Research On The Detection Method Of Complex Solution Component Concentration Based On Hyperspectral Image

Because the traditional spectral analysis method has not obtained effective results in the detection of chemical components in complex solutions with a large number of scattering particles,this study uses the spatial information changes caused by the absorption and scattering characteristics of complex solutions to increase the spectral information,and proposes and uses The method of hyperspectral image detects and analyzes the concentration of complex solution components.This study improves the accuracy and reliability of modeling and significantly enhances the ability to analyze complex solution concentrations.The main research work is as follows:(1)Firstly,in order to simultaneously obtain the absorption and scattering information exhibited by the non-circumferentially symmetric hyperspectral image of a complex solution,a wedge-shaped sample pool was designed and fabricated.In this study,a complex solution hyperspectral image detection platform was built based on light source,wedge-shaped sample pool and Headwall Photonics G4-332 imaging spectrometer.The correctness and stability of the experimental system are proved by the parameter setting of the hyperspectral image acquisition and the stability analysis and test of the experimental system.(2)Secondly,in order to prove that the complex solution hyperspectral image detection platform can effectively detect the concentration of components containing the scattering characteristic solution,this study conducted a two-dimensional intensity distribution data by experimenting with a solution containing scattering characteristics through a complex solution hyperspectral image detection platform.In this study,the two-dimensional intensity distribution data collected and analyzed are modeled and analyzed.Compared with other modeling methods,the two-dimensional intensity distribution modeling method increases the correlation coefficient by 2.48% and the mean square error by 6.89%.The experimental results show that the spectral quantitative analysis method based on the wedge-shaped sample pool can detect the concentration of the solution containing the scattering characteristic solution,which lays a research foundation for the next multi-dimensional hyperspectral solution quantitative analysis method.(3)Thirdly,in order to improve the detection accuracy of the complex solution concentration,this study obtained the absorption and scattering information of the sample to be tested constructed by multi-wavelength two-dimensional intensity distribution through the hyperspectral image complex solution detection platform and multi-wavelength detection method.Multi-dimensional hyperspectral imagery enables fast simultaneous acquisition of multi-position and multi-dimensional hyperspectral images of complex solutions.In this study,the dimensionality-reduced data was modeled and analyzed by two multi-dimensional hyperspectral image dimensionality reduction methods.By comparing the data dimensionality reduction method based on wavelength recombination and the multi-dimensional spectroscopy data dimensionality reduction method,the correlation coefficient is increased by 1.27%,and the mean square error is reduced by 3.46%.The experimental results show that the data dimensionality reduction method based on wavelength recombination can effectively improve the accuracy and speed of complex solution component concentration detection in analytical chemistry.The results of this study show that the complex solution concentration detection method based on hyperspectral image realizes the advantages of fast,non-contact,non-polluting and on-line detection.The number of wavelengths of the model is extended,the multi-dimensional hyperspectral data dimension is reduced,and the prediction accuracy of the model is improved,which is of great significance in the field of complex solution detection and analysis.