Soil pollution assessment method and system based on artificial intelligence

Abstract

The invention relates to the technical field of soil pollution detection, and discloses a soil pollution assessment method and system based on artificial intelligence. The artificial intelligence-based soil pollution assessment method comprises the following steps: step S101, determining a sensitive wave band and a contribution degree value of a pollutant; s102, judging whether the pollutants are included in a pollutant list or not according to the second characteristic value; step S103, constructing an incidence matrix; step S104, correcting the hyperspectral data of the pollutants in the pollutant list; step S105, obtaining the concentration of the pollutants through the pollutant concentration prediction model; and S106, judging whether the soil is polluted or not according to the comprehensive pollution index. According to the method, the enhanced absorption effect between the pollutants is converted into incidence matrix representation by utilizing the characteristics of the sensitive wave bands of the pollutants in the hyperspectral data, and the precision of pollutant concentration measurement is improved by utilizing the nonlinear learning capability of the pollutant concentration prediction model.

Description

Technical Field

[0001] The present invention relates to the technical field of soil pollution detection, and more specifically, to a soil pollution assessment method and system based on artificial intelligence. Background Art

[0002] With the acceleration of industrialization and the increase in the use of fertilizers and pesticides, the concentration of soil pollutants has continued to rise, endangering the ecological environment and human health. At present, soil pollutant detection mostly relies on traditional manual field sampling and experimental analysis methods. Although these methods can provide accurate pollutant concentrations for local areas, they have problems such as poor timeliness, large workload and cumbersome measurement process. In addition, it is difficult to achieve large-scale and real-time soil pollutant concentration detection, resulting in an inability to fully assess the spread of soil pollution.

[0003] With the development of remote sensing technology, especially the emergence of hyperspectral data, the existing method is to extract the spectral reflectance related to soil pollutants in hyperspectral data through correlation coefficients, and then establish an inversion model between it and the measured value of pollutant concentration through polynomials, so as to achieve large-scale and real-time soil pollutant concentration detection. However, the above scheme does not take into account the enhanced absorption effect between pollutants, that is, when multiple pollutants coexist in the soil, the spectral absorption characteristics of these pollutants will interact with each other, resulting in the enhancement or suppression of the spectral reflectance of certain bands. For example, heavy metal elements will form complexes with organic matter, clay minerals, oxides, etc. in the soil, thereby changing the hyperspectral data and affecting the accuracy of pollutant concentration measurement. Summary of the invention

[0004] The present invention provides a soil pollution assessment method and system based on artificial intelligence, which solves the technical problem that the existing technology in the related art does not take into account the influence of the enhanced absorption effect between pollutants on hyperspectral data, resulting in low accuracy in pollutant concentration measurement.

[0005] The present invention provides a soil pollution assessment method based on artificial intelligence, comprising the following steps: Step S101, determining the sensitive bands of different pollutants in the soil to be detected in the hyperspectral data and the corresponding contribution values ​​according to the hyperspectral fingerprint library; The sum of the contribution values ​​of the same pollutant corresponding to different sensitive bands in the hyperspectral data is 1; The total number of sensitive bands M of each pollutant in the hyperspectral data in the hyperspectral fingerprint library is 3; Step S102, calculating a first eigenvalue based on the sensitive band of the pollutant in the hyperspectral data, and calculating a second eigenvalue based on the contribution value corresponding to the sensitive band, and if it is determined that the second eigenvalue is greater than or equal to a preset characteristic threshold, the pollutant is included in the pollutant list; Step S103, constructing a correlation matrix according to the peak spectral reflectance of the sensitive bands of all pollutants in the pollutant list in the hyperspectral data; The correlation matrix includes A rows and A columns, A represents the total number of pollutants in the pollutant list, and the element value of the i-th row and j-th column is represented by the three sets of weight coefficients and bias coefficients of the impact of the i-th pollutant on the j-th pollutant. When i=j, all weight coefficients and bias coefficients of the i-th row and j-th column are assigned 1 and 0 respectively, 1≤i≤A, 1≤j≤A; Step S104, correcting the sensitive bands of all pollutants in the pollutant list in the hyperspectral data according to the correlation matrix, and extracting the characteristic parameters corresponding to each pollutant; Step S105, inputting the characteristic parameters corresponding to each pollutant into the pollutant concentration prediction model, and outputting the concentration of each pollutant; Step S106, a comprehensive pollution index is calculated based on the concentration of each pollutant, and if it is greater than or equal to a preset pollution index threshold, it indicates that the soil to be tested is polluted.

