CN113866204A - Bayesian regularization-based soil heavy metal quantitative analysis method - Google Patents
Bayesian regularization-based soil heavy metal quantitative analysis method Download PDFInfo
- Publication number
- CN113866204A CN113866204A CN202111132874.9A CN202111132874A CN113866204A CN 113866204 A CN113866204 A CN 113866204A CN 202111132874 A CN202111132874 A CN 202111132874A CN 113866204 A CN113866204 A CN 113866204A
- Authority
- CN
- China
- Prior art keywords
- neural network
- heavy metal
- soil
- regularization
- metal elements
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 229910001385 heavy metal Inorganic materials 0.000 title claims abstract description 39
- 239000002689 soil Substances 0.000 title claims abstract description 32
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000004445 quantitative analysis Methods 0.000 title claims abstract description 13
- 238000013528 artificial neural network Methods 0.000 claims abstract description 34
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000001228 spectrum Methods 0.000 claims abstract description 15
- 238000010606 normalization Methods 0.000 claims abstract description 6
- 238000012937 correction Methods 0.000 claims abstract description 4
- 238000002189 fluorescence spectrum Methods 0.000 claims abstract description 3
- 230000003595 spectral effect Effects 0.000 claims abstract description 3
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000011985 exploratory data analysis Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 description 8
- 238000004876 x-ray fluorescence Methods 0.000 description 8
- 239000011159 matrix material Substances 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 238000003062 neural network model Methods 0.000 description 5
- 238000011478 gradient descent method Methods 0.000 description 4
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 229910052785 arsenic Inorganic materials 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910052797 bismuth Inorganic materials 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 229910052802 copper Inorganic materials 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 231100001240 inorganic pollutant Toxicity 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 239000013077 target material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
- G01N23/22—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material
- G01N23/223—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material by irradiating the sample with X-rays or gamma-rays and by measuring X-ray fluorescence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N7/00—Computing arrangements based on specific mathematical models
- G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/07—Investigating materials by wave or particle radiation secondary emission
- G01N2223/076—X-ray fluorescence
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/10—Different kinds of radiation or particles
- G01N2223/101—Different kinds of radiation or particles electromagnetic radiation
- G01N2223/1016—X-ray
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2223/00—Investigating materials by wave or particle radiation
- G01N2223/60—Specific applications or type of materials
- G01N2223/616—Specific applications or type of materials earth materials
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Linguistics (AREA)
- Probability & Statistics with Applications (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Algebra (AREA)
- Mathematical Analysis (AREA)
- Molecular Biology (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Analysing Materials By The Use Of Radiation (AREA)
Abstract
The invention provides a Bayesian regularization-based quantitative analysis method for soil heavy metals, which is characterized in that discrete wavelets are used for denoising and background subtraction of an X fluorescence spectrum sample of a soil sample to obtain processed spectrum information; calculating the component information of the heavy metal elements and the preset interference elements in the processed spectral information by utilizing a Compton normalization method, respectively taking the component information of the heavy metal elements and the component information of the interference elements as the input of a BP (back propagation) neural network, and taking the actual content of the corresponding heavy metal elements as the output of a model; the hyper-parameters for completing the training of the BP neural network are determined through Bayesian regularization, and the BP neural network adopts an error function of regularization correction as a target function in the training process. The method provided by the invention can effectively improve the accuracy of quantitative analysis of the heavy metal elements in the soil, and has advantages compared with the traditional BP neural network method.
Description
Technical Field
The invention relates to an X-ray fluorescence spectrum analysis technology, in particular to a quantitative analysis technology based on Bayesian regularization.
Background
Soil is one of the basic elements of an ecosystem, is closely related to grain safety, ecological safety and human health, and soil heavy metals become an important component of inorganic pollutants of soil due to the reasons of easy accumulation, high treatment difficulty, poor chemical stability and the like. Therefore, the method improves the accurate quantitative analysis of the heavy metals in the soil and has important significance for guiding pollution control. Among the methods for detecting the elemental composition content, the X-ray fluorescence spectrum XRF is widely used in various fields because of its rapidity, low cost, and suitability for large-area monitoring.
