The influence of illumination parameters on the performances of color sorting machines

Abstract

The color recorded in the image is not an inherent value of observed object, because it is also influenced by the illumination properties, as well as geometry and surfaces of neighboring objects. Numerous studies have developed and investigated image processing procedures in color sorting machines, where as not many of them have considered the influence of illumination on numerical values related to the colors. In this paper, parameters related to the color of corn have been examined. In each image, the corn has been illuminated by one of four different types of light sources. And different luminous intensities of each type of light source have been applied. The processing of obtained images has been performed in MATLAB, and parameters of images in RGB, CIE L*a*b* and HSV color space have been analyzed. Further descriptive statistics analysis has been performed by IBM SPSS. The variations of parameters with the change of light intensity, showed no statistical significance. The change of the type of the light source has a significant impact on all analyzed features.