[0006] Further, the starting wavelength spectral reflectance and the peak spectral reflectance of the sensitive band of the pollutant in the hyperspectral data are obtained, the band between the starting wavelength spectral reflectance and the peak spectral reflectance is equally spaced into N bands, and the first eigenvalue is obtained according to the spectral reflectance of the N bands, where N is a custom parameter; First eigenvalue The calculation formula is as follows: ; in and Respectively represent the starting wavelength and ending wavelength of the nth band, and They represent the spectral reflectance of the starting wavelength and the ending wavelength of the nth band respectively. and They represent the peak spectral reflectance and standard spectral reflectance of the sensitive band respectively, is a custom parameter.

[0007] Furthermore, the second eigenvalue The calculation formula is as follows: ; in represents the first eigenvalue corresponding to the mth sensitive band, Indicates the contribution value corresponding to the mth sensitive band.

[0008] Furthermore, constructing the correlation matrix includes the following steps: Step S201, randomly selecting H pollutants from the hyperspectral fingerprint library, and randomly generating concentration values ​​of each pollutant within upper and lower limits to form a soil composite; Among them, H and the upper and lower limits of each pollutant are custom parameters; Step S202, obtaining the peak spectral reflectance of the sensitive bands of H pollutants in the soil composite in the hyperspectral data as sample data of the training sample; Step S203, respectively obtaining peak spectral reflectances of sensitive bands of H individual pollutants in the hyperspectral data as sample labels of training samples; Step S204, repeating steps S201 to S203 until a training data set consisting of G training samples is obtained, and the spectral correction model is trained by the training data set; Where G is a custom parameter; Step S205, respectively input the peak spectral reflectance of the sensitive band of each pollutant in the pollutant list in the hyperspectral data into the trained spectral correction model, and output the weight coefficient and bias coefficient of the influence of other pollutants on the pollutant to form a correlation matrix.

[0009] Furthermore, the calculation formula of the spectral correction model is as follows: ; in and They represent the peak spectral reflectance of the hth pollutant in the mth sensitive band in the hyperspectral data before and after correction, and They represent the weight coefficient and bias coefficient of the h-th pollutant in the m-th sensitive band in the hyperspectral data respectively; The calculation formula of the loss function Loss of the spectral correction model is as follows: ; in It represents the peak spectral reflectance of the hth pollutant corresponding to the sample data of the gth training sample after correction in the mth sensitive band of the hyperspectral data through the spectral correction model. Represents the peak spectral reflectance of the h-th pollutant in the m-th sensitive band in the hyperspectral data corresponding to the sample label of the g-th training sample.

[0010] Furthermore, the starting wavelength spectral reflectance of the sensitive band of the pollutant in the hyperspectral data and the corrected peak spectral reflectance are divided into K bands at equal intervals, and the normalized reflectance index of each band is calculated as a characteristic parameter, where K is a custom parameter; Normalized reflectance index of the kth band of the mth sensitive band after correction The calculation formula is as follows: ,in and They respectively represent the starting wavelength spectral reflectance and the ending wavelength spectral reflectance of the kth band of the mth sensitive band after correction, 1≤m≤M=3.

[0011] Furthermore, the pollutant concentration prediction model is composed of M hidden layers, each hidden layer includes K hidden units, the kth hidden unit of the mth hidden layer inputs the normalized reflectance index of the kth band of the mth sensitive band corrected in the feature parameter, and outputs an update vector, wherein the number of dimensions of the update vector is a custom parameter; The update vector output by the Kth hidden unit of the Mth hidden layer is input into the classifier, whose category space represents the concentration of the pollutant.

[0012] Furthermore, the calculation formula of the pollutant concentration prediction model includes: ; ; ; ; ; in represents the update vector for the output of the kth hidden unit in the mth hidden layer, represents the update vector for the output of the k-1th hidden unit in the mth hidden layer, represents the update vector for the output of the kth hidden unit in the m-1th hidden layer, Represents the normalized reflectance index of the kth band of the mth sensitive band corrected in the characteristic parameter input to the kth hidden unit of the mth hidden layer, , , and denote the reset vector, update gate value, intermediate conversion vector and hidden vector of the kth hidden unit of the mth hidden layer, respectively. , , and denote the first, second, third, and fourth weight parameters of the kth hidden unit of the mth hidden layer, respectively. , , and denote the first, second, third, and fourth bias parameters of the kth hidden unit of the mth hidden layer, respectively. Represents point-by-point multiplication, sigmoid represents the sigmoid activation function, Swish represents the Swish activation function, and Mish represents the Mish activation function.