Since the content of heavy metals to be detected is usually very low, while the traditional XRF method can be interfered by the background and the signals of the neighboring elements when predicting heavy metal elements in soil, the characteristic peak of the element is more easily submerged in the background noise. Compared with the traditional method for researching the relationship between the element component information and the content, the BP neural network has strong robustness on noise data and the capability of approaching any nonlinear relationship. Because of the advantages, the accurate determination of the heavy metal content in the soil by using artificial intelligence algorithms such as a neural network becomes a research hotspot in the XRF analysis field, and the research is mainly focused on two points: searching a proper neural network data modeling method; how to design the analysis process and further optimize the modeling method.
The existing standard BP neural network algorithm model has the problems of overfitting, low convergence speed, local convergence and the like, and the mean square error is adopted as a target function of the model, so that the analysis effect of the heavy metal content of the soil is poor, and the requirement on accuracy is difficult to meet.
Disclosure of Invention
The invention aims to provide a method for improving the accuracy of analysis of soil heavy metal elements in XRF.
The invention adopts the technical scheme that the soil heavy metal quantitative analysis method based on Bayesian regularization comprises the following steps:
1) denoising and background subtraction are carried out on the X fluorescence spectrum sample of the soil sample by using discrete wavelets to obtain processed spectrum information;
2) calculating the component information of heavy metal elements and preset interference elements in the processed spectrum information by using a Compton normalization method, respectively using the component information of the heavy metal elements and the interference elements as the input of a BP neural network,
and the actual content of the corresponding heavy metal element is used as the output of the model;
the hyper-parameters for completing the training of the BP neural network are determined by Bayesian regularization, and the BP neural network adopts an error function of regularization correction as a target function F (W) in the training process, wherein F (W) is beta ED+αEW；
EDAs the original error function:
EWfor the decay function:
alpha and beta are hyper-parameters, and are determined by a Bayesian algorithm:
m is the total number of parameters of the BP neural network, gamma is the number of valid parameters, WMPIs the minimum point where the objective function f (w) is zero in gradient.
The Bayesian regularization correction target function can improve the generalization capability of the neural network, and the introduction of Bayesian regularization has multiple advantages: (1) the effective value of the network is ensured to be as small as possible under the condition that the network training error is as small as possible, which is equivalent to automatically reducing the scale of the network; (2) under the condition that the size of the training sample set is certain, the scale of the neural network is far smaller than that of the training sample, so that the over-training opportunity is reduced, and the generalization capability is improved. The invention adopts Bayesian algorithm to determine the hyperparameter, so that the hyperparameter can be adjusted in a self-adaptive manner in the network training to achieve the optimal size.
The invention has the beneficial effects that: according to the invention, a quantitative analysis method based on Bayesian regularization is adopted, the BP neural network improved through Bayesian regularization quantitatively analyzes the heavy metal elements in XRF, and the Bayesian algorithm is adopted to determine the hyperparameters, so that the size of the hyperparameters can be adjusted in a self-adaptive manner in the training of the network, and the hyperparameters are optimized, and further the generalization capability of the neural network is improved. Experimental data show that the method provided by the invention can effectively improve the accuracy of quantitative analysis of the heavy metal elements in the soil, and has advantages compared with the traditional BP neural network method.
Drawings
FIG. 1 is a flow chart of an embodiment;
FIG. 2 is a spectrum of a soil sample in the example;
FIG. 3 is a comparison between the predicted content and the actual content of lead Pb by the Bayesian regularization-improved BP neural network model in the embodiment;
FIG. 4 is a comparison between the predicted content and the actual content of Pb by the BP neural network based on the gradient descent method in the example.
