Computer vision system and graphic design tools for obtention of segmented and averaged food samples images in CIE-L*a*b* coordinates

Abstract

To obtain segmented and averaged food samples images in CIE-L*a*b* coordinates a lighting system assembly was carried out. A CCD digital camera was selected and conditions for the images acquisition were established. The color of 23 samples of food was determined. Preprocessing, region-based segmentation and averaged techniques for images were employed by means of the Adobe® Photoshop® CS3 Extended software and were compared with a reference instrument. Coincidences and differences among the images obtained, the appreciations of the observers and indicated color coordinates values in the literature in relation to the luminosity and the negative or positive character of the coordinates a* and b* were determined. The used proceeding permitted the obtention of uniform and non-uniform food samples images of dark and clear colors, averaged in CIE-L*a*b* coordinates; presented applications in the color variations comparison among different samples or not of a same food, in crops ripening stages evaluation, in the color coordinates results transformation of other studies to images and can be implemented in the determination of physical characteristics of quality in samples of food.