UV-Vis Spectroscopy For Escherichia Coli In Milk Study On Accuracy Of Bacterial Population Detection

Escherichia coli is the food-borne pathogenic bacteria with the highest detection rate in milk,E.coli has important significance in accurately capturing the specific information,detecting the number of its flora accurately and quickly,and improving the detection accuracy of dairy products.In the field of microbial detection,the spectral detection method is widely used due to its high efficiency and non-destructive advantages.However,because its detection spectrum is easily affected by complex background components,the spectral characteristic peaks are red-shifted and covered.In order to improve the detection results Accuracy,to solve the problem that the one-dimensional spectrum cannot accurately pick up the characteristic wavelength of the microbe-specific information in the spectrum during detection.This article uses two-dimensional correlation analysis to pick up and analyze the spectral characteristics of E.coli in milk;The nonlinear problems caused and the difficulty in selecting model parameters were established by PLS method,RBF method and GA-RBF method to establish the prediction model of the number of Escherichia coli bacteria in milk for optimization.The main research contents of the paper are as follows:Aiming at the problem of baseline shift caused by uneven light distribution of milk in the UV-Vis spectrum acquisition experiment,the standard normal variable transformation was used to correct the baseline shift;for the noise generated by the instrument in the experiment,wavelet was used to remove The noise method effectively filters out.The experiment was designed to obtain the UV-Vis spectrum of E.coli in different concentrations in milk,and it was calibrated using the national standard plate counting method to obtain the true value of E.coli flora.Analyze the experimental characteristics,select several pre-processing methods to pre-process the original spectrum data,through comparison,we can conclude that the above two methods can more effectively filter out the noise interference of the experiment,reduce the measurement error,improve the detection accuracy,and enhance the UVvisible spectrum Signal-to-noise ratio.In order to obtain the characteristic wavelength more representative of the specific information of E.coli,the two-dimensional correlation analysis method is used to extract the feature of the pre-processed spectrum to separate the overlapping spectrum of milk and E.coli.According to Noda rules,in the two-dimensional synchronous correlation spectrum Select characteristic wavelengths that characterize E.coli specific information.Explore the correlation between the UV-visible spectrum and the number of E.coli bacteria in milk and the accuracy of the model.Based on the PLS method,RBF method and GA-RBF method of E.coli in the milk,the characteristic wavelength corresponding absorbance value and the number of E.coli bacteria in the milk prediction model.The results show that,compared with the traditional linear regression modeling PLS model(86.954%),the RBF model can fully consider the nonlinear problems caused by the complex milk components(91.034%),and the GARBF model that optimizes the RBF model parameters,More effectively improve the prediction accuracy of the number of E.coli bacteria in milk(96.052%).