[0013] Furthermore, the calculation formula of the comprehensive pollution index comp is as follows: ; Where D represents the total number of pollutants in the pollutant inventory, , and They respectively represent the concentration, standard concentration and weight coefficient of the dth pollutant in the pollutant list. E represents the customized default concentration, where the standard concentration and weight coefficient of the pollutant are both customized parameters. Max represents the maximum value.

[0014] The present invention provides a soil pollution assessment system based on artificial intelligence, comprising: The first module is used to determine the sensitive bands and corresponding contribution values ​​of different pollutants in the soil to be detected in the hyperspectral data according to the hyperspectral fingerprint library; The second module is used to calculate the first eigenvalue according to the sensitive band of the pollutant in the hyperspectral data, and calculate the second eigenvalue in combination with the contribution value corresponding to the sensitive band, and if it is greater than or equal to the preset characteristic threshold, the pollutant is included in the pollutant list; The third module is used to construct a correlation matrix according to the peak spectral reflectance of the sensitive bands of all pollutants in the pollutant inventory in the hyperspectral data; The fourth module is used to correct the sensitive bands of all pollutants in the pollutant list in the hyperspectral data according to the correlation matrix, and extract the characteristic parameters corresponding to each pollutant; The fifth module is used to input the characteristic parameters corresponding to each pollutant into the pollutant concentration prediction model and output the concentration of each pollutant; The sixth module is used to calculate the comprehensive pollution index according to the concentration of each pollutant, and determine whether it is greater than or equal to a preset pollution index threshold, which indicates that the soil to be tested is polluted.

[0015] The beneficial effects of the present invention are as follows: the present invention utilizes the characteristics of the sensitive bands of pollutants in hyperspectral data to convert the enhanced absorption effect between pollutants into a correlation matrix representation, corrects the hyperspectral data of pollutants through the correlation matrix, and utilizes the nonlinear learning ability of the pollutant concentration prediction model to achieve mapping with the pollutant concentration, thereby improving the accuracy of pollutant concentration measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1is a flow chart of a soil pollution assessment method based on artificial intelligence of the present invention; Figure 2 is a flow chart of constructing an association matrix of the present invention; Figure 3 is a schematic diagram of a soil pollution assessment system based on artificial intelligence of the present invention; Figure 4 It is a comparison chart of lead concentration detection of the present invention; Figure 5 It is a comparison diagram of cadmium concentration detection of the present invention; Figure 6 This is a comparison chart of perfluorooctane sulfonic acid concentration detection of the present invention.

[0017] In the figure: a first module 301 , a second module 302 , a third module 303 , a fourth module 304 , a fifth module 305 , and a sixth module 306 . DETAILED DESCRIPTION

[0018] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0019] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing in front of the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] like Figure 1 to Figure 6 As shown, a soil pollution assessment method based on artificial intelligence includes the following steps: Step S101, determining the sensitive bands of different pollutants in the soil to be detected in the hyperspectral data and the corresponding contribution values ​​according to the hyperspectral fingerprint library; The sum of the contribution values ​​of the same pollutant corresponding to different sensitive bands in the hyperspectral data is 1; The total number of sensitive bands M of each pollutant in the hyperspectral data in the hyperspectral fingerprint library is 3; Step S102, calculating a first eigenvalue based on the sensitive band of the pollutant in the hyperspectral data, and calculating a second eigenvalue based on the contribution value corresponding to the sensitive band, and if it is determined that the second eigenvalue is greater than or equal to a preset characteristic threshold, the pollutant is included in the pollutant list; Step S103, constructing a correlation matrix according to the peak spectral reflectance of the sensitive bands of all pollutants in the pollutant list in the hyperspectral data; The correlation matrix includes A rows and A columns, A represents the total number of pollutants in the pollutant list, and the element value of the i-th row and j-th column is represented by the three sets of weight coefficients and bias coefficients of the impact of the i-th pollutant on the j-th pollutant. When i=j, all weight coefficients and bias coefficients of the i-th row and j-th column are assigned 1 and 0 respectively, 1≤i≤A, 1≤j≤A; Step S104, correcting the sensitive bands of all pollutants in the pollutant list in the hyperspectral data according to the correlation matrix, and extracting the characteristic parameters corresponding to each pollutant; Step S105, inputting the characteristic parameters corresponding to each pollutant into the pollutant concentration prediction model, and outputting the concentration of each pollutant; Step S106, a comprehensive pollution index is calculated based on the concentration of each pollutant, and if it is greater than or equal to a preset pollution index threshold, it indicates that the soil to be tested is polluted.

