Detection of total bilirubin based on multi-wavelength excitation fluorescence spectroscopy

Abstract

Due to the overlapping absorption lines of all components in the blood, the spectrum of total bilirubin is largely obscured by strongly absorbing components, making it difficult to obtain high-precision detection results. Therefore, this article proposes a detection method for multi wavelength excitation fluorescence spectroscopy. The experiment collected fluorescence spectra of 325 samples at four wavelengths and established a total bilirubin model. In the experimental results, compared with the single wavelength method, the multi wavelength method reduced the root mean square error of the modeling set and prediction set by 50.5 % and 55.7 %, respectively, and reduced the mean absolute error of the sample prediction results to 2.121 $\mu mol/L$. The experimental results show that this article improves the spectral line difference coefficient between target and non-target components by increasing the information carried by the spectrum and the constraint equation of data analysis, while providing a complete research idea and method for the detection of other blood components with fluorescence effects.