A novel digital color analysis method for rapid glucose detection

Abstract

The spectrophotometer is the most used analysis equipment in traditional colorimetric methods. However, the operation of using a spectrophotometer is time-consuming and labor-intensive, which presents practical difficulties in rapid detection. To this end, we present a digital color analysis method, using the typical 3,5-dinitrosalicylic acid (DNS) method for glucose detection as an example. The primary colors from 3 color spaces (Red-Green-Blue, Hue-Saturation-Value, Hue-Saturation-Intensity) were studied as quantitative analytical parameters for the glucose concentration and the red color (from the Red-Green-Blue colorspace) of the assay image provides superior prediction precision (>99.8%). Combined with the color analysis, two calculation algorithms, nonlinear regression and artificial neural networks, were compared for the detection of a high concentration of glucose. Then a microtiter plate (48-well plate) platform, based on the color analysis, was set up. Compared to existing methods using a spectrophotometer, the digital color analysis method has a large detection range (0–10 g L⁻¹), high accuracy (0.07 g L⁻¹) and fast detection rate (150 samples detected within about 15 min). It also shows great promise for use in a variety of reducing sugar measurements such as xylose, fructose and maltose. These aforementioned features render this newly developed method highly suitable for quick detection applications.