[0021] It should be noted that hyperspectral data are represented by the spectral reflectance corresponding to the band, and the sensitive bands of different pollutants in the hyperspectral data in the hyperspectral fingerprint library and the contribution values ​​corresponding to the sensitive bands are all custom parameters, where the sensitive bands represent the bands with the most significant spectral reflectance of the pollutants in the hyperspectral data, and the contribution values ​​represent the relative importance of the spectral reflectance corresponding to different sensitive bands to the identification of pollutants. For example, lead (Pb) includes three sensitive bands, namely 530-550nm, 830-850nm and 1660-1700nm, and the corresponding contribution values ​​are set to 0.25, 0.45 and 0.3 respectively; for another example, cadmium (Cd) includes three sensitive bands, namely 520-540nm, 910-930nm and 2250-2350nm, and the corresponding contribution values ​​are set to 0.2, 0.5 and 0.3 respectively.

[0022] It should be noted that the present invention first uses the hyperspectral fingerprint library to determine the type of pollutant in the soil to be detected, and then corrects the sensitive band of the pollutant in the hyperspectral data (the spectral reflectance is corrected), and finally realizes the pollutant concentration measurement through the pollutant concentration prediction model. In addition, in order to improve the accuracy of the measurement, a pollutant type can be set to correspond to a pollutant concentration prediction model, but the structure and calculation formula of all pollutant concentration prediction models are the same.

[0023] In one embodiment of the present invention, the starting wavelength spectral reflectance and the peak spectral reflectance of the sensitive band of the pollutant in the hyperspectral data are obtained, the band between the starting wavelength spectral reflectance and the peak spectral reflectance is equally divided into N bands, and the first eigenvalue is calculated according to the spectral reflectance of the N bands, where N is a custom parameter, for example, N is set to 20; First eigenvalue The calculation formula is as follows: ; in and Respectively represent the starting wavelength and ending wavelength of the nth band, and They represent the spectral reflectance of the starting wavelength and the ending wavelength of the nth band respectively. and They represent the peak spectral reflectance and standard spectral reflectance of the sensitive band respectively, For custom parameters, such as Set to 100.

[0024] In one embodiment of the present invention, the second eigenvalue The calculation formula is as follows: ; in represents the first eigenvalue corresponding to the mth sensitive band, Indicates the contribution value corresponding to the mth sensitive band.

[0025] It should be noted that a sensitive band of a pollutant in the hyperspectral data corresponds to a first eigenvalue, and the calculation formulas for the first eigenvalue and the second eigenvalue do not involve dimension calculations, and the preset characteristic threshold is also a custom parameter. The second eigenvalue can be calculated based on the sensitive band of pollutants with low concentrations (for example, 10% of the pollution standard) in the hyperspectral data as the preset characteristic threshold.

[0026] In one embodiment of the present invention, Figure 2 As shown, constructing the association matrix includes the following steps: Step S201, randomly selecting H pollutants from the hyperspectral fingerprint library, and randomly generating concentration values ​​of each pollutant within upper and lower limits to form a soil composite; Among them, H and the upper and lower limits of each pollutant are custom parameters; Step S202, obtaining the peak spectral reflectance of the sensitive bands of H pollutants in the soil composite in the hyperspectral data as sample data of the training sample; Step S203, respectively obtaining peak spectral reflectances of sensitive bands of H individual pollutants in the hyperspectral data as sample labels of training samples; Step S204, repeating steps S201 to S203 until a training data set consisting of G training samples is obtained, and the spectral correction model is trained by the training data set; Where G is a custom parameter; Step S205, respectively input the peak spectral reflectance of the sensitive band of each pollutant in the pollutant list in the hyperspectral data into the trained spectral correction model, and output the weight coefficient and bias coefficient of the influence of other pollutants on the pollutant to form a correlation matrix.

