Measuring the sugar content in sugar beet is challenging because of the labor, expenses, and chemicals required. Hence, there is a necessity for a more efficient method to measure this content. This study aims to establish an efficient, non-invasive method for detecting sugar content in sugar beets using hyperspectral imaging, which could revolutionize quality control in the sugar beet industry. This study used hyperspectral imaging to analyze 400-950 nm sugar beet paste. Pre-processing techniques such as SNV (Standard Normal Variate) and SG (Savitzky-Golay), along with wavelength selection methods such as SPA (Successive Projection Algorithm) and CARS (competitive adaptive reweighted sampling), were applied. Furthermore, various regression models including MLR (multiple linear regression), PLS (Partial Least Squares regression ( , and SVR (Support Vector Regression) were employed for prediction. Evaluating these models based on R2 and RMSE criteria, the PLS regression model with SNV pre-processing and SPA wavelength selection stood out, achieving an R2 value of 0.91% and RMSE of 0.24. These findings suggest the potential of hyperspectral imaging as a rapid and accurate means of determining sugar content in sugar beet across the VIS-NIR spectrum.

Keywords