| As the lifeblood of contemporary industry,oil spills occur during the process of oil extraction,transportation,and use,which damage the ecological environment.Therefore,how to quickly detect the types and relative content of pollutants in sea surface oil spills is of great significance for sea surface oil spill detection and management.Based on the threedimensional fluorescence spectrum detection technology and Laser-induced fluorescence detection technology,this paper conducts the research on the quantitative and qualitative detection of oil spill pollutants on the sea surface based on the optimized parallel factor algorithm and neural network algorithm.Firstly,explain the principle of fluorescence spectroscopy detection and the feasibility of fluorescence detection for oil spills on the sea surface;This paper introduces the spectrum of the three-dimensional fluorescence spectrum technology and Laser-induced fluorescence spectrum technology used in this paper,as well as the use of instruments(Laser-induced fluorescence instrument is built by the laboratory itself);Preprocessing the spectral data used lays the foundation for further detection of oil spills on the sea surface.Secondly,for the detection of secondary substances in mixed oil samples,it is necessary to analyze their fluorescence spectra.In order to solve the problem that the number of samples has a great impact on the trilinear decomposition of three-dimensional fluorescence spectrum by the second-order correction algorithm PARAFAC,this paper uses the linear interpolation method to expand the number of samples,and on this basis,the accuracy of quantitative analysis results is improved.Thirdly,for the detection of the main substances in mixed oil samples,it is not only necessary to consider the issue of fluorescence quenching,but also to consider how to perform peak pickup in complex spectral situations.This article proposes to use the 2DPCA algorithm to reconstruct complex spectra for different situations.The fluorescence spectra reconstructed by 2DPCA reduce redundant information.Based on this,peak picking is combined with an optimized GRNN network to achieve quantitative analysis of samples,providing a new approach for the quantitative analysis of complex fluorescence spectra.Finally,in order to address the issue of inconvenient portability of three-dimensional fluorescence spectrometers and the gap between them and the actual detection requirements for oil spills at sea,LIF technology was introduced.The LIF system built in the laboratory was used to achieve the identification of oil types in sea surface oil spill water containing oil emulsions,and research on the emulsification stage of sea surface oil spill pollutants was developed. |