[0027] It should be noted that the number of pollutants selected from the hyperspectral fingerprint library each time may not be a fixed value. The upper and lower limits are set for the concentration value of each pollutant to ensure the rationality of random generation. The purpose of the spectral correction model is to construct a mapping relationship between the peak spectral reflectance of multiple pollutants in the sensitive bands of hyperspectral data and the peak spectral reflectance of a single pollutant in the sensitive bands of hyperspectral data through learnable weight coefficients and bias coefficients, thereby achieving the effect of hyperspectral correction. In addition, the larger the number of training samples G, the better the training effect.

[0028] In one embodiment of the present invention, the calculation formula of the spectrum correction model is as follows: ; in and They represent the peak spectral reflectance of the hth pollutant in the mth sensitive band in the hyperspectral data before and after correction, and They represent the weight coefficient and bias coefficient of the h-th pollutant in the m-th sensitive band in the hyperspectral data respectively; The calculation formula of the loss function Loss of the spectral correction model is as follows: ; in It represents the peak spectral reflectance of the hth pollutant corresponding to the sample data of the gth training sample after correction in the mth sensitive band of the hyperspectral data through the spectral correction model. Represents the peak spectral reflectance of the h-th pollutant in the m-th sensitive band in the hyperspectral data corresponding to the sample label of the g-th training sample.

[0029] It should be noted that the gradient optimizer is used to reversely update the weight coefficients and bias coefficients in the spectral correction model, so as to minimize the loss value calculated forward using the loss function. Commonly used gradient optimizers include AMSGrad, AdaBelief, etc. In addition, the method of correcting the sensitive bands of all pollutants in the pollutant list in the hyperspectral data according to the correlation matrix is ​​the same as the calculation formula of the spectral correction model.

[0030] In one embodiment of the present invention, the starting wavelength spectral reflectance of the sensitive band of the pollutant in the hyperspectral data and the corrected peak spectral reflectance are divided into K bands at equal intervals, and the normalized reflectance index of each band is calculated as a characteristic parameter, where K is a custom parameter, for example, K is set to 10; Normalized reflectance index of the kth band of the mth sensitive band after correction The calculation formula is as follows: ,in and They respectively represent the starting wavelength spectral reflectance and the ending wavelength spectral reflectance of the kth band of the mth sensitive band after correction, 1≤m≤M=3.

[0031] It should be noted that the characteristic parameters are represented in the form of a matrix of 3 rows and K columns, that is, the rows of the matrix correspond to 3 sensitive bands, and the columns of the matrix correspond to K bands.

[0032] In one embodiment of the present invention, the pollutant concentration prediction model is composed of M hidden layers, each hidden layer includes K hidden units, the kth hidden unit of the mth hidden layer inputs the normalized reflectance index of the kth band of the mth sensitive band corrected in the feature parameter, and outputs an update vector, wherein the number of dimensions of the update vector is a custom parameter; The update vector output by the Kth hidden unit of the Mth hidden layer is input into the classifier, whose category space represents the concentration of the pollutant.

[0033] In one embodiment of the present invention, the calculation formula of the pollutant concentration prediction model includes: ; ; ; ; ; in represents the update vector for the output of the kth hidden unit in the mth hidden layer, represents the update vector for the output of the k-1th hidden unit in the mth hidden layer, represents the update vector for the output of the kth hidden unit in the m-1th hidden layer, Represents the normalized reflectance index of the kth band of the mth sensitive band corrected in the characteristic parameter input to the kth hidden unit of the mth hidden layer, , , and denote the reset vector, update gate value, intermediate conversion vector and hidden vector of the kth hidden unit of the mth hidden layer, respectively. , , and denote the first, second, third, and fourth weight parameters of the kth hidden unit of the mth hidden layer, respectively. , , and denote the first, second, third, and fourth bias parameters of the kth hidden unit of the mth hidden layer, respectively. Represents point-by-point multiplication, sigmoid represents the sigmoid activation function, Swish represents the Swish activation function, and Mish represents the Mish activation function.

