Industrial beets are an emerging feedstock for biofuel and bioproducts industry in the US. Milling of industrial beets is the primary step in front end processing for ethanol production. Milled beets undergo color change during juice extraction. A custom designed computer vision system for measurement of milled beet color kinetics, consisting of a digital camera, custom-designed non-reflective enclosure, and color calibration method was developed in the present study. Beet samples of five different total soluble solids (TSS) contents were prepared by washing with cold water for color kinetics measurement. An artificial neural network model was used for converting the red, green, and blue (RGB) values of the acquired sample images to L*a*b* values. Seven color parameters (L*, a*, b*, hue, chroma, browning index (BI), and total color change (ΔE)) were analyzed. Page, user-defined polynomial, and fractional conversion models gave better fits for the experimental data with color parameters than the zeroth order, first order, exponential, and Peleg kinetic models. Of the color parameters studied, L* (Page R2>0.99, user-defined polynomial R2>0.97, and fractional R2>0.93) and ΔE (Page R2>0.98, user-defined polynomial R2>0.94, and fractional R2>0.85) gave the best description for the color change kinetics of milled beets. Developed TSS prediction models from the color measurements based on Page model constant, kp with color parameter, L* gave good prediction (R2=0.99), also did the simple linear model based on direct color values (R2=0.98). Measurement and mathematical modeling of milled industrial beets color kinetics will serve as important quality assessment tool in processing the beets for various renewable fuel and products applications.
Read full abstract