on historical survey data or auxiliary environmental data.

Fig. 6 is a schematic structural diagram of an electronic device according to the present invention, as shown in fig. 6, the electronic device may include: processor 610, communication interface (Communications Interface) 620, memory 630, and communication bus 640, wherein processor 610, communication interface 620, memory 630 communicate with each other via communication bus 640. The processor 610 may invoke logic instructions in the memory 630 to perform the soil heavy metal laboratory test spot screening method provided by the methods described above, the method comprising:

Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which when executed by a processor is implemented to perform the soil heavy metal laboratory test sample screening method provided by the above methods, the method comprising:

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for screening detection sample points of a soil heavy metal laboratory, which is characterized by comprising the following steps:

2. The method for screening soil heavy metal laboratory detection samples according to claim 1, wherein calculating XRF detection error indexes of each soil sampling point based on XRF detection results of heavy metal to be detected of each soil sampling point in a research area comprises:

3. The method for screening soil heavy metal laboratory detection samples according to claim 2, wherein the calculating the XRF detection error index of each soil sampling point of the research area according to the naive bayes principle based on the prior knowledge and the research area XRF detection data specifically comprises:

4. The method for screening soil heavy metal laboratory detection samples according to claim 1, wherein calculating the agricultural land risk zone misdemarcation error index of each soil sampling point based on the XRF detection result of the heavy metal to be detected of each soil sampling point of the research area comprises:

5. The method for screening the detection sample points of the soil heavy metal laboratory according to claim 1, wherein the step of calculating the XRF quantitative limit error index of each soil sampling point based on the XRF detection result of the heavy metal to be detected of each soil sampling point of the research area comprises the following steps:

6. The method for screening laboratory detection sample points of soil heavy metals according to claim 1, wherein the method for determining laboratory detection target sample points from the laboratory detection primary screening sample point set by using a multi-target simulated annealing algorithm with the optimization objective of minimizing environmental similarity and maximizing spatial distribution uniformity comprises the following steps:

7. The method for screening soil heavy metal laboratory detection spots according to claim 1, wherein the XRF detection result is obtained by detecting each soil sampling spot in a soil monitoring survey spot arrangement scheme;

8. Soil heavy metal laboratory detects sample point sieving mechanism, its characterized in that, the device includes:

The primary screening module is used for carrying out primary screening based on the soil investigation comprehensive error indexes of the soil sampling points to obtain a laboratory detection primary screening sample point set;

9. An electronic device comprising a memory, a processor and a computer program stored on the memory and running on the processor, wherein the processor, when executing the program, implements the soil heavy metal laboratory test spot screening method of any one of claims 1 to 7.

10. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the soil heavy metal laboratory test spot screening method according to any one of claims 1 to 7.