[0034] It should be noted that the number of dimensions of the update vector can be set to 16, so the number of dimensions of the intermediate conversion vector and the hidden vector must also be 16, and the update gating value is a real value, then the first weight parameter needs to be designed as a matrix of (1+16)×1, and the first bias parameter can be designed as a real value. The second weight parameter needs to be designed as a matrix of (16+16)×16, then the second bias parameter needs to be designed as a vector of 1×16, and the third weight parameter needs to be designed as a matrix of 1×16. The number of dimensions of the reset vector is 16, the third bias parameter needs to be designed as a vector of 1×16, the fourth weight parameter needs to be designed as a matrix of (1+16)×16, and the fourth bias parameter needs to be designed as a vector of 1×16.

[0035] In one embodiment of the present invention, the calculation formula of the comprehensive pollution index comp is as follows: ; Where D represents the total number of pollutants in the pollutant inventory, , and They respectively represent the concentration, standard concentration and weight coefficient of the dth pollutant in the pollutant list. E represents the customized default concentration. For example, E is set to 0.01, where the standard concentration and weight coefficient of the pollutant are both customized parameters. max represents the maximum value.

[0036] It should be noted that the calculation formula of the comprehensive pollution index does not involve dimensional calculations, and the preset pollution index threshold is a custom parameter. In addition, the concentration of each pollutant can be directly compared with the corresponding standard concentration. If it is greater than the standard concentration, an alarm message will be generated. I will not go into details here.

[0037] In one embodiment of the present invention, Figure 3 As shown, the present invention provides a soil pollution assessment system based on artificial intelligence, comprising: The first module 301 is used to determine the sensitive bands and corresponding contribution values ​​of different pollutants in the soil to be detected in the hyperspectral data according to the hyperspectral fingerprint library; The second module 302 is used to calculate the first characteristic value according to the sensitive band of the pollutant in the hyperspectral data, and calculate the second characteristic value in combination with the contribution value corresponding to the sensitive band, and if it is greater than or equal to a preset characteristic threshold, the pollutant is included in the pollutant list; The third module 303 is used to construct a correlation matrix according to the peak spectral reflectance of the sensitive bands of all pollutants in the pollutant list in the hyperspectral data; The fourth module 304 is used to correct the sensitive bands of all pollutants in the pollutant list in the hyperspectral data according to the correlation matrix, and extract the characteristic parameters corresponding to each pollutant; The fifth module 305 is used to input the characteristic parameters corresponding to each pollutant into the pollutant concentration prediction model and output the concentration of each pollutant; The sixth module 306 is used to calculate the comprehensive pollution index according to the concentration of each pollutant, and determine whether it is greater than or equal to a preset pollution index threshold, which indicates that the soil to be tested is polluted.

[0038] In one embodiment of the present invention, Figure 4 As shown, 10 groups of lead concentrations (unit: mg / kg) were detected and compared using the soil pollutant concentration detection method (spectral correction model and pollutant concentration prediction model) provided by the present invention and the inversion model of the prior art (the polynomial model mentioned in the background technology).

[0039] In one embodiment of the present invention, Figure 5As shown, 10 groups of cadmium concentrations (unit: mg / kg) were detected and compared using the soil pollutant concentration detection method (spectral correction model and pollutant concentration prediction model) provided by the present invention and the inversion model of the prior art (the polynomial model mentioned in the background technology).

[0040] It should be noted that the concept of the present invention can also be used to measure the concentration of other compounds in the soil, such as perfluoroalkyl and polyfluoroalkyl compounds (PFAS), which are widely used in various consumer and industrial products due to their excellent thermal stability and chemical stability. Common PFAS include perfluorooctanoic acid (PFOA), perfluorooctane sulfonic acid (PFOS), etc.

[0041] In one embodiment of the present invention, Figure 6 As shown, 10 groups of PFOS concentrations (unit: μg / kg) were detected and compared using the soil pollutant concentration detection method (spectral correction model and pollutant concentration prediction model) provided by the present invention and the inversion model of the prior art (the polynomial model mentioned in the background technology).

[0042] It should be noted that, through observation, it is found that the soil pollutant concentration detection method provided by the present invention has higher detection accuracy.

[0043] It should be noted that the interval and threshold size are set for the convenience of comparison, where the size of the threshold depends on the amount of sample data and the number of bases set by technicians in this field for each group of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. In addition, the above formulas are all calculations of removing dimensions and taking their values. The formulas are all obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0044] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.

