Research On XRF Spectral Overlapping Peak Analysis And Content Inversion Model Of Soil Heavy Metals

With the development of human economy and society,soil heavy metal pollution has become more and more serious.Soil heavy metal pollution seriously threatens human living environment and human health,so the detection of soil heavy metals is very important.The traditional detection method has cumbersome steps and poor timeliness,which is not conducive to rapid quantitative detection of soil heavy metals.X-ray fluorescence spectroscopy(XRF)can realize fast online detection of soil heavy metals due to its simple operation,fast detection speed,and no secondary pollution.Based on the analysis and research of soil heavy metal detection data using energy dispersive X-ray analyzer,the main research work is as follows:Firstly,starting from the theoretical basis of studying X-ray physics,the inevitable relationship between the intensity of X-ray spectrum and the content of heavy metals in soil is analyzed;the basis of X-ray physics and the theory of quantitative and qualitative analysis are introduced,and different experiments are designed.In the experiment,by optimizing the detection conditions,the influence of some external conditions on the detection was effectively eliminated,and the characteristics of the obtained data were analyzed to provide theoretical and data support for constructing the soil heavy metal content inversion model under the influence of moisture content.Secondly,in view of the spectral overlap between some heavy metals found in the feature analysis,a Chaos Particle Swarm Optimization(CPSO)optimization analysis model based on Gaussian Mixture Statistical Model(GMSM)is proposed.Establish a GMSM mathematical model for the spectral overlapping peaks in the region of interest.This model combined with statistical ideas can explain the physical characteristics of X-rays.The GMSM can be optimized globally through CPSO to achieve the analysis of spectral overlapping peaks,and finally the results will be analyzed.Compared with the actual results,a more accurate overlapped peak analytical correction model was established.Finally,a non-negative matrix factorization(NMF)model is proposed for moisture content,a factor that cannot be eliminated through experimental conditions.The original energy spectrum matrix is decomposed in sections according to the multiplicative iterative update rule,the peak signal-to-noise ratio is used for evaluation,and finally the fusion spectrum is synthesized.After NMF processing,the spectral data set is regressed using partial least squares(PLS),and then the net peak area and water content of the test set are input into the fitting function,and the output predicted heavy metal content is compared with the actual content.The result is very small.Compared with other moisture content removal algorithms,the accuracy is higher and the effect is better.Finally,the relationship between the net peak area and the content of heavy metals in the soil was established to realize the establishment of the inversion model of the content of heavy metals in the soil under the influence of moisture content.