Measurement of the geometrical features and surface color of rapeseeds using digital image analysis

Digital image analysis was applied to determine the geometrical features and color of rape seeds surface, and to discriminate some impurities, that are difficult to separate in the cleaning process. The paper notices on methodological aspects, and the experiment described constitutes the first stage of studies on the possibility of applying digital image analysis to rapeseed quality estimation, so the results obtained should be treated as preliminary. The geometrical features of seeds and their color were analyzed using the LUCIA G ver. 4.8 software. It was found that variation in geometrical dimensions of seeds was much lower than in color of their surface, so minimum sample size utilized for color measurements should be larger. The surface color of seeds was feature that insufficiently differentiates seeds of different dimensions. Only small seeds were characterized by somewhat changed distribution of color on their surface. An analysis of color of rape and stickywilly seeds in RGB (red/green/blue) model showed distinct differences in value ranges, enabling to distinguish between these seeds. Surface color of mature, immature and broken seeds cannot be used to distinguish these fractions. (c) 2005 Elsevier Ltd. All rights reserved.