Claims

1. A soil pollution assessment method based on artificial intelligence, characterized in that: The following steps are involved: Step S101, determining the sensitive bands of different pollutants in the soil to be detected in the hyperspectral data and the corresponding contribution values ​​according to the hyperspectral fingerprint library; The sum of the contribution values ​​of the same pollutant corresponding to different sensitive bands in the hyperspectral data is 1; The total number of sensitive bands M of each pollutant in the hyperspectral data in the hyperspectral fingerprint library is 3; Step S102, calculating a first eigenvalue based on the sensitive band of the pollutant in the hyperspectral data, and calculating a second eigenvalue based on the contribution value corresponding to the sensitive band, and if it is determined that the second eigenvalue is greater than or equal to a preset characteristic threshold, the pollutant is included in the pollutant list; Step S103, constructing a correlation matrix according to the peak spectral reflectance of the sensitive bands of all pollutants in the pollutant list in the hyperspectral data; The correlation matrix includes A rows and A columns, A represents the total number of pollutants in the pollutant list, and the element value of the i-th row and j-th column is represented by the three sets of weight coefficients and bias coefficients of the impact of the i-th pollutant on the j-th pollutant. When i=j, all weight coefficients and bias coefficients of the i-th row and j-th column are assigned 1 and 0 respectively, 1≤i≤A, 1≤j≤A; Step S104, correcting the sensitive bands of all pollutants in the pollutant list in the hyperspectral data according to the correlation matrix, and extracting the characteristic parameters corresponding to each pollutant; Step S105, inputting the characteristic parameters corresponding to each pollutant into the pollutant concentration prediction model, and outputting the concentration of each pollutant; Step S106, a comprehensive pollution index is calculated based on the concentration of each pollutant, and if it is greater than or equal to a preset pollution index threshold, it indicates that the soil to be tested is polluted.

2. The soil pollution assessment method based on artificial intelligence according to claim 1 is characterized in that: Obtain the starting wavelength spectral reflectance and the peak spectral reflectance of the sensitive band of the pollutant in the hyperspectral data, divide the band between the starting wavelength spectral reflectance and the peak spectral reflectance into N bands at equal intervals, and calculate the first eigenvalue according to the spectral reflectance of the N bands, where N is a custom parameter; First eigenvalue The calculation formula is as follows: ； in and Respectively represent the starting wavelength and ending wavelength of the nth band, and They represent the spectral reflectance of the starting wavelength and the ending wavelength of the nth band respectively. and They represent the peak spectral reflectance and standard spectral reflectance of the sensitive band respectively, is a custom parameter.

3. The soil pollution assessment method based on artificial intelligence according to claim 1 is characterized in that: The second eigenvalue The calculation formula is as follows: ； in represents the first eigenvalue corresponding to the mth sensitive band, Indicates the contribution value corresponding to the mth sensitive band.

4. The soil pollution assessment method based on artificial intelligence according to claim 1 is characterized in that: Constructing the correlation matrix includes the following steps: Step S201, randomly selecting H pollutants from the hyperspectral fingerprint library, and randomly generating concentration values ​​of each pollutant within upper and lower limits to form a soil composite; Among them, H and the upper and lower limits of each pollutant are custom parameters; Step S202, obtaining the peak spectral reflectance of the sensitive bands of H pollutants in the soil composite in the hyperspectral data as sample data of the training sample; Step S203, respectively obtaining peak spectral reflectances of sensitive bands of H individual pollutants in the hyperspectral data as sample labels of training samples; Step S204, repeating steps S201 to S203 until a training data set consisting of G training samples is obtained, and the spectral correction model is trained by the training data set; Where G is a custom parameter; Step S205, respectively input the peak spectral reflectance of the sensitive band of each pollutant in the pollutant list in the hyperspectral data into the trained spectral correction model, and output the weight coefficient and bias coefficient of the influence of other pollutants on the pollutant to form a correlation matrix.

5. The soil pollution assessment method based on artificial intelligence according to claim 4 is characterized in that: The calculation formula of the spectral correction model is as follows: ； in and They represent the peak spectral reflectance of the hth pollutant in the mth sensitive band in the hyperspectral data before and after correction, and They represent the weight coefficient and bias coefficient of the h-th pollutant in the m-th sensitive band in the hyperspectral data respectively; The calculation formula of the loss function Loss of the spectral correction model is as follows: ； in It represents the peak spectral reflectance of the hth pollutant corresponding to the sample data of the gth training sample after correction in the mth sensitive band in the hyperspectral data through the spectral correction model. Represents the peak spectral reflectance of the h-th pollutant in the m-th sensitive band in the hyperspectral data corresponding to the sample label of the g-th training sample.

6. The soil pollution assessment method based on artificial intelligence according to claim 1 is characterized in that: The starting wavelength spectral reflectance of the sensitive band of the pollutant in the hyperspectral data and the corrected peak spectral reflectance are divided into K bands at equal intervals, and the normalized reflectance index of each band is calculated as the characteristic parameter, where K is a custom parameter; Normalized reflectance index of the kth band of the mth sensitive band after correction The calculation formula is as follows: ,in and They respectively represent the starting wavelength spectral reflectance and the ending wavelength spectral reflectance of the kth band of the mth sensitive band after correction, 1≤m≤M=3.

7. The soil pollution assessment method based on artificial intelligence according to claim 6 is characterized in that: The pollutant concentration prediction model consists of M hidden layers, each hidden layer includes K hidden units, the kth hidden unit of the mth hidden layer inputs the normalized reflectance index of the kth band of the mth sensitive band corrected in the feature parameters, and outputs an update vector, where the number of dimensions of the update vector is a custom parameter; The update vector output by the Kth hidden unit of the Mth hidden layer is input into the classifier, whose category space represents the concentration of the pollutant.

8. The soil pollution assessment method based on artificial intelligence according to claim 7 is characterized in that: The calculation formula of the pollutant concentration prediction model includes: ； ； ； ； ； in represents the update vector for the output of the kth hidden unit in the mth hidden layer, represents the update vector for the output of the k-1th hidden unit in the mth hidden layer, represents the update vector for the output of the kth hidden unit in the m-1th hidden layer, Represents the normalized reflectance index of the kth band of the mth sensitive band corrected in the characteristic parameter input to the kth hidden unit of the mth hidden layer, , , and denote the reset vector, update gate value, intermediate conversion vector and hidden vector of the kth hidden unit of the mth hidden layer, respectively. , , and denote the first, second, third, and fourth weight parameters of the kth hidden unit of the mth hidden layer, respectively. , , and denote the first, second, third, and fourth bias parameters of the kth hidden unit of the mth hidden layer, respectively. Represents point-by-point multiplication, sigmoid represents the sigmoid activation function, Swish represents the Swish activation function, and Mish represents the Mish activation function.

9. The soil pollution assessment method based on artificial intelligence according to claim 1 is characterized in that: The calculation formula of the comprehensive pollution index comp is as follows: ； Where D represents the total number of pollutants in the pollutant inventory, , and They respectively represent the concentration, standard concentration and weight coefficient of the dth pollutant in the pollutant list. E represents the customized default concentration, where the standard concentration and weight coefficient of the pollutant are both customized parameters. Max represents the maximum value.

10. A soil pollution assessment system based on artificial intelligence, characterized in that: Executing a soil pollution assessment method based on artificial intelligence as claimed in any one of claims 1 to 9, comprising: The first module is used to determine the sensitive bands and corresponding contribution values ​​of different pollutants in the soil to be detected in the hyperspectral data according to the hyperspectral fingerprint library; The second module is used to calculate the first eigenvalue according to the sensitive band of the pollutant in the hyperspectral data, and calculate the second eigenvalue in combination with the contribution value corresponding to the sensitive band, and if it is greater than or equal to the preset characteristic threshold, the pollutant is included in the pollutant list; The third module is used to construct a correlation matrix according to the peak spectral reflectance of the sensitive bands of all pollutants in the pollutant inventory in the hyperspectral data; The fourth module is used to correct the sensitive bands of all pollutants in the pollutant list in the hyperspectral data according to the correlation matrix, and extract the characteristic parameters corresponding to each pollutant; The fifth module is used to input the characteristic parameters corresponding to each pollutant into the pollutant concentration prediction model and output the concentration of each pollutant; The sixth module is used to calculate the comprehensive pollution index according to the concentration of each pollutant, and determine whether it is greater than or equal to a preset pollution index threshold, which indicates that the soil to be tested is polluted.

Images