Detailed Description
Considering that the X-ray fluorescence spectrum of the soil has serious background interference, firstly, denoising and background subtraction are carried out on an obtained X-ray fluorescence spectrum sample by using discrete wavelets. And calculating and processing component information of heavy metal elements and other elements which can possibly generate interference by a Compton normalization method, wherein the component information is respectively used as the input of the BP neural network improved by Bayesian regularization, the actual content of the corresponding heavy metal elements is used as the output of the model, and the regularization is to improve the generalization capability of the BP neural network through a target function. The method can ensure that the effective value of the network is as small as possible under the condition that the network training error is as small as possible, which is equivalent to automatically reducing the scale of the network, and under the condition that the size of the training sample set is certain, the scale of the neural network is far smaller than that of the training sample, so that the occurrence probability of over-training is reduced, the generalization capability is improved, and the accurate analysis of the heavy metal content of the soil is realized.
Based on the above thought, the embodiment provides a quantitative analysis method based on bayesian regularization, the workflow is shown in fig. 1, and the specific steps are as follows:
step 1: collecting X-ray fluorescence spectrum information of soil samples by an X-ray fluorescence spectrometer, and collecting X-ray fluorescence spectrum information g of n soil samples i1,2, n, the spectrogram of the soil sample is shown in fig. 2;
step 2: the discrete wavelet transform is used for removing noise and deducting background from the spectrum information to obtain processed spectrum information g1i,i＝1,2,...,n；
And step 3: calculating and processing the composition information x of heavy metal elements and other several possibly interfering elements by a Compton normalization methodi,i＝1,2,...,n；
And 4, step 4: the samples were divided into N training samples and (N-N) prediction samples, for which D ═ xi,ti) N, heavy metal elements and several other elements that may interfere with each other, i 1,2iCarrying out data standardization on variable data serving as input nodes, and carrying out data standardization on actual content t of corresponding heavy metal elementsiAs an output of the model;
and 5: determining the BP neural network structure, and initializing hyper-parameters alpha, beta and weight. After the first training step, the parameters of an objective function F (W) are restored to the initial setting, and the established objective function comprises an error function and an attenuation function of the network:
in the formula, EDAs an error function of the network, EWAs a decay function, f (x)iW) is the output value of the network, tiThe actual value is N, the total number of training samples is W, the parameter vector of the neural network is W, and the hyper-parameters alpha and beta are the distribution form of the control connection weight and the threshold value.
Step 6: and minimizing an object function F (W) by using a Levenberg-Marquardt algorithm, wherein a network weight iteration formula used in the training process is as follows:
Wh+1＝Wh-[J(Wh)TJ(Wh)+μI]-1J(Wh)Te
in the formula, WhAnd f, the h iteration network weight vector, mu is an adaptive scalar, e is an error matrix, when mu is small, the Levenberg-Marquardt algorithm is changed into a Newton method, and when mu is large, the gradient descent method is adopted. The Levenberg-Marquardt algorithm is one of the existing optimization algorithms. The optimization is to find the parameter vector that minimizes the function value.
And 7: calculating the number gamma of effective parameters and the updated estimation values of the hyper-parameters alpha and beta of the objective function F (W);
wherein M is the number of the total network parameters, tr is the trace of the matrix,
is a vector differential operator, a Hessian matrix
Is obtained by approximation by a Gauss-Newton method, J is a Jacobian matrix of the training set sample errors,IMis EWA jacobian matrix. The optimal values of the hyper-parameters alpha and beta are determined by a Bayesian algorithm, namely, the objective function F (W) is subjected to the minimum value point W with zero gradientMPPerforming Taylor series expansion nearby, neglecting high-order terms, and performing Bayesian rule on WMPPoint optimization regularization parameters:
and 8: repeating 6 and 7 until convergence;
and step 9: and completing parameter training by using the Bayesian regularization improved BP neural network, and testing a network model by using a prediction sample.
The following embodiments are further described with reference to specific examples, namely quantitative analysis of the heavy metal Pb in the soil by using a Bayesian regularization improved BP neural network model. In order to ensure the stability of soil spectrum data, a soil mode is selected for the spectrometer within an allowable environment temperature, after the optimal test parameters are adjusted, the spectrometer works under the light tube voltage of 45kV and the light tube current of 25uA during the test, the peak forming time of the multi-channel acquisition system is set to be 0.8us, and the test time is 90 s. In order to eliminate the influence of errors and other environmental factors in the testing process, the same soil sample is tested for 3 times and the average value is taken to obtain the final spectral data of the corresponding sample.
The first step is as follows: sequentially collecting X-ray fluorescence spectrum data g of the 57 national standard soil samples i1, 2.., 57. As can be seen from the figure, due to the existence of redundant information such as noise and a substrate, the identification difficulty of the characteristic peak is increased, so that the accurate calculation of the peak area is influenced, and the precision of quantitative analysis is reduced. Therefore, a coif3 wavelet is used for denoising, and a sym4 wavelet is used for background subtraction.
The second step is that: the target material of the spectrum analyzer is an Ag target, so the Compton normalization method is adopted, namely the component information of heavy metal element Pb and other elements (As, Cu and Bi) which can interfere the heavy metal element Pb and the heavy metal element Pb, which are obtained by dividing the counting of the target element Pb by the Compton counting of the Ag peak, forms a 57 multiplied by 4 element component information matrix.
The third step: the method comprises the steps of randomly dividing element component information into 45 parts of training samples and 12 parts of prediction samples, taking component information of Pb elements and related elements in a training set as input of a Bayesian regularization improved BP neural network model, taking Pb element content as output, determining parameter values of the model through preliminary experiments, namely completing training, and testing the network model by using the prediction samples.
And 5-9, finally obtaining the comparison between the predicted content and the actual content of the heavy metals in the Pb soil by using the Bayesian regularized improved BP neural network model. As can be seen from fig. 3, the predicted value and the actual value in the test sample set of the Pb model have a better contact ratio, which indicates that the BP neural network model improved by bayes regularization has higher accuracy, and is suitable for determining the content of heavy metal elements in soil, thereby proving the effectiveness of the embodiment.
To further illustrate the superiority of the method of the present invention, fig. 4 shows the comparison of the predicted values and actual values in the test sample set using a conventional gradient descent method-based BP neural network. Comparing with fig. 3, it is clear that the example works better in the determination of the content of heavy metal element Pb in soil. Meanwhile, the running time of the embodiment and the traditional BP neural network method based on the gradient descent method is 2.446 seconds and 3.212 seconds respectively, and the embodiment is more efficient to calculate and is about 1.313 times of the latter.
Claims (1)
1. A soil heavy metal quantitative analysis method based on Bayesian regularization is characterized by comprising the following steps:
1) denoising and background subtraction are carried out on the X fluorescence spectrum sample of the soil sample by using discrete wavelets to obtain processed spectrum information;
2) calculating the component information of the heavy metal elements and the preset interference elements in the processed spectral information by utilizing a Compton normalization method, respectively taking the component information of the heavy metal elements and the component information of the interference elements as the input of a BP (back propagation) neural network, and taking the actual content of the corresponding heavy metal elements as the output of a model;
the hyper-parameters for completing the training of the BP neural network are determined by Bayesian regularization, and the BP neural network adopts an error function of regularization correction as a target function F (W) in the training process, wherein F (W) is beta ED+αEW；
EDAs the original error function:
EWfor the decay function:
alpha and beta are hyper-parameters, and are determined by a Bayesian algorithm:
m is the total number of parameters of the BP neural network, gamma is the number of valid parameters, WMPIs the minimum point where the objective function f (w) is zero in gradient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111132874.9A CN113866204A (en) | 2021-09-27 | 2021-09-27 | Bayesian regularization-based soil heavy metal quantitative analysis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111132874.9A CN113866204A (en) | 2021-09-27 | 2021-09-27 | Bayesian regularization-based soil heavy metal quantitative analysis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113866204A true CN113866204A (en) | 2021-12-31 |
Family
ID=78991029
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111132874.9A Pending CN113866204A (en) | 2021-09-27 | 2021-09-27 | Bayesian regularization-based soil heavy metal quantitative analysis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113866204A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115112699A (en) * | 2022-06-10 | 2022-09-27 | 电子科技大学 | XRF soil heavy metal element quantitative analysis method |
| CN116539647A (en) * | 2023-04-10 | 2023-08-04 | 电子科技大学 | Trace element content prediction method based on hierarchical deep neural network |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103425743A (en) * | 2013-07-17 | 2013-12-04 | 上海金自天正信息技术有限公司 | Steam pipe network prediction system based on Bayesian neural network algorithm |
| WO2016101182A1 (en) * | 2014-12-23 | 2016-06-30 | 清华大学 | Interval type indicator forecasting method based on bayesian network and extreme learning machine |
| JP2016186489A (en) * | 2016-06-01 | 2016-10-27 | コミシリア ア レネルジ アトミック エ オ エナジーズ オルタネティヴズ | Spectrometric analysis method and related device |
| CN107255647A (en) * | 2017-05-10 | 2017-10-17 | 中国科学院合肥物质科学研究院 | Microelement contents of soil analyzing and predicting method based on X-ray fluorescence spectra and depth confrontation study |
| CN109409614A (en) * | 2018-11-16 | 2019-03-01 | 国网浙江瑞安市供电有限责任公司 | A kind of Methods of electric load forecasting based on BR neural network |
| CN112748140A (en) * | 2020-12-23 | 2021-05-04 | 电子科技大学 | Trace element XRF determination method based on iterative discrete wavelet background subtraction |
| WO2021154747A1 (en) * | 2020-01-27 | 2021-08-05 | Lam Research Corporation | Performance predictors for semiconductor-manufacturing processes |
| CN114047212A (en) * | 2021-11-11 | 2022-02-15 | 中国科学院地质与地球物理研究所 | Method for in-situ nondestructive testing of stone source of stone cultural relics |
| US20220179121A1 (en) * | 2019-03-11 | 2022-06-09 | Schlumberger Technology Corporation | Estimating mineralogy and reconstructing elements of reservoir rock from spectroscopy data |
| CN114660104A (en) * | 2022-03-22 | 2022-06-24 | 台州高视通智能科技有限公司 | X-ray fluorescence analyzer device convenient for construction of Internet of things |
| US20220252744A1 (en) * | 2019-05-28 | 2022-08-11 | Commissariat A L'energie Atomique Et Aux Energies Alternatives | Method and device for identifying atomic species emitting x- or gamma radiation |
| CN115112699A (en) * | 2022-06-10 | 2022-09-27 | 电子科技大学 | XRF soil heavy metal element quantitative analysis method |
-
2021
- 2021-09-27 CN CN202111132874.9A patent/CN113866204A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103425743A (en) * | 2013-07-17 | 2013-12-04 | 上海金自天正信息技术有限公司 | Steam pipe network prediction system based on Bayesian neural network algorithm |
| WO2016101182A1 (en) * | 2014-12-23 | 2016-06-30 | 清华大学 | Interval type indicator forecasting method based on bayesian network and extreme learning machine |
| JP2016186489A (en) * | 2016-06-01 | 2016-10-27 | コミシリア ア レネルジ アトミック エ オ エナジーズ オルタネティヴズ | Spectrometric analysis method and related device |
| CN107255647A (en) * | 2017-05-10 | 2017-10-17 | 中国科学院合肥物质科学研究院 | Microelement contents of soil analyzing and predicting method based on X-ray fluorescence spectra and depth confrontation study |
| CN109409614A (en) * | 2018-11-16 | 2019-03-01 | 国网浙江瑞安市供电有限责任公司 | A kind of Methods of electric load forecasting based on BR neural network |
| US20220179121A1 (en) * | 2019-03-11 | 2022-06-09 | Schlumberger Technology Corporation | Estimating mineralogy and reconstructing elements of reservoir rock from spectroscopy data |
| US20220252744A1 (en) * | 2019-05-28 | 2022-08-11 | Commissariat A L'energie Atomique Et Aux Energies Alternatives | Method and device for identifying atomic species emitting x- or gamma radiation |
| WO2021154747A1 (en) * | 2020-01-27 | 2021-08-05 | Lam Research Corporation | Performance predictors for semiconductor-manufacturing processes |
| CN112748140A (en) * | 2020-12-23 | 2021-05-04 | 电子科技大学 | Trace element XRF determination method based on iterative discrete wavelet background subtraction |
| CN114047212A (en) * | 2021-11-11 | 2022-02-15 | 中国科学院地质与地球物理研究所 | Method for in-situ nondestructive testing of stone source of stone cultural relics |
| CN114660104A (en) * | 2022-03-22 | 2022-06-24 | 台州高视通智能科技有限公司 | X-ray fluorescence analyzer device convenient for construction of Internet of things |
| CN115112699A (en) * | 2022-06-10 | 2022-09-27 | 电子科技大学 | XRF soil heavy metal element quantitative analysis method |
Non-Patent Citations (3)
| Title |
|---|
| HZ CHENG 等: "Genetic algorithm-optimized BP neural network model for prediction of soil heavy metal content in XRF" * |
| 兰强太: "基于主成分分析和 BP 神经网络算法的综合选股实证研究" * |
| *** 等: "基于遗传算法的BP神经网络模型在岩心扫描仪测定海洋沉积物多种组分中的应用研究" * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115112699A (en) * | 2022-06-10 | 2022-09-27 | 电子科技大学 | XRF soil heavy metal element quantitative analysis method |
| CN116539647A (en) * | 2023-04-10 | 2023-08-04 | 电子科技大学 | Trace element content prediction method based on hierarchical deep neural network |
| CN116539647B (en) * | 2023-04-10 | 2024-01-23 | 电子科技大学 | Trace element content prediction method based on hierarchical deep neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113866204A (en) | Bayesian regularization-based soil heavy metal quantitative analysis method | |
| CN109884892B (en) | Process industrial system prediction model based on cross correlation time-lag grey correlation analysis | |
| CN108447057B (en) | SAR image change detection method based on significance and depth convolution network | |
| CN113687433A (en) | Bi-LSTM-based magnetotelluric signal denoising method and system | |
| CN117434153B (en) | Road nondestructive testing method and system based on ultrasonic technology | |
| CN110879927A (en) | Sea clutter amplitude statistical distribution field modeling method for sea target detection | |
| CN115640744A (en) | Method for predicting corrosion rate outside oil field gathering and transportation pipeline | |
| CN113126038A (en) | High-frequency ground wave radar working frequency optimization method, system, storage medium and application | |
| CN112149502A (en) | Unfavorable geology positioning forecasting method based on convolutional neural network | |
| CN112518425A (en) | Intelligent machining cutter wear prediction method based on multi-source sample migration reinforcement learning | |
| CN115796351A (en) | Rainfall shorthand prediction method and device based on variational modal decomposition and microwave attenuation | |
| CN115112699A (en) | XRF soil heavy metal element quantitative analysis method | |
| CN113988210A (en) | Method and device for restoring distorted data of structure monitoring sensor network and storage medium | |
| Zhang et al. | Real-time monitoring of the corrosion behaviour of the 304SS in HCl solution using BPNN with joint image recognition and electrochemical noise | |
| CN113743008A (en) | Fuel cell health prediction method and system | |
| CN117556198A (en) | Apparent resistivity denoising calculation method based on full-waveform well ground method | |
| CN110542748B (en) | Knowledge-based robust effluent ammonia nitrogen soft measurement method | |
| CN114839210B (en) | XRF element quantitative analysis method based on SSA-BP neural network | |
| CN115389743A (en) | Method, medium and system for predicting content interval of dissolved gas in transformer oil | |
| CN113868957B (en) | Residual life prediction and uncertainty quantitative calibration method under Bayes deep learning | |
| CN112616160B (en) | Intelligent short-wave frequency cross-frequency-band real-time prediction method and system | |
| CN115014451A (en) | Gas-solid two-phase flow measuring method of multi-network characteristic fusion model | |
| CN112801936B (en) | Self-adaptive background subtraction method for X-ray fluorescence spectrum | |
| CN113065498A (en) | Abnormal data detection method based on improved EMD and neural network model | |
| CN112032578B (en) | Method and equipment for measuring parameters of corrosion point in pipeline |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211231 |