Application of digital images to determine color in honey samples from Argentina

Honey color is the first quality attribute evaluated by consumers and is an important sensory property in the beekeeping market. The most commonly used method for color determination is based on optical comparison, employing a Pfund colorimeter. This method is laborious, time consuming and is based on a subjective assessment. In order to obtain a fast and objective measurement of honey color, an analytical method based on digital images analysis combined with multivariate calibration (Partial Least Squares) was developed. Three color models, RGB (red-green-blue), HSB (hue-saturation-brightness) and Grayscale, were used to analyze the digital images. The optimum results were obtained by using HSB color model (r2 = 0.97 and RMSEP = 2.46), indicating that the digital image analysis combined with multivariate calibration is an excellent strategy for quantifying color in honey samples.