RESUMEN
The wedge-shaped sample cell, by offering a comprehensive representation of scattering information in turbid media, significantly enhances the informational content conveyed by spectral images compared to flat sample cells. To further refine the accuracy of turbid medium component detection utilizing wedge-shaped sample cells, this work undertakes modeling and analysis of the influence of different wedge angles on detection precision. In this study, employing a 5° gradient in the incident angle of light, we investigate the impact of incident angles ranging from 10 to 45° on the turbid medium component analysis. Validation experiments are performed by utilizing solutions of Indian ink and fat emulsion at varying ratios. Experimental findings demonstrate that under identical experimental conditions, the wedge-shaped sample cell model at an incident angle of 35° yields optimal analysis results. Utilizing partial least-squares regression (PLSR) for the corresponding optical parameters, the highest value of Rp reached 0.980, with an RMSEP of 0.002. When compared to the model with a 30° incident angle, Rp increased by 0.033, and RMSEP decreased by 0.008. In comparison to the flat sample cell model, Rp increased by 0.041, and RMSEP decreased by 0.004. This study, through continuous variation of wedge angles and PLSR modeling and prediction, further enhances the accuracy of turbid medium component detection, laying an experimental foundation for subsequent analysis of turbid medium components based on wedge-shaped